STM32F103 logic analyzer and more
buck50 is open-source firmware that turns a “Blue Pill” STM32F103 development board (widely available for approx. US$1.50) into a multi-purpose test and measurement instrument, including:
configure
commandlogic
commandoscope
commanddump
commandtrigger
commandmonitor
commandlines
configurationadcN
configurationsnumbers
command(*)gpio
commandusart
commandspi
commandi2c
commandipc
configurationpty
configurationsocket
configurationpulse
command(*)reset
commandbuck50: Test and measurement firmware for “Blue Pill” STM32F103 development board
Copyright (C) 2019,2020 Mark R. Rubin aka “thanks4opensource”
This file is part of buck50.
The buck50 program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
The buck50 program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with the buck50 program. If not, see https://www.gnu.org/licenses/gpl.html
buck50 is provided as ready-to-flash files in ELF (buck50.elf), binary (buck50.bin), and Intel hex (buck50.hex) formats in addition to source code (see Building from source, below).
Instructions on how to flash code onto microcontrollers such as the STM32F103 are beyond the scope of this document. Briefly, the STM32F103 cannot be flashed via its USB port (except 3), so other techniques must be employed.
Probably the best method is via the “Blue Pill” SWD port (see Hardware connections, below) using either a Black Magic Probe (website, wiki) or an ST-Link (website) or clone thereof (Google search). Note that the former is used with GDB from the GNU Arm Embedded Toolchain and the latter either with GDB or stlink
(GitHub repository) among other software front ends.
Alternately, the “Blue Pill” can be flashed via its USART1 serial port (RX on port PA10, TX on PA9, see below). This typically requires a USB-to-serial converter
(Google search) (or, ironically, another already-flashed “buck50” 4) as modern host computers rarely include an RS-232 port. Instructions and software required for doing so can be found at the stm32flash
SourceForge website, this tutorial web page, and/or a GitHub wiki, among other resources. If using a serial port method remember to change the BOOT0 jumper (see directly below) back to its “0” position after flashing with it set to “1”.
The following is a highly simplified, edited version of the incredibly useful image created by Rasmus Friis Kjeldsen 5, available at http://reblag.dk/stm32/ (direct link: http://reblag.dk/wordpress/wp-content/uploads/2016/07/The-Generic-STM32F103-Pinout-Diagram.pdf). The edited and original versions are included in this repository under terms in compliance with the original’s license.
WARNING! Incorrect hookups to the “Blue Pill” (or any other microcontroller’s) ports may cause permanent damage. Examples: Connecting a push-pull output port directly to a positive voltage source or ground sink, or to another push-pull output on the same or different MCU set to the opposite polarity. Readers of this document and users of the firmware it describes assume all responsibility for their actions.
The buck50 firmware configures and uses the “Blue Pill” STM32F103 GPIO ports as follows. Note that many pins have multiple uses depending on which buck50.py
host command is in effect, including changing roles to be either inputs, outputs, or bidirectional.
Pin(s) Type Purpose buck50.py commands
====== ============================== ================================= ==================
PB4-11 input weak pull-down digital input logic, monitor
PB10,11 in/out open-drain I2C i2c, monitor
PA0-7 output push-pull digital output numbers, gpio
PA0-7 input analog analog input oscope, monitor
PA1-3 output push-pull digital pulse outputs pulse
PA0 input weak pull-up usart CTS usart
PA1 output push-pull usart RTS usart
PA2 output push-pull usart TX usart
PA3 input weak pull-up usart RX usart
PA4 output push-pull SPI select output (master) spi, monitor
PA4 input floating SPI select input (slave) spi, monitor
PA5 output push-pull SPI clock (master) spi, monitor
PA5 input floating SPI clock (slave) spi, monitor
PA6 input floating/pull-up/-down SPI MISO spi, monitor
PA7 output push-pull SPI MOSI spi, monitor
PA8 output push-pull USART clock (synchronous) usart, monitor
PA9 output push-pull USART TX usart, monitor
PA10 input weak pullup USART RX usart, monitor
PB13 output push-pull sampling indicator, sync logic, oscope
PB14 output open-drain external/ganged trigger logic, oscope
PB15 output open-drain external/ganged trigger ack logic, oscope
PC13 output open-drain status/fault/sampling/triggering <all>, logic, oscope
PA11 output open-drain USB negative (do not connect) <all>
PA12 output open-drain USB positive (do not connect) <all>
See logic, monitor, numbers, spi, and usart, below.
reset blink
command (used to identify physical device)buck50.py
’s help reset ganged=
, help reset ext-trig
, and help trigger
, and logic command and reset command, below.buck50.py is (obviously) written in Python. Python3 (7) (3.6 to be precise). It was developed under Linux and should be portable any Linux distribution or other UNIX-like operating system. External library and file system assumptions (/dev/ttyACMn
, /dev/pts/NN
) may require modification for use under macOS. It is unlikely that buck50.py
will function in a Microsoft Windows® environment without fairly extensive code modifications.
(Note to the impatient: If reading sequentially, skip ahead to TL;DNR (“Too Long, Did Not Read”), below.)
Good question. The original intent for buck50 was to integrate the host-based driver into an existing open source waveform viewer program. This was abandoned due to limitations in the ones I could find (see Waveform viewing software, below), and also as the scope of buck50 grew beyond it’s original beginnings as a logical analyzer and evolved into the many-headed Hydra it eventually became. (Its capabilities no longer fit into the basic design of a logic analyzer program.)
In addition, many of those capabilities are inherently text based (monitoring, SPI and USART bridges, etc). Sure, any decent GUI library provides a scrolling text widget, but a text-based program inherently leverages the features of the many excellent Linux terminal emulator programs available. As an analogy, consider the difference between the typical on-screen GUI calculator program that mimics a handheld device with only one numeric display line compared to a full featured text-based program like bc
(or, for that matter, Python).
The buck50.py
interface was expressly written for fast and easy usage in an active test and measurement environment. Think of it as an ugly-looking GUI with many, many “hot key” shortcuts. (Just like the GUI calculators: What’s faster? Typing numbers on a keyboard or mousing around clicking virtual GUI buttons? Also 8.) Try buck50.py
, and if dissatisfied but still want to use the firmware, feel free to write a GUI driver (respect the GPL license above). A wrapper layer around the buck50.py
module would be fairly easy to implement.
All the above notwithstanding, buck50.py
does attempt to integrate nicely with external waveform viewing software (see below). If gnuplot
is installed on the host system, the code will by default to automatically dump captured digital or analog data to a temporary file and load that (see gnuplot, below) with the appropriate commands 9. Other viewers may be manually configured/enabled as well.
The buck50 firmware and buck50.py
driver implement many complex capabilities. The interface is thus necessarily complex but has been designed to allow simple default usage when fine-grained control is not needed.
Output 4Hz incrementing binary counter on output ports PA0 through PB77:
$1.50: numbers
<ENTER> to halt ...
Monitor input ports PB4 through PB11:
$1.50: monitor
<ENTER> to end
0.000 00111001 ... ... blu grn yel ... ... brn
0.250 00111010 ... ... blu grn yel ... red ...
0.500 00111011 ... ... blu grn yel ... red brn
0.750 00111100 ... ... blu grn yel org ... ...
1.000 00111101 ... ... blu grn yel org ... brn
1.250 00111110 ... ... blu grn yel org red ...
1.500 00111111 ... ... blu grn yel org red brn
1.750 01000000 ... vio ... ... ... ... ... ...
2.000 01000001 ... vio ... ... ... ... ... brn
2.250 01000010 ... vio ... ... ... ... red ...
2.500 01000011 ... vio ... ... ... ... red brn
2.750 01000100 ... vio ... ... ... org ... ...
3.000 01000101 ... vio ... ... ... org ... brn
3.250 01000110 ... vio ... ... ... org red ...
Logic analyzer, capture first 32 signal level changes on ports PB4 through PB11:
$1.50: logic edges=32
logic: edges=32
Waiting for sampling finish (<ENTER> to abort) ...
Triggered at state #0: 32 samples (6.26MHz) in 0.00 seconds. Stopped by number of samples.
The captured sample data will be automatically uploaded and an external viewer program launched if configured (see above). Otherwise (or in addition), data can be manually uploaded and saved, for example:
$1.50: dump digital-frmt=vcd file=capture
logic samples: 0...31 of 32 @ 6.26MHz ... saving to file capture.vcd
Logic analyzer, as above, but trigger on rising edge of input port PB8
$1.50: trigger 0=xxx0xxxx-1-0 # wait for PB8 low
$1.50: trigger 1=xxx1xxxx-0-1 # wait for PB8 high
$1.50: logic edges=32
logic: edges=32
Waiting for sampling finish (<ENTER> to abort) ...
Triggered at state #1: 32 samples (6.26MHz) in 5.18 seconds. Stopped by number of samples.
logic samples: 0...31 of 32 @ 6.26MHz ... saving to file /tmp/temp.csv
$1.50: usart
<ENTER> to input, then input ascii text with optional "\n", "\t", and "\x<two hex digits>" escape sequences ...
received data
more received data
-- <ENTER> key pressed here --
"END" to finish: data to send
yet more received data
-- <ENTER> key pressed here --
"END" to finish: END
flushing USB CDC-ACM pipe for 1 second (<ENTER> to abort) ...
or alternately:
$1.50: usart ipc external pty enable
ipc: external
pty: enable
pty /dev/pts/41 ready
I/O active, press <ENTER> to exit ...
and then run an external terminal application (screen
, miniterm
), or more simply:
$ cat /dev/pts/41 & cat > /dev/pts/41
I2C bridge (slave @ address 31, initialize register 0x5b with 1, register 0x2a with 0x39, then read 6 byte wide register 0x33 twice):
$1.50: i2c master
i2c: master
Input format: <slave address> <rx size> <tx data <2hex|3dec|1ascii>) ...> [; ...]
"0 0" to end: 0x1f 0 5b 01 ; 0x1f 0 2a 39
stat= OK tx=2 rx=0 :
stat= OK tx=2 rx=0 :
"0 0" to end: 0x1f 6 33
stat= OK tx=1 rx=6 : 255-ff-. 215-d7-. 254-fe-. 044-2c-, 003-03-. 208-d0-.
"0 0" to end: 0x1f 6 33
stat= OK tx=1 rx=6 : 255-ff-. 212-d4-. 254-fe-. 043-2b-+ 003-03-. 212-d4-.
"0 0" to end: 0 0
Continuously read register after above I2C address 31 device initialization:
$1.50: monitor pb4-11=disabled rate=1s i2c master addr=31 tx-data=33 rx-size=6
i2c: master
monitor: pb4-11=disabled
monitor: rate=1Hz=1s
i2c: addr=31
i2c: tx-data=33
i2c: rx-size=6
<ENTER> to end
0.001 i:ff.d9.fe.2e.03.d5
1.002 i:ff.d5.fe.2c.03.d4
2.003 i:ff.d7.fe.2f.03.d5
3.003 i:ff.d1.fe.28.03.cf
4.004 i:ff.d2.fe.31.03.d4
5.005 i:ff.d9.fe.2a.03.dd
6.006 i:ff.d5.fe.2c.03.d9
7.007 i:ff.d2.fe.27.03.d0
8.008 i:ff.cd.fe.2a.03.da
...
Output 1kHz square waves on port PA1 through PA3, rising edges at 250, 500, and 750 ms respectively in total period:
$1.50: pulse enable
Capture analog values from inputs PA4 and PA5, triggering on positive slope at 1.65V on channel PA4
$1.50: oscope trgr-chnl=4 scnd-chnl=5 trigger=0:3.3:0.15:0.01
oscope: trgr-chnl=4
oscope: scnd-chnl=5
Waiting for sampling finish (<ENTER> to abort) ...
3257 samples, 2 channels (PA4,PA5) at 239.5+12.5@12MHz->47.6kHz in 3.83s wall clock time. Stopped by number of samples.
Samples 0...3256 (3257 total) 2 channels (PA4,PA5) per sample at 239.5+12.5@12MHz->47.6kHz
After reading the above, an reasonable reaction might be: You want me to type all those complicated commands? How is that “fast and easy” usage?
A valid criticism … but there is hope. That’s why you, the sophisticated user, chose to read this section. ;) Remember all that stuff that got printed when the program started?
buck50.py 0.9.2
Copyright 2020 Mark R. Rubin aka "thanks4opensource"
This is free software with ABSOLUTELY NO WARRANTY.
For details type "warranty" or "help warranty".
Type "using" for program usage.
Type "help" for commands, configurations, parameters, values.
...
Well, after reading the full license (you did type “warranty” or “help warranty”, didn’t you?), those with inquisitive minds might try:
$1.50: using
Input:
- Type "<command> ..." followed by ENTER key
- Abbreviate any command, configuration, parameter, or value to its unique prefix
Example:
m d=i r=4k a w=2 spe=6
same as:
monitor duration=infinite rate=4kHz adc weight=2 speed=600kHz=1.66667μs
- "#" in line: rest of line ignored
Completion:
key(s) result:
------ -------
<TAB> Completes partially typed command, configuration, parameter,
or value (if unique prefix), else ...
<TAB><TAB> ... sounds terminal bell and second <TAB> prints list
of possible completions
...
The using
help natters on for quite a while (violating this section’s "Not Too Much More” promise) with examples, information on input editing, history, and more. But the two important points are:
TAB
key completion.#2 is the “FTW” secret. If you take away nothing else from reading this document, remember:
TAB
KEY!!! (10)That, along with the the program’s built-in help system, largely make this entire README.md
redundant. The help system is hierarchical, starting with general information:
$1.50: help
Help:
Commands, Configurations, Actions, Parameters, Values:
<command> [[<action>] [<parameter>=<value>] ...] [<configuration> [<action>]
[<parameter>=<value>] ...] ... [# comment ...]
Syntax:
symbol explanation
------ -----------
<command> command name
[...] optional
[<action>] optional action name
[<parameter>=<value>] optional parameter=value names
<configuration> configuration name
... optional repeats
[# comment ...] text from "#" to end of line ignored
Verbose explanation:
- Program executes Commands
- Commands use Parameter=Value settings
- Parameter=Value settings are grouped into Configurations
- Configurations may have optional Actions
- Parameters have default Values
- Parameters retain their Values until changed by new Parameter=Value setting,
"configure load", or program exit/restart.
- Commands have a primary Configuration, with same name as the Command
- Commands may have one or more secondary Configurations
- Primary configuration is initial current Configuration
- Configuration remains current until changed by explicit new Configuration
- Only current Configuration's Parameters=Values can be set
- Configurations may be used by multiple Commands
- Only one Action allowed per Configuration per Command
...
and continuing on with examples and hints on how to dive deeper into the help hierarchy.
Beyond that, experimenting with the commands and their configurations and parameters (and the TAB
key!) should suffice to get started. Or continue with reading the next section …
Kudos to those who ended up (or skipped directly to) here. A brief review 11, particularly for the latter, of the buck50.py
program’s usage features …
First, any input can be fully typed, or abbreviated to its unique prefix. Each of the following sets of lines produces an identical result:
$1.50 h
$1.50 he
$1.50 help
$1.50 m
$1.50 mon
$1.50 monitor
$1.50: n o=p r=1k m=b
$1.50: num op=pu ra=1kh mo=bin
$1.50: numbers open-pull=push-pull rate=1kHz mode=binary
Note that all commands echo any changed parameters, both as a reminder and more importantly to confirm the expansion of any abbreviations:
$1.50: n o=p r=1k m=b
numbers: open-pull=push-pull
numbers: rate=1kHz=1ms
numbers: mode=binary
$1.50: num op=pu ra=1kh mo=bin
numbers: open-pull=push-pull
numbers: rate=1kHz=1ms
numbers: mode=binary
$1.50: numbers open-pull=push-pull rate=1kHz mode=binary
numbers: open-pull=push-pull
numbers: rate=1kHz=1ms
numbers: mode=binary
Second, the keyboard TAB
key completes any unambiguous partially-typed entry, possibly up a common prefix of multiple valid choices. It sounds the system “bell” otherwise, and when pressed a second time presents a list of valid choices (if any). TAB
can be used with blank space to give a full list of valid entries in the current context.
$1.50: s<TAB>
$1.50: spi <TAB><TAB>
ascii-num= endian= mode= pull= snoop= xmit-only=
baud= idle= nss-time= rate= socket
busy-wait= ipc nss= rx-wait= speed=
disable master phase= select= spi
end= miso= pty slave tx-data=
$1.50: spi b<TAB><TAB>
baud= busy-wait=
$1.50: spi ba<TAB>
$1.50: spi baud=
1.125MHz 2.25MHz 36MHz 562.5KHz
18MHz 281.25KHz 4.5MHz 9MHz
$1.50: spi baud=2<TAB><TAB>
2.25MHz 281.25KHz
$1.50: spi baud=28<TAB>
$1.50: spi baud=281.25KHz
The painstaking example above is more complicated and time-consuming to go through than actually using the TAB
key in the buck50.py
itself. Just keep pressing TAB
and follow the suggestions.
As described in the previous section, buck50.py
features an extensive build-in help system. The following is a (largely depth-first) guided tour through it, with occasional brief additional comments.
First, however, the program needs to be started from a UNIX shell …
Commandline arguments:
$ ./buck50.py --help
usage: buck50.py [-h] [-v] [-q] [-f [FILE]] [-x]
[--halt {monitor,gpio,usart,spi,i2c,numbers}]
[acm]
positional arguments:
acm CDC/ACM device (with or without "/dev" or "/dev/tty"
prefix) (default: /dev/ttyACM0)
optional arguments:
-h, --help show this help message and exit
-v, --version Print program/protocol version and exit (default:
False)
-q, --quiet Don't print startup info (default: False)
-f [FILE], --file [FILE]
Config file to load at startup (see "help configure
load" (default: None)
-x, --no-auto-viewer Disable automatic, pre-"--file" search for waveform
viewing software and setting of "logic autodump=",
"oscope autodump=", and "dump auto-digital= auto-
analog= viewer-csv= viewer-vcd= digital-frmt= analog-
frmt=" parameters. See "help xxx ..." for those
commands/configurations/parameters. (default: False)
--halt {monitor,gpio,usart,spi,i2c,numbers}
Experimental. Reset firmware if buck50.py exit while
command in progress. Must be done before buck50.py
without "--halt", and must specify correct command
name, else firmware/hardware reset and/or power cycle,
and USB re-enumeration required. (default: none)
Normal startup:
$ buck50.py
buck50.py 0.9.2
Copyright 2020 Mark R. Rubin aka "thanks4opensource"
This is free software with ABSOLUTELY NO WARRANTY.
For details type "warranty" or "help warranty".
Type "using" for program usage.
Type "help" for commands, configurations, parameters, values.
Connecting to buck50 device (press CTRL-C to abort ... )
Firmware identity match: 0xea017af5
Firmware version match: 0.9.2
Device serial number: 123456789abcdef987654321
Found /usr/bin/pulseview and /usr/bin/gnuplot. Have now set:
configure logic autodump=enabled oscope autodump=enabled dump output=file auto-digital=enabled auto-analog=enabled digital-frmt=csv viewer-csv=gnuplot viewer-vcd=pulseview
Note that the correct /dev
CDC/ACM USB special file corresponding to the STM32F103 “Blue Pill” device running the buck50 firmware must given as a commandline argument if it is not the default /dev/ttyACM0
. On Linux systems the default should be correct if no other CDC/ACM USB devices are connected (or the buck50 device was connected first, before the others).
If not, the correct /dev
file is reported when the “Blue Pill” is connected via USB. The utilities dmesg
, journalctl
, and others will show output similar to:
usb m-n.p.q: new full-speed USB device number N using xhci_hcd
usb m-n.p.q: New USB device found, idVendor=0483, idProduct=5740
usb m-n.p.q: New USB device strings: Mfr=1, Product=2, SerialNumber=3
usb m-n.p.q: Product: STM32 Virtual COM Port
usb m-n.p.q: Manufacturer: STMicroelectronics
usb m-n.p.q: SerialNumber: 1234567890abcdef01234567
cdc_acm m-n.p.q:1.0: ttyACMn: USB ACM device
where “m”, “n”, “p”, “q”, “N”, the SerialNumber
, and the “n” in ttyACMn
are arbitrary values.
If the “Blue Pill” is already plugged in and more than one /devttyACMn
file exists, various Linux utilities can be used to determine which is correct. For example:
$ for NAME in /dev/ttyACM* ; do
> echo $NAME
> udevadm info --name=$NAME | egrep '0483|5740|ID_SERIAL_SHORT'
> done
Each /dev/tttyACMn
corresponding to a buck50 “Blue Pill” will output lines of the form:
/dev/ttyACMn
E: ID_MODEL_ID=5740
E: ID_SERIAL_SHORT=1234567890abcdef01234567
E: ID_VENDOR_ID=0483
while non-“Blue Pill”/non-buck50 devices will not have the E: ID_MODEL_ID=5740
and ID_VENDOR_ID=0483
lines.
If multiple buck50 “Blue Pill”s are connected (which can be very useful 12) (type help reset ganged=
for a usage example), once again the easiest way to tell which /dev/ttyACMn
file corresponds to which is to plug them into USB sequentially and watch the dmesg
, etc. output.
After multiple buck50.py
programs are connected to the multiple buck50 devices via the appropriate /dev/ttyACM0
files, which one is which can be determined by the serial numbers reported at program startup. The serial numbers can matched to physical hardware by noting the dmesg
, etc. output, or by the udevadm
command described above.
Finally (and easier), the reset blink
command will blink the user LED of the connected “Blue Pill” device for identification purposes.
On macOS the device special files should be of the form /dev/tty.usbmodemWXYZ...
for some value of “WXYZ…” 13. As noted above, it is unlikely that the buck50.py
program will function in a Microsoft Windows® environment without significant revisions.
Review the license:
$1.50: warranty
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 3 of the License , or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, write to
The Free Software Foundation, Inc.
51 Franklin Street, Fifth Floor
Boston, MA 02110-1335 USA
$1.50: help
Help:
Commands, Configurations, Actions, Parameters, Values:
<command> [[<action>] [<parameter>=<value>] ...] [<configuration> [<action>]
[<parameter>=<value>] ...] ... [# comment ...]
Syntax:
symbol explanation
------ -----------
<command> command name
[...] optional
[<action>] optional action name
[<parameter>=<value>] optional parameter=value names
<configuration> configuration name
... optional repeats
[# comment ...] text from "#" to end of line ignored
Verbose explanation:
- Program executes Commands
- Commands use Parameter=Value settings
- Parameter=Value settings are grouped into Configurations
- Configurations may have optional Actions
- Parameters have default Values
- Parameters retain their Values until changed by new Parameter=Value setting,
"configure load", or program exit/restart.
- Commands have a primary Configuration, with same name as the Command
- Commands may have one or more secondary Configurations
- Primary configuration is initial current Configuration
- Configuration remains current until changed by explicit new Configuration
- Only current Configuration's Parameters=Values can be set
- Configurations may be used by multiple Commands
- Only one Action allowed per Configuration per Command
Examples:
logic
- Command with no Parameter=Value settings, Actions, or secondary
Configurations
usart baud=19.2kHz
- Command with Parameter=Value setting in primary Configuration
pulse time=125ms mode1=rise time1=15%
- Command with multiple Parameters=Values settings in primary Configuration
monitor lines 6=clock 7=data
- Command with multiple Parameter=Value settings in secondary Configuration
reset halt
- Command with Action in primary Configuration
monitor spi master
- Command with Action in secondary Configuration
monitor rate=10Hz pb4-11=disabled adc0 enable
s/h+adc=239.5+12.5@12MHz->47.6kHz spi slave miso=push-pull monitor
file=monitored.txt output=both
- Command with multiple Parameters=Values settings in primary Configuration,
two secondary Configurations each with a Parameter=Value setting and an
Action, then explicit primary Configuration with two more Parameter=Value
settings
Commands:
configure, trigger, logic, oscope, dump, monitor, pulse, gpio, usart, spi,
i2c, numbers, reset, warranty, help, using, quit
Configurations:
configure, monitor, pulse, numbers, gpio, usart, spi, i2c, adc0, adc1, adc2,
adc3, adc4, adc5, adc6, adc7, logic, oscope, dump, lines, reset, socket,
pty, ipc
Time/Frequency Errors:
- Many time and/or frequency parameters have limited precision due to their
STM32F103 hardware implementations (32 bit register values, etc). Arbitrary
Parameter=Values are rounded to the nearest achievable value and the error
difference reported. Examples:
31.125μs(error:-5.5ns)
- time is 5.5ns shorter than requested
3.13043MHz=319.444ns(error:-28.8452kHz,+2.91663ns)
- frequency is 28.8452kHz lower, time 2.91663ns longer, than requested
monitor
- Command with no Parameters=Values settings, Actions, or Configurations.
All current Parameter=Value settings in all Configurations remain in
effect.
More help:
help using
help <command>
help <command> <parameter>
help <command> <configuration>
help <command> <configuration> <parameter>
Commands, configurations, actions, parameters, values: What could be simpler? ;) Hopefully further help, below, experimenting using the program (and, as always 11, using the tab key) will clarify.
$1.50: help using
Help for command "using":
Input:
- Type "<command> ..." followed by ENTER key
- Abbreviate any command, configuration, parameter, or value to its unique
prefix
Example:
m d=i r=4k a w=2 spe=6
same as:
monitor duration=infinite rate=4kHz adc weight=2 speed=600kHz=1.66667μs
- "#" in line: rest of line ignored
Completion:
key(s) result:
------ -------
<TAB> Completes partially typed command, configuration, parameter,
or value (if unique prefix), else ...
<TAB><TAB> ... sounds terminal bell and second <TAB> prints list
of possible completions
Example:
n<TAB> i<TAB>3 o<TAB>p<TAB> m<TAB><TAB>g<TAB> l<TAB><TAB>o<TAB>16
completes:
numbers increment=3 open-pull=push-pull mode=gray low=16
(note could use abbreviations instead):
n i=3 o=p m=g lo=16
Example:
usart p<TAB><TAB>
prints list:
parity= phase= ports= pty
then:
usart pa<TAB>
completes:
usart parity=
Example (note space after "usart", before double <TAB> press):
spi <TAB><TAB>
prints list:
ascii-num= endian= mode= pull= snoop= xmit-only=
baud= idle= nss-time= rate= socket
busy-wait= ipc nss= rx-wait= speed=
disable master phase= select= spi
end= miso= pty slave tx-data=
then:
spi x<TAB>
completes:
spi xmit-only=
and:
spi xmit-only=<TAB><TAB>
prints list:
disabled enabled
and finally:
spi xmit-only=e<TAB>
results in:
spi xmit-only=enabled
Editing, history:
(only basics listed, web search "GNU readline" for more)
key result:
------------------------- -------
<LEFT-ARROW> or <CTRL+b> cursor left
<RIGHT-ARROW> or <CTRL+f> cursor right
<BACKSPACE> or <DELETE> delete character
<HOME> or <CTRL+a> cursor to start of line
<END> or <CTRL+f> cursor to end of line
<UP-ARROW> or <CTRL+p> previous input line
<DOWN-ARROW> or <CTRL+n> next input line
<CTRL+r> search input history
<CTRL+g> clear input line
Alternate "?" help syntax:
"?" anywhere in line same as "help <line>" (line truncated at "?")
"?" can immediately follow command/configuration/parameter/value or be
separated by space
Examples ("?", " ?", and " ? " always equivalent):
"?" syntax same as:
---------- --------
usart? help usart
usart ? help usart
numbers rate? help numbers rate
numbers rate=? help numbers rate=
numbers rate=2Hz? help numbers rate=2Hz
monitor adc3? help monitor adc3
configure usart baud? help configure usart baud
configure pulse time=1kHz spi? help configure spi
spi mode=? endian=lsb help spi mode=
monitor duration=? adc weight=3 help monitor duration
configure gpio end=XX rate=? pulse time=1s help gpio rate
More help:
Type "help" for general help on commands, configurations, actions, and
parameters
There it is: Abbreviations, the full TAB
key magic, the rest of the line editing capabilities, and a quicker way to get help. Commands and other elements in the following examples are fully typed, but remember that abbreviations or TAB
completion can always be used instead.
Once more, it cannot be overemphasized:
TAB
KEY OFTEN!!! 14See Future work, below, for small problem with “?” alternate help syntax implementation.
Also note that the non-intuitive command names (gpio
instead of parallel
, numbers
instead of counter
) were chosen in an attempt to make commands (and other strings) unique in their first letter (or first few) for easy TAB
completion.
Continuing the help system tour …
configure
commandThe configure
command displays, changes, saves, and or loads any parameters in any or all configurations.
$1.50: help configure
Help for command "configure":
View, change, load, and/or save settings
- View any or all configurations' parameters:
- Save or load all configurations' parameters to/from disk file
- Adjust all time output values (trim hardware CPU clock)
Command usage:
configure [<parameter>=<value> ...] [<action>] [<configuration>
<parameter>=<value ...] [<action>]
Configurations (primary="configure"):
configure, lines, numbers, gpio, usart, adc0, adc1, adc2, adc3, adc4, adc5,
adc6, adc7, spi, i2c, pulse, monitor, logic, oscope, dump, reset, socket,
pty, ipc
Type "help configure <configuration>" for configuration description,
parameters, and actions
Type "help configure <configuration> <parameter>" for parameter description
configure
configurationAs with all commands, configure
has its own default configuration, also named “configure” 15, with a set of parameters and actions:
$1.50: help configure configure
Help for configuration "configure" (e.g. "configure configure"):
Load and/or save all configuration parameters to file
- Will append ".b50" to file name if missing
- Supports <TAB> completion of filenames
Command/configuration usage:
configure [<parameter>=<value> ...] [<action>]
Actions:
save Save all configurations/parameters to file, name...
load Load all configurations/parameters to file, name...
adjust Adjust reported times by "trim" parameter value
Parameters:
file= Name of ".b50" file to load or save configuratio...
trim= Trim all reported times by factor of X/Y where X...
Configuration "configure" used by commands:
configure
Type "help configure" for command description
Type "help configure configure <parameter>" for parameter description
Type "help configure configure <action>" for action description
configure
parameters and actionsConfiguration help (for any configuration) truncates long explanations. To see the full text:
$1.50: help configure save
Help for action "save" (e.g. "configure configure save"):
- Save all configurations/parameters to file, name specified with"file="
parameter
Type "help configure configure" for list of configure configuration actions
Type "help configure" for command description
$1.50: help configure load
Help for action "load" (e.g. "configure configure load"):
- Load all configurations/parameters to file, name specified with"file="
parameter
Type "help configure configure" for list of configure configuration actions
Type "help configure" for command description
$1.50: help configure adjust
Help for action "adjust" (e.g. "configure configure adjust"):
- Adjust reported times by "trim" parameter value
Type "help configure configure" for list of configure configuration actions
Type "help configure" for command description
$1.50: help configure file=
Help for parameter "file=" (e.g. "configure configure file="):
- Name of ".b50" file to load or save configurations and parameters. ".b50"
suffix will be appended if missing. See "load" and "save" actions.
Current value:
Valid values: <any valid file or path name>
Type "help configure configure" for list of configure configuration parameters
Type "help configure" for command description
$1.50: help configure trim=
Help for parameter "trim=" (e.g. "configure configure trim="):
- Trim all reported times by factor of X/Y where X is currently reported time
and Y is desired. Use "adjust" action to effect.
Current value: 1.00000/1.00000
Valid values: <X/Y with X and Y floating point numbers>
Type "help configure configure" for list of configure configuration parameters
Type "help configure" for command description
Customization of buck50.py
can be done by changing any of the various commands' configuration parameters and saving with configure file=<filename> save
16. The customizations can manually loaded with configure file=<filename> load
, or automatically at buck50.py
startup with the --file=
commandline option.
Note that buck50.py
has no explicit “reset defaults” command or action. Again, this can be easily implemented via configure file=<filename> save
before any configuration changes (and without commandline --file=
) and reloaded as desired.
Note that the configure
command is useful not only for changing its own configuration parameters, but also for changing other command’s configurations' parameters without executing those commands. Examples:
$1.50: configure trim=1.00/0.97
configure: trim=1.00000/0.970000
configure:
file : default.b50
trim : 1.00000/0.970000
$1.50: configure monitor duration=90s
monitor: duration=1.5m
monitor:
rate : 4Hz=250ms
duration : 1.5m
pb4-11 : enabled
file :
output : both
printf : %7.3f
$1.50: configure lines 4=NSS 5=SCK 6=MISO 7=MOSI spi slave miso=push-pull tx-data=12.34.56
spi: slave
lines: 4=NSS
lines: 5=SCK
lines: 6=MISO
lines: 7=MOSI
spi: miso=push-pull
spi: tx-data=12.34.56
lines:
4 : NSS
5 : SCK
6 : MISO
7 : MOSI
8 : grn
9 : blu
10 : vio
11 : gry
spi:
mode : slave
xmit-only : disabled
snoop : disabled
select : software
baud : 281.25KHz
endian : msb
idle : low
phase : 1st
miso : push-pull
pull : floating
speed : 2MHz
nss : floating
tx-data : 12.34.56
rate : unlimited
busy-wait : 1ms
rx-wait : 10ms
nss-time : zero
ascii-num : numeric
end : END
Moving on to the actual functional, results–producing commands and their configurations …
logic
command $1.50: help logic
Help for command "logic":
Digital logic analyzer.
- capture samples of logic levels on ports PB4(lsb) through PB11(msb)
- samples recorded only if level changes on any port
- samples also recorded every 0.233016889 seconds regardless if any change
- sampling starts when trigger condition met (see "help trigger")
- sampling continues until first of:
. user interrupt (<ENTER> key)
. duration elapsed ("logic duration=")
. number of samples ("logic edges=") (incl. extra samples @ 233ms)
. memory full
- see "help logic mode" for sampling speed
- see "help trigger" for triggering
- see "help reset ganged=" and "help reset ext-trig" for external triggering and
multiple device synchronization
- on-board user LED off at start of triggering, back on when sampling finished
- port PB13 toggles at c. 4.29 Hz during sampling, can monitored by external
device, drive LED at 3mA max, and/or aid post-processing of multiple device
captures
- Use "dump" command to upload, print, save, and/or view samples.
Command usage:
logic [<parameter>=<value> ...] [<action>]
Configuration: logic
Type "help logic logic" for configuration description, parameters, and actions
Type "help logic <parameter>" for parameter description
logic
configuration $1.50: help logic logic
Help for configuration "logic" (e.g. "logic logic"):
Digital (logic analyzer) capture parameters.
Command/configuration usage:
logic [<parameter>=<value> ...]
Parameters:
mode= Sampling mode...
duration= Sampling time limit
edges= Maximum number of digital samples (including ext...
code-mem= Sampling code memory bank. "ram" faster, "flash"...
autodump= Automatically (CAUTION!) do "dump" command after...
Configuration "logic" used by commands:
configure
logic
Type "help logic" for command description
Type "help logic <parameter>" for parameter description
logic
parameters $1.50: help logic mode=
Help for parameter "mode=" (e.g. "logic logic mode="):
- Sampling mode
- digital digital ports PB4(lsb)...PB11(msb)
- all clock counts @ 72MHz main CPU)
- see https://github.com/thanks4opensource/buck50/#why_so_slow
- if "code-mem=flash":
6.26MHz : sequence 5@11+1@14 CPU clocks/sample (72*6/69 = 6.26MHz)
irregular : alternating 11 + 14 " " " (72*2/25 = 5.76MHz)
uniform : constant 15 " " " (72 /15 = 4.80MHz)
4MHz : constant 18 " " " (72 /18 = 4.00MHz)
- if "code-mem=ram":
6.26MHz : sequence 5@13+1@17 CPU clocks/sample (72*6/69 = 5.27MHz)
irregular : alternating 13 + 17 " " " (72*2/30 = 4.80MHz)
uniform : constant 17 " " " (72 /17 = 4.24MHz)
4MHz : constant 18 " " " (72 /18 = 3.43MHz)
Current value: 6.26MHz
Valid values: "6.26MHz", "irregular", "uniform", or "4MHz"
Type "help logic logic" for list of logic configuration parameters
Type "help logic" for command description
Limitations of the digital sampling speeds is a long and unhappy topic. See Why so slow?, below.
$1.50: help logic duration=
Help for parameter "duration=" (e.g. "logic logic duration="):
- Sampling time limit
Current value: infinite
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", or "ns" suffix in range [910.222μs ... 59.6514s], or
"infinite"
Type "help logic logic" for list of logic configuration parameters
Type "help logic" for command description
Note the funky our
abbreviation for “hour” in the mandatory units suffix for time values. Units are required for all time and time-or-frequency parameters (see below for one of many examples of the latter). The rationale is that hour
is very rarely used while Hz
is very common – implementing our
instead of hour
avoids having to explicitly type “hz” vs “ho” to disambiguate between the two in abbreviations and TAB
completion. 17
$1.50: help logic edges=
Help for parameter "edges=" (e.g. "logic logic edges="):
- Maximum number of digital samples (including extra timing samples at c. 4Hz)
Current value: unlimited
Valid values: "unlimited" or integer in range [0 ... 65535]
Type "help logic logic" for list of logic configuration parameters
Type "help logic" for command description
$1.50: help logic code-mem=
Help for parameter "code-mem=" (e.g. "logic logic code-mem="):
- Sampling code memory bank. See "help logic logic" and
https://github.com/thanks4opensource/buck50/#flash_vs_ram
Current value: flash
Valid values: "ram" or "flash"
Type "help logic logic" for list of logic configuration parameters
Type "help logic" for command description
Sampling code can be run either directly from flash memory, or copied to RAM and executed there. Current tests show flash being faster despite generally accepted wisdom to the contrary, so there is little reason to use the code-mem=ram
option (it also slightly reduces available sample memory). See below for more details.
$1.50: help logic autodump=
Help for parameter "autodump=" (e.g. "logic logic autodump="):
- Automatically (CAUTION!) do "dump" command after completion. See "CAUTION!" in
"help dump auto-digital."
Current value: enabled
Valid values: "enabled" or "disabled"
Type "help logic logic" for list of logic configuration parameters
Type "help logic" for command description
oscope
commandThe oscope
command implements a primitive two channel digital storage oscilloscope. Bandwidth, sampling rate, and particularly sample storage depth are limited.
$1.50: help oscope
Help for command "oscope":
Primitive DSO (digital storage oscilloscope)
- samples one or two analog channels chosen from ports PB0...PB7
- primary/trigger channel: "trgr-chnl="
- second channel: "scnd-chnl=", (set to "none" for single channel operation)
- triggered at specified level (with hysteresis) and slope
. see "help oscope trigger=" and "help oscope slope="
. hysteresis implemented due to ADC conversion noise (otherwise slowly rising
or falling signals will trigger near level regardless of "slope=" setting)
(as noise dithers around level)
. trigger when signal rises above level after first falling below level minus
hysteresis if "slope=positive"
. trigger when signal falls below level after first rising above level plus
hysteresis if "slope=negative"
. use "slope=disabled" to trigger immediately without regard to level
- samples recorded continuously at rate specified by "s/h+adc="
- sampling continues number specified by "samples=" parameter
. triggering and/or sampling interruptible via <ENTER> key
- on-board user LED off, and output port PB13 toggles at c. 4 Hz, during
triggering and sampling
- Use "dump" command to upload, print, save, and/or view samples.
- See "help reset ganged=" and "help reset ext-trig" for external/sync'd
triggering.
Command usage:
oscope [<parameter>=<value> ...] [<action>] [<configuration>
<parameter>=<value ...] [<action>]
Configurations (primary="oscope"):
oscope, adc0, adc1, adc2, adc3, adc4, adc5, adc6, adc7
Type "help oscope <configuration>" for configuration description, parameters,
and actions
Type "help oscope <configuration> <parameter>" for parameter description
oscope
configuration $1.50: help oscope oscope
Help for configuration "oscope" (e.g. "oscope oscope"):
Analog DSO (digital storage oscilloscope) parameters
Command/configuration usage:
oscope [<parameter>=<value> ...]
Parameters:
samples= Maximum number of single or dual analog samples
trgr-chnl= Port number PA<x> (x=0...7) of trigger/first/sin...
scnd-chnl= Port number PA<x> (x=0...7) of second channel, o...
trigger= Triggering scaling, level, and hysteresis. Level...
slope= Analog triggering slope (or "disabled" for immed...
s/h+adc= Sampling rate, sample and hold time plus analog ...
time-scale= Scale factor for time values in "dump" command o...
printf= C format string for time values in "dump" comman...
autodump= Automatically do "dump" command after completion
Configuration "oscope" used by commands:
configure
oscope
dump
Type "help oscope" for command description
Type "help oscope oscope <parameter>" for parameter description
oscope
parameters $1.50: help oscope samples=
Help for parameter "samples=" (e.g. "oscope oscope samples="):
- Maximum number of single or dual analog samples
Current value: unlimited
Valid values: "unlimited" or integer in range [0 ... 65535]
Type "help oscope oscope" for list of oscope configuration parameters
Type "help oscope" for command description
$1.50: help oscope trgr-chnl=
Help for parameter "trgr-chnl=" (e.g. "oscope oscope trgr-chnl="):
- Port number PA<x> (x=0...7) of trigger/first/single channel
Current value: 0
Valid values: decimal or hex integer in range [0 ... 7]
Type "help oscope oscope" for list of oscope configuration parameters
Type "help oscope" for command description
$1.50: help oscope scnd-chnl=
Help for parameter "scnd-chnl=" (e.g. "oscope oscope scnd-chnl="):
- Port number PA<x> (x=0...7) of second channel, or "none" for single channel
sampling
Current value: none
Valid values: "none" or integer in range [0 ... 15]
Type "help oscope oscope" for list of oscope configuration parameters
Type "help oscope" for command description
$1.50: help oscope trigger=
Help for parameter "trigger=" (e.g. "oscope oscope trigger="):
- Triggering scaling, level, and hysteresis. Level and hysteresis in scaling
(low-to-high) units/range.
Current value: 0:3.3:1.65:0.05
Valid values: <four ":"-separated floating point values>
(<low>:<high>:<level>:<hysteresis>)
Type "help oscope oscope" for list of oscope configuration parameters
Type "help oscope" for command description
As noted in help oscope
above, hysteresis is implemented due ADC input and conversion noise. If slope=positive
, the signal must fall to level - hysteresis
and then rise back above level
to trigger. If slope=negative
, above level + hysteresis
then below level
. Without hysteresis or with it set too low, slowly changing signals will trigger immediately as they approach level
regardless of slope
(as noise sends the measurement above and below the level
value).
Level and hysteresis are expressed in trigger=
low-to-high units/range.
$1.50: help oscope slope=
Help for parameter "slope=" (e.g. "oscope oscope slope="):
- Analog triggering slope (or "disabled" for immediate start of sampling)
Current value: positive
Valid values: "disabled", "positive", or "negative"
Type "help oscope oscope" for list of oscope configuration parameters
Type "help oscope" for command description
$1.50: help oscope s/h+adc=
Help for parameter "s/h+adc=" (e.g. "oscope oscope s/h+adc="):
- Sampling rate, sample and hold time plus analog conversion time. Limited set
of values supported by hardware.
Current value: 239.5+12.5@12MHz->47.6kHz
Valid values: "1.5+12.5@12MHz->857kHz", "7.5+12.5@12MHz->600kHz",
"13.5+12.5@12MHz->462kHz", "28.5+12.5@12MHz->293kHz",
"41.5+12.5@12MHz->222kHz", "55.5+12.5@12MHz->176kHz",
"71.5+12.5@12MHz->143kHz", or "239.5+12.5@12MHz->47.6kHz"
Type "help oscope oscope" for list of oscope configuration parameters
Type "help oscope" for command description
$1.50: help oscope time-scale=
Help for parameter "time-scale=" (e.g. "oscope oscope time-scale="):
- Scale factor for time values in "dump" command output
Current value: 1000000.0
Valid values: <any floating point number>
Type "help oscope oscope" for list of oscope configuration parameters
Type "help oscope" for command description
$1.50: help oscope printf=
Help for parameter "printf=" (e.g. "oscope oscope printf="):
- C format string for time values in "dump" command
Current value: %10.2f μs
Valid values: <"printf"-style format string with <= 64 total characters for use
with single float argument. Must have single "%" followed by
valid printf format character. May contain arbitrary additional
text before and/or after printf format, including optional escape
sequences in the form \C character, \xMN hexadecimal, \ABC octal,
and/ or \uABCD unicode escapes) (use "\x20" or "\040" for space)
(use "%%" for single literal non-format "%" character). Examples:
"%.2f", "%10.4e", "%g seconds", "control: %6.4f \x03bcV",
"%8.3f%%"
Type "help oscope oscope" for list of oscope configuration parameters
Type "help oscope" for command description
$1.50: help oscope autodump=
Help for parameter "autodump=" (e.g. "oscope oscope autodump="):
- Automatically do "dump" command after completion
Current value: enabled
Valid values: "enabled" or "disabled"
Type "help oscope oscope" for list of oscope configuration parameters
Type "help oscope" for command description
dump
commandThe dump
command is used for uploading, printing, and exporting captured samples.
$1.50: help dump
Help for command "dump":
Upload most recent captured samples (from "logic" or "oscope" commands) to
terminal and/or file.
Digital ("logic") samples:
- Initial line, range of samples. Example:
logic samples: 0...31 @ of 32 (max 4842)
- One line per sample. Example:
23 0xd3 11010011 gry vio ... grn ... ... red brn 622 8.63889μs
\--/ \--/ \------/ \-----------------------------/ \------/ \-------/
| | | | | |
| | | +-- port name ("configure | |
| | | names <bit>=<name>") if | |
| | | port high else "..." | |
| | +-- state of ports PB11(msb) to PB4(lsb) | |
| +-- state of ports PB11(msb) to PB4(lsb) | |
+-- sample number | |
number of 72MHz ticks since --+ |
previous sample |
elapsed time since --+
previous sample
Analog ("oscope") samples:
- Initial line, range of samples. Example:
Samples 0...59 of 60 (max 4842) 2 channels (PA0,PA1) per sample at
239.5+12.5@12MHz->47.6kHz
- One line per one- or two-channel samples. Example:
37 1554.00 μs PA0 1.631V PA1 77.9°F
\--/ \----------/ \--------/ \--------/
| | | |
| | | +-- second channel value
| | +-- trigger channel value
| +-- sample time
+-- sample number
Command usage:
dump [<parameter>=<value> ...] [<action>] [<configuration>
<parameter>=<value ...] [<action>]
Configurations (primary="dump"):
dump, lines, oscope, adc0, adc1, adc2, adc3, adc4, adc5, adc6, adc7
Type "help dump <configuration>" for configuration description, parameters, and
actions
Type "help dump <configuration> <parameter>" for parameter description
dump
configurationThe dump
configuration:
$1.50: help dump dump
Help for configuration "dump" (e.g. "dump dump"):
Dump samples to terminal and/or file.
Command/configuration usage:
dump [<parameter>=<value> ...]
Parameters:
begin= First sample number
count= Number of samples to upload, starting with "begi...
file= File to write samples to
digital-frmt= Digital dump file format
analog-frmt= Analog dump file format
actives= List of "logic" ports to dump to file (does not ...
tick-units= Time units per tick for digital "vcd" format. No...
per-tick= Counts per tick for digital "vcd" format. If usi...
output= Dump output destination
viewer-csv= External viewer program for CSV files
viewer-vcd= External viewer program for VCD files
other-csv= External viewer program for CSV files when "view...
other-vcd= External viewer program for VCD files when "view...
auto-digital= Automatically (CAUTION!) run external viewer pro...
auto-analog= Automatically run external viewer program. See "...
warn-pulseview= Warn if estimated "digital-frmt=vcd" memory usag...
linewidth= analog gnuplot line width
pointtype= analog gnuplot point type
pointsize= analog gnuplot point size (0 for no points)
Configuration "dump" used by commands:
configure
dump
Type "help dump" for command description
Type "help dump dump <parameter>" for parameter description
dump
parameters $1.50: help dump begin=
Help for parameter "begin=" (e.g. "dump dump begin="):
- First sample number
Current value: 0
Valid values: decimal or hex integer in range [0 ... 65535]
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump count=
Help for parameter "count=" (e.g. "dump dump count="):
- Number of samples starting with "begin="
Current value: unlimited
Valid values: "unlimited" or integer in range [1 ... 65535]
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump file=
Help for parameter "file=" (e.g. "dump dump file="):
- File to write samples to
Current value:
Valid values: <any valid file or path name>
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump digital-frmt=
Help for parameter "digital-frmt=" (e.g. "dump dump digital-frmt="):
- Digital dump file format
Current value: vcd
Valid values: "csv" or "vcd"
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump analog-frmt=
Help for parameter "analog-frmt=" (e.g. "dump dump analog-frmt="):
- Analog dump file format
Current value: csv
Valid values: csv
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump actives=
Help for parameter "actives=" (e.g. "dump dump actives="):
- List of "logic" ports to dump to file (does not affect terminal or "oscope"
output)
Current value: 4.5.6.7.8.9.10.11
Valid values: 1 to 8 "."-separated integers in range 4...11
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump tick-units=
Help for parameter "tick-units=" (e.g. "dump dump tick-units="):
- Time units per tick for digital "vcd" format. Note only "us", "ns", and "fs"
valid for `pulseview`. See "help dump per-tick"
Current value: ns
Valid values: "s", "ms", "us", "ns", "ps", "fs", "as", "zs", or "ys"
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump per-tick=
Help for parameter "per-tick=" (e.g. "dump dump per-tick="):
- Counts per tick for digital "vcd" format. If using "pulseview", set this
parameter and "tick-units" as large as possible without exceeding the shortest
anticipated time period between samples. See
https://github.com/thanks4opensource/buck50/#pulseview
Current value: 125
Valid values: decimal or hex integer in range [1 ... 1000000000000000]
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
See buck50.py VCD/pulseview parameters, below, for discussion regarding per-tick
parameter.
$1.50: help dump output=
Help for parameter "output=" (e.g. "dump dump output="):
- Dump output destination
Current value: file
Valid values: "terminal" or "file"
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump viewer-csv=
Help for parameter "viewer-csv=" (e.g. "dump dump viewer-csv="):
- External viewer program for CSV files
Current value: gnuplot
Valid values: "gnuplot", "pulseview", "gtkwave", or "other"
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump viewer-vcd=
Help for parameter "viewer-vcd=" (e.g. "dump dump viewer-vcd="):
- External viewer program for VCD files
Current value: pulseview
Valid values: "gnuplot", "pulseview", "gtkwave", or "other"
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump other-csv=
Help for parameter "other-csv=" (e.g. "dump dump other-csv="):
- External viewer program for CSV files when "viewer-vcd=other". May contain
"%s" character for filename. Use \040 for spaces between commandline arguments
Current value:
Valid values: <any string <= 1024 chars including optional \C character, \xMN
hexadecimal, \ABC octal, and/ or \uABCD unicode escapes) (use
"\x20" or "\040" for space)>
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump other-vcd=
Help for parameter "other-vcd=" (e.g. "dump dump other-vcd="):
- External viewer program for VCD files when "viewer-vcd=other". May contain
"%s" character for filename. Use \040 for spaces between commandline arguments
Current value:
Valid values: <any string <= 1024 chars including optional \C character, \xMN
hexadecimal, \ABC octal, and/ or \uABCD unicode escapes) (use
"\x20" or "\040" for space)>
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump auto-digital=
Help for parameter "auto-digital=" (e.g. "dump dump auto-digital="):
- Automatically (CAUTION!) run external viewer program. See "viewer_vcd=" and
"viewer_csv=" parameters. CAUTION! If "viewer_vcd=pulseview", pathological
"logic" sampling captures may cause almost unlimited host memory consumption
and possible system lockup. See
https://github.com/thanks4opensource/buck50/#pulseview
Current value: enabled
Valid values: "enabled" or "disabled"
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump auto-analog=
Help for parameter "auto-analog=" (e.g. "dump dump auto-analog="):
- Automatically run external viewer program. See "viewer_vcd=" and "viewer_csv="
parameters.
Current value: enabled
Valid values: "enabled" or "disabled"
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump warn-pulseview=
Help for parameter "warn-pulseview=" (e.g. "dump dump warn-pulseview="):
- Warn if estimated "digital-frmt=vcd" memory usage exceeds value and might
trigger host system slowdown or lockup when file loaded into `pulseview`. See
"dump auto-digital=" and
https://github.com/thanks4opensource/buck50/#pulseview
Current value: 100000000.0
Valid values: <any floating point number>
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump linewidth=
Help for parameter "linewidth=" (e.g. "dump dump linewidth="):
- analog gnuplot line width
Current value: 0.0
Valid values: <any floating point number>
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump pointtype=
Help for parameter "pointtype=" (e.g. "dump dump pointtype="):
- analog gnuplot point type
Current value: none
Valid values: "none", "plus", "x", "asterisk", "square-open", "square-filled",
"circle-open", "circle-filled", "delta-open", "delta-filled",
"nabla-open", "nabla-filled", "diamond-open", "diamond-filled",
"arrow-open", or "arrow-filled"
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
$1.50: help dump pointsize=
Help for parameter "pointsize=" (e.g. "dump dump pointsize="):
- analog gnuplot point size (0 for no points)
Current value: 0.0
Valid values: <any floating point number>
Type "help dump dump" for list of dump configuration parameters
Type "help dump" for command description
trigger
commandThe trigger
command allows viewing and modification of the triggering state machine used by the logic
command.
Far more than a GUI, buck50.py
needs a custom language and compiler for defining its trigger state machines. The state machine mechanism supports very sophisticated triggering scenarios, but defining the states is error prone and somewhat non-intuitive. Particularly difficult is implementing logical-OR (“A or B or C”) conditions. Hopefully the following description and examples will help:
$1.50: help trigger
Help for command "trigger":
Digital capture/sampling triggering.
Usage:
trigger list trigger states
trigger <n>=<abcdefgh>-<pass>-<fail> [text] define trigger state
trigger check check trigger states
trigger delete n[-m] [p[-q] ... delete trigger state(s)
trigger undo undo last define or delete
where:
state number 0...255
a,b,c,d,e,f,g,h state bits:
X = don't care alternate chars: pPUu.*
0 = port low alternate chars: oO
1 = port high alternate chars: iI
pass state number, 0...255
fail state number, 0...255
text optional comments
Examples:
trigger 0=xxxx101x-1-0
trigger 17=11001111-23-15
Trigger state machine pseudocode:
set state to 0
loop while not triggered:
if connected to other units and PB14 external trigger is low:
triggering complete, exit loop, start sample capture
else:
set input to ports PB4...PB11 (lsb...msb)
set begin_state to state
loop while checking input:
if input matches state's bits (masked by "don't care" bits):
if state's "pass" state is 0:
triggering complete, exit both loops, start sample capture
else:
set state to state's "pass" state
exit inner "checking input" loop, continue outer loop
else:
set state to state's "fail" state
if state is same as begin_state:
exit inner "checking trigger" loop, continue outer loop
else:
continue inner "checking trigger" loop
Triggers:
- at least one trigger state, state #0, required
- default state machine: one trigger, "0=xxxxxxxx-0-0
- trigger state numbers arbitrary in range [0 ... 255]
- state numbers need not be contiguous (slight download speed advantage if so)
- "pass" and "fail" states must point to existing/defined trigger states
- chain of "pass" states back to state 0 (triggering complete) required
- logical "OR" implemented by chains of "pass" states looping back first in
chain or ending with "xxxxxxxx-X-Y" all "don't care" state
- warnings for above and other requirements reported by "trigger check" and at
start of "logic" command
- cannot trigger on implicitly clocked asynchronous (UART) serial data, only
synchronous/clocked (USART, SPI, etc)
- see "help logic mode" for sampling speed after triggering
- note triggering speed is slower than sampling speed, especially if multiple
logical "OR" states
- note minor additional slowdown if multiple units connected (see "help reset
ganged=")
Triggers examples:
Trigger immediately (default):
t 0=xxxxxxxx-0-0 any input matches, go to state 0, triggered
Trigger on bit pattern:
t 0=x1x101x0-0-0 trigger if pattern matches, else continue checking input
Trigger on rising edge on port 9
t 0=xx0xxxxx-1-0 wait until port PA9 is low
t 1=xx1xxxxx-0-1 trigger when PA9 goes high
Trigger on one bit pattern followed by second bit pattern
t 0=xxxx01xx-1-0 go to state 1 if pattern matches, else recheck
t 1=111xxxxx-0-1 trigger if new pattern matches, else recheck
Trigger on one bit pattern followed immediately by second bit pattern
t 0=xxxxxxxx-1-0 go to wait for state 1 match
t 1=xxxx01xx-2-1 go to state 2 if pattern matches, else recheck
t 2=111xxxxx-0-3 trigger if new pattern matches, else ...
t 3=xxxxxxxx-1-1 ... recheck first pattern
Trigger on serial 1-0-1 bitstream on PA8, clocked on rising edge on PA6
t 0=xxxxxxxx-1-0 start state machine
t 1=xxx0xxxx-2-1 wait until clock line PA8 low
t 2=xxx1xxxx-3-2 wait until clock line PA8 high
t 3=xxx1x1xx-4-0 continue if clocked PA6 bit is 1, else restart
t 4=xxx0xxxx-5-4 wait until clock line PA8 low
t 5=xxx1xxxx-6-5 wait until clock line PA8 high
t 6=xxx1x0xx-7-0 continue if clocked PA6 bit is 0, else restart
t 7=xxx0xxxx-8-7 wait until clock line PA8 low
t 8=xxx1xxxx-9-8 wait until clock line PA8 high
t 9=xxx1x1xx-0-0 trigger if clocked PA6 bit is 1, else restart
Trigger on binary 3, 4, or 10 on ports PA4(lsb) through PA7(msb)
t 0=xxxxxxxx-1-0 go to check for 3, 4, or 10
t 1=xxxx0011-0-2 trigger if binary 3, else check if binary 4
t 2=xxxx0100-0-3 trigger if binary 4, else check if binary 10
t 3=xxxx1010-0-1 trigger if binary 10, else end 3, 4, or 10 check
Trigger on 7-bit ascii "u", "$", or 'Z", followed immediately by "~", on
PA4(lsb) through PA10, all data clocked on falling edge on PA11
t 0=xxxxxxxx-1-0 start state machine
t 1=1xxxxxxx-2-1 wait for clock high
t 2=0xxxxxxx-3-2 wait for clock low
t 3=01110101-7-4 check if "u", else check if "$"
t 4=00100100-7-5 check if "$", else check if "Z"
t 5=01011010-7-6 check if "Z", else end "u" or "$" or "Z" check
t 6=xxxxxxxx-1-7 end "u" or "$" or "Z" check
t 7=1xxxxxxx-8-7 wait for clock high
t 8=0xxxxxxx-9-8 wait for clock low
t 9=01111110-0-0 trigger if "~", else restart from beginning
Never trigger (useful if only external triggering desired, see below):
t 0=xxxxxxxx-1-0 start state machine
t 1=00000000-2-1 go to state 2 on success
t 2=11111111-1-2 go to state 1 on success
External/sync'd triggering:
- See "help reset ganged=" and "help reset ext-trig"
Command usage:
trigger [<parameter>=<value> ...] [<action>]
Configuration: triggers
Type "help trigger triggers" for configuration description, parameters, and
actions
Type "help trigger <parameter>" for parameter description
Type "help trigger <action>" for action description
Regarding trigger check
, I’ll admit my “graph theory fu” is not quite at the ninja 18 level I’d like it to be. The attempt is to warn about trigger state machines that cannot trigger regardless of the inputs fed to them. Such a state machine will cause the buck50 firmware on the “Blue Pill” to pause indefinitely, although because …
logic
command does an automatic check before starting and issues the warning if necessary, and …ENTER
key), and …logic duration=
parameter is not infinite
, triggering will timeout and abort automatically… bad state machines are not a serious problem. Also, in certain circumstances it may be useful to purposely define such a state machine, as per the “external trigger only” example in the help text. If anything, trigger check
is likely overly restrictive in the warnings it issues.
trigger
configuration and actions $1.50: help trigger triggers
Help for configuration "triggers" (e.g. "trigger triggers"):
Command usage:
trigger <state number> <abcdefgh>-<pass state>-<fail state> [text]
where:
state number 0...255
a,b,c,d,e,f,g,h X = don't care alternate chars: pPUu.*
0 = port low alternate chars: oO
1 = port high alternate chars: iI
pass state 0...255
fail state 0...255
text optional comments
Type "help trigger"
Command/configuration usage:
trigger triggers [<parameter>=<value> ...] [<action>]
Actions:
check Check triggers for conflicts
delete Delete one or more triggers: n[-m] [p[-q] ...
undo Undo last trigger command change or delete
Type "help trigger" for command description
TAB
completion supports re-editing existing trigger states, and modifying them into new ones without excessive typing (19). The “alternate characters” are supported in case they make entering “1”, “0”, and “x” bits easier (adjacent keys on keyboard).
Moving past triggering, sampling, and dumping …
monitor
command $1.50: help monitor
Help for command "monitor":
Monitor and/or log digital, analog, spi, and/or usart inputs
Continue until "duration=<time>" (if not "infinite") elapsed, or user interrupt
via <ENTER> key.
Prints text line with monitored data to terminal ("output=terminal"), file
("file=<filename> output=file) or both ("file=<filename> output=both")
Lines printed if rate limiting ("rate=<time>") time elapsed and any of the
following conditions met:
digital:("pb4-11=enabled")
- level change on any digital port PB4 through PB11
analog: ("adcN enable")
- analog voltage change on enabled analog configs "adc<N>" for N in 0...7
- values filtered by "adcN exponent=", "weight=", and "hysteresis="
parameters
via formula:
filtered_voltage = filtered_voltage * (pow(2,exponent) - weight)
+ current_voltage * ( weight)
but only if:
abs(filtered_voltage - previous_printed_voltage) > hysteresis
- values printed using "adcN scale-hyst=" scaling, with "name=" name, and
"printf=" printf formula including optional additional text
- Example:
Port PA4 connected to analog temperature sensor with 0 to 3.3V output
corresponding to -20 to 160 degrees Fahrenheit:
monitor adc4 enable scale-hyst=-20:160:0.5 exponent=3 weight=5
name=outdoors printf=%5.1f\x20F
If scaled current value filtered with previous value using ratio of 5/8 to
3/8 is 82 and previous printed value was less than 81.5 or greater than 82.5
will print:
Xoutdoors: 82.0 F
(see below for "x" error status character, normally blank space)
usart: ("usart enable")
- "usart synchro=disabled": if RX data received (note also "usart
tx-data=<byte>" sent on TX port each "rate=<time>" period)
- "usart synchro=enabled": if read RX data read on port different than
previous
spi:
"mode=master": Send "tx-data=" on MOSI once per "rate=" period. Print
received MISO data if different than previous printed
"mode=slave": Print received MOSI data (having sent "tx-data=" on MISO
one-to-one with MOSI bytes)
i2c:
"mode=master": Send "tx-data=" on per "rate=" period. Print received slave
TX data if different than previous printed
"mode=slave": Print received master data. Also sends "tx-data=" (if any), as
slave-tx-to-master.
Prints one or more fields depending on enabled configurations/parameters:
all:
004.491
\-----/ seconds since command start ("monitor printf=<format code>")
digital: ("pb4-11=enabled") state of digital ports PB4 to PB11. Example:
10111011 gry ... blu grn yel ... red brn
\------/ \-----------------------------/
| |
| +-- port name ("configure names <bit>=<name>")
| if port high else "..."
+-- state of ports, PB11(msb) to PB4(lsb)
analog: ("adcN enable") voltage on analog port PA<N>. Example:
(see above for more details)
PA2:1.963V
|||||||||||
|||||++++++-- "scale-hyst=" scaled value, "printf=" formula
||||+-------- ascii ":" character
|+++-------- "name=" name
+----------- error code, normally blank (see below)
usart: ("usart enable=enabled"), RX data on port PA10
u:pr.rs...
|||||||||||
|||++++++++-- RX data bytes, length of "spi tx-data=",
||| two hexadecimal digits per byte, separated by "."
||+---------- ascii ":" character
|+----------- ascii "u" character
+------------ error code, normally blank (see below)
spi:
s:pr.rs...
|||||||||||
|||++++++++-- MISO or MOSI data bytes, length of "spi tx-data=",
||| two hexadecimal digits per byte, separated by "."
||+---------- ascii ":" character
|+----------- ascii "s" character
+------------ error code, normally blank (see below)
i2c:
mi:pr.rs...
||||||||||||
||||++++++++-- received data bytes (master or slave),
|||| two hexadecimal digits per byte, separated by "."
|||+---------- ascii ":" character
||+----------- ascii "i" character
|+------------ master:ascii "m" slave:"0"/"1"/"2"==gen-call/oar1/oar2
+------------- error code, normally blank (see below)
all:
(<ENTER> to abort) ...
error codes:
" " no error
"E" receive data register empty (usart, spi)
"B" peripheral busy
"O" receive data overrun
"T" send or receive data hardware timeout
"F" generic error
"L" line break (usart)
"N" noise (usart)
"P" parity (usart)
"U" framing (usart)
"S" start (i2c)
"A" address (i2c)
"N" nack (i2c)
"X" missing BTF ("byte transfer finished") flag (i2c)
"R" missing RXNE ("receive register not empty") flag (i2c)
"P" missing STOP (i2c)
Command usage:
monitor [<parameter>=<value> ...] [<action>] [<configuration>
<parameter>=<value ...] [<action>]
Configurations (primary="monitor"):
monitor, lines, usart, spi, i2c, adc0, adc1, adc2, adc3, adc4, adc5, adc6,
adc7
Type "help monitor <configuration>" for configuration description, parameters,
and actions
Type "help monitor <configuration> <parameter>" for parameter description
monitor
is probably the most complex and configurable of all the buck50.py
commands. Many different options. Its various output formats should be fairly self-explanatory in actual use. The spi
and usart
configuration parameters will be described below in the sections on the spi and usart commands.
See lines configuration, below, for info on the lines
configuration, also used by the monitor
command.
monitor
configuration $1.50: help monitor monitor
Help for configuration "monitor" (e.g. "monitor monitor"):
Command usage:
monitor [<parameter>=<value> ...]
Parameters:
rate= Max update frequency
duration= Halt after time elapsed
pb4-11= Monitor/log digital inputs
file= Log file name
output= Log to file ("file" or "both") only if "file=" i...
printf= Print format for time (seconds). Single C-style ...
Configuration "monitor" used by commands:
configure
monitor
Type "help monitor" for command description
Type "help monitor monitor <parameter>" for parameter description
monitor
parameters $1.50: help monitor rate=
Help for parameter "rate=" (e.g. "monitor monitor rate="):
- Max update frequency
Current value: "4Hz=250ms"
Valid values: floating point number with "y", "d", "h", "min", "s", "ms",
"us", "μs", "ns", "Hz", "kHz", "MHz", or "GHz" suffix in range
[72MHz=13.8889ns ... 0.1249nHz=253.881y], or "unlimited"
Type "help monitor monitor" for list of monitor configuration parameters
Type "help monitor" for command description
$1.50: help monitor duration=
Help for parameter "duration=" (e.g. "monitor monitor duration="):
- Halt after time elapsed
Current value: "infinite"
Valid values: floating point number with "y", "d", "h", "min", "s", "ms",
"us", "μs", or "ns" suffix in range [13.8889ns ... 253.881y], or
"infinite"
Type "help monitor monitor" for list of monitor configuration parameters
Type "help monitor" for command description
The pb4-11=
parameter allows disabling monitoring of those digital input lines if not needed which can be useful to reduce output clutter if analog
, usart
, and/or spi
monitoring is enabled in their respective configurations.
$1.50: help monitor pb4-11=
Help for parameter "pb4-11=" (e.g. "monitor monitor pb4-11="):
- Monitor/log digital inputs
Current value: "enabled"
Valid values: "enabled" or "disabled"
Type "help monitor monitor" for list of monitor configuration parameters
Type "help monitor" for command description
More monitor
parameters:
$1.50: help monitor file=
Help for parameter "file=" (e.g. "monitor monitor file="):
- Log file name
Current value: ""
Valid values: <any valid file or path name>
Type "help monitor monitor" for list of monitor configuration parameters
Type "help monitor" for command description
$1.50: help monitor output=
Help for parameter "output=" (e.g. "monitor monitor output="):
- Log to file ("file" or "both") only if "file=" is also set)
Current value: "both"
Valid values: "terminal", "file", or "both"
Type "help monitor monitor" for list of monitor configuration parameters
Type "help monitor" for command description
$1.50: help monitor printf=
Help for parameter "printf=" (e.g. "monitor monitor printf="):
- C format string for time (seconds).
Current value: %7.3f
Valid values: <"printf"-style format string with <= 64 total characters for use
with single float argument. Must have single "%" followed by
valid printf format character. May contain arbitrary additonal
text before and/or after printf format, including optional escape
sequences in the form \C character, \xMN hexadecimal, \ABC octal,
and/ or \uABCD unicode escapes) (use "\x20" or "\040" for space)
(use "%%" for single literal non-format "%" character). Examples:
"%.2f", "%10.4e", "%g seconds", "control: %6.4f \x03bcV",
"%8.3f%%"
Type "help monitor monitor" for list of monitor configuration parameters
Type "help monitor" for command description
lines
configurationThe lines
configuration parameters are used by dump
and monitor
commands to associate default or user chosen names with the PB4 through PB11 digital input ports. The names are also output to the terminal by monitor
and dump
, and included in some dump
output file formats for use when viewing with external Waveform viewing software.
As with all configurations/parameters, they can be viewed and changed using configure
, so help for their configuration can be accessed via that command:
$1.50: help configure lines lines
Help for configuration "lines" (e.g. "configure lines"):
Names for digital input ports PB4 through PB11.
Command usage:
configure lines [<parameter>=<value> ...]
Parameters:
4= Symbolic name for input pin PB4
5= Symbolic name for input pin PB5
6= Symbolic name for input pin PB6
7= Symbolic name for input pin PB7
8= Symbolic name for input pin PB8
9= Symbolic name for input pin PB9
10= Symbolic name for input pin PB10
11= Symbolic name for input pin PB11
Configuration "lines" used by commands:
configure
dump
monitor
Type "help configure" for command description
Type "help configure lines <parameter>" for parameter description
lines
parameters: $1.50: help configure lines 4=
Help for parameter "4=" (e.g. "configure lines 4="):
- Symbolic name for input pin PB4
Current value: "brn"
Valid values: <any string <= 16 chars, \xMN hex escapes allowed>
Type "help configure lines" for list of names configuration parameters
Type "help configure" for command description
and similarly for 5=
though 11=
.
The default names are:
$1.50: configure lines
lines:
4 : brn
5 : red
6 : org
7 : yel
8 : grn
9 : blu
10 : vio
11 : gry
Apologies that this omits “black” from the standard resistor code color-to-numeral mapping, but being a simple-minded software engineer, “black” irrevocably means “ground” in my mind (my buck50 “Blue Pills” are color-coded as above). The resistor color codes 20 are not as firmly burned into my brain as they are for EEs, and I’ll use as a further excuse the fact that the buck50 PB ports also don’t start at “0” (see Port assignments, below).
In any case, the names can be modified to suit (likely with signal names instead of colors – “clock”, “data”, etc) and saved/reloaded via configure save file=<filename>
and configure save ...
). Or the buck50.py
source code can be easily modified (open-source software and all that).
adcN
configurationsThe adc0
through adc7
configurations store parameters for ADC conversion on input ports PA0 through PA7, respectively. Several commands rely on some or all of these parameters.
Help for the adc0
configuration is listed below (adc1
through adc7
are analogous):
$1.50: help configure adc0
Help for configuration "adc0" (e.g. "configure adc0"):
Analog-to-digital input on port PA0.
Filtering/scaling:
- Filtering for "monitor" command (computed by firmware on STM32F103):
new_filtered_value = prev_filtered_value * (pow(2,exponent) - weight)
+ current_adc_value * ( weight)
but then not reported unless:
abs(new_filtered_value - prev_reported_value) > hysteresis
- Scaling for "monitor" and "dump" commands (computed by software on host):
normalized = value / float(max_12_bit_adc_value)
scaled = (1.0 - normalized) * lowscale + normalized * highscale
Command/configuration usage:
configure adc0 [<parameter>=<value> ...] [<action>]
Actions:
enable Set parameter "active=enabled"
disable Set parameter "active=disabled"
Parameters:
active= For "monitor" command
s/h+adc= Sampling rate, sample and hold time plus analog ...
scale-hyst= Scaling units/range for "monitor" and "dump" com...
exponent= Filter width. See "analog" section in "help moni...
weight= Filter width. See "analog" section in "help moni...
name= Port/channel name for "monitor", "oscope", and "...
printf= C format string to print scaled value. For "moni...
Configuration "adc0" used by commands:
configure
oscope
dump
monitor
Type "help configure" for command description
Type "help configure adc0 <parameter>" for parameter description
Type "help configure adc0 <action>" for action description
adcN
parameters and actions $1.50: help configure adc0 enable
Help for action "enable" (e.g. "configure adc0 enable"):
- Set parameter "active=enabled"
Type "help configure adc0" for list of adc0 configuration actions
Type "help configure" for command description
$1.50: help configure adc0 disable
Help for action "disable" (e.g. "configure adc0 disable"):
- Set parameter "active=disabled"
Type "help configure adc0" for list of adc0 configuration actions
Type "help configure" for command description
$1.50: help configure adc0 active=
Help for parameter "active=" (e.g. "configure adc0 active="):
- For "monitor" command
Current value: disabled
Valid values: "enabled" or "disabled"
Type "help configure adc0" for list of adc0 configuration parameters
Type "help configure" for command description
$1.50: help configure adc0 s/h+adc=
Help for parameter "s/h+adc=" (e.g. "configure adc0 s/h+adc="):
- Sampling rate, sample and hold time plus analog conversion time. Limited set
of values supported by STM32F103.
Current value: 239.5+12.5@12MHz->47.6kHz
Valid values: "1.5+12.5@12MHz->857kHz", "7.5+12.5@12MHz->600kHz",
"13.5+12.5@12MHz->462kHz", "28.5+12.5@12MHz->293kHz",
"41.5+12.5@12MHz->222kHz", "55.5+12.5@12MHz->176kHz",
"71.5+12.5@12MHz->143kHz", or "239.5+12.5@12MHz->47.6kHz"
Type "help configure adc0" for list of adc0 configuration parameters
Type "help configure" for command description
$1.50: help configure adc0 scale-hyst=
Help for parameter "scale-hyst=" (e.g. "configure adc0 scale-hyst="):
- Scaling units/range for "monitor" and "dump" commands, and hysteresis value
for "monitor"
Current value: 0:3.3:0
Valid values: <three ":"-separated floating point values>
(<low>:<high>:<hysteresis>)
Type "help configure adc0" for list of adc0 configuration parameters
Type "help configure" for command description
$1.50: help configure adc0 exponent=
Help for parameter "exponent=" (e.g. "configure adc0 exponent="):
- Filter width. See "analog" section in "help monitor".
Current value: 1
Valid values: decimal or hex integer in range [0 ... 8]
Type "help configure adc0" for list of adc0 configuration parameters
Type "help configure" for command description
$1.50: help configure adc0 weight=
Help for parameter "weight=" (e.g. "configure adc0 weight="):
- Filter width. See "analog" section in "help monitor".
Current value: 1
Valid values: decimal or hex integer in range [0 ... 255]
Type "help configure adc0" for list of adc0 configuration parameters
Type "help configure" for command description
$1.50: help configure adc0 name=
Help for parameter "name=" (e.g. "configure adc0 name="):
- Port/channel name for "monitor", "oscope", and "dump" commands' output.
Current value: PA0
Valid values: <any string <= 16 chars including optional \C character, \xMN
hexadecimal, \ABC octal, and/ or \uABCD unicode escapes) (use
"\x20" or "\040" for space)>
Type "help configure adc0" for list of adc0 configuration parameters
Type "help configure" for command description
$1.50: help configure adc0 printf=
Help for parameter "printf=" (e.g. "configure adc0 printf="):
- C format string to print scaled value. For "monitor" and "oscope" commands
Current value: %5.3fV
Valid values: <"printf"-style format string with <= 32 total characters for use
with single float argument. Must have single "%" followed by
valid printf format character. May contain arbitrary additonal
text before and/or after printf format, including optional escape
sequences in the form \C character, \xMN hexadecimal, \ABC octal,
and/ or \uABCD unicode escapes) (use "\x20" or "\040" for space)
(use "%%" for single literal non-format "%" character). Examples:
"%.2f", "%10.4e", "%g seconds", "control: %6.4f \x03bcV",
"%8.3f%%"
Type "help configure adc0" for list of adc0 configuration parameters
Type "help configure" for command description
numbers
commandThe poorly-name numbers
command (see above, should be “counter”) outputs sequential values on the “Blue Pill” GPIOB port lines:
$1.50: help numbers
Help for command "numbers":
Output binary values on ports PB-4 (lsb) to PB-11 (msb).
- Configurable rate, or unlimited (maximum firmware speed)
- Incrementing/decrementing values
- High/low limits
- Binary or gray code
Command usage:
numbers [<parameter>=<value> ...] [<action>]
Configuration: numbers
Type "help numbers numbers" for configuration description, parameters, and
actions
Type "help numbers <parameter>" for parameter description
numbers
configuration $1.50: help numbers numbers
Help for configuration "numbers" (e.g. "numbers numbers"):
"limited=" duration used if "rate=unlimited"
"unlimited=" duration otherwise
Command/configuration usage:
numbers [<parameter>=<value> ...]
Parameters:
mode= Numeric sequence encoding
increment= Change between successive values
low= Low value of low-to-high values range
high= High value of low-to-high values range
rate= Frequency/speed if not "rate= unlimited". See "T...
gpio-speed= Hardware slew rate
open-pull= Hardware port mode
limited= Command duration if "rate=limited"
unlimited= Command duration if "rate=unlimited"
Configuration "numbers" used by commands:
configure
numbers
Type "help numbers" for command description
Type "help numbers <parameter>" for parameter description
numbers
parameters $1.50: help numbers mode=
Help for parameter "mode=" (e.g. "numbers numbers mode="):
- Numeric sequence encoding
Current value: binary
Valid values: "binary" or "gray"
Type "help numbers numbers" for list of numbers configuration parameters
Type "help numbers" for command description
$1.50: help numbers increment=
Help for parameter "increment=" (e.g. "numbers numbers increment="):
- Change between successive values
Current value: 1
Valid values: decimal or hex integer in range [-127 ... 127]
Type "help numbers numbers" for list of numbers configuration parameters
Type "help numbers" for command description
$1.50: help numbers low=
Help for parameter "low=" (e.g. "numbers numbers low="):
- Low value of low-to-high values range
Current value: 0
Valid values: decimal or hex integer in range [0 ... 255]
Type "help numbers numbers" for list of numbers configuration parameters
Type "help numbers" for command description
$1.50: help numbers high=
Help for parameter "high=" (e.g. "numbers numbers high="):
- High value of low-to-high values range
Current value: 255
Valid values: decimal or hex integer in range [0 ... 255]
Type "help numbers numbers" for list of numbers configuration parameters
Type "help numbers" for command description
$1.50: help numbers rate=
Help for parameter "rate=" (e.g. "numbers numbers rate="):
- Frequency/speed if not "rate= unlimited". See "Time/Frequency Errors" section
in "help". Additional rate inaccuracies due to firmware speed limitations
at rates greater than approx. 100kHz.
Current value: 4Hz=250ms
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", "ns", "Hz", "kHz", "MHz", or "GHz" suffix in range
[72MHz=13.8889ns ... 16.7638milliHz=59.6523s], or "unlimited"
Type "help numbers numbers" for list of numbers configuration parameters
Type "help numbers" for command description
$1.50: help numbers gpio-speed=
Help for parameter "gpio-speed=" (e.g. "numbers numbers gpio-speed="):
- Hardware slew rate
Current value: 2MHz
Valid values: "2MHz", "10MHz", or "50MHz"
Type "help numbers numbers" for list of numbers configuration parameters
Type "help numbers" for command description
$1.50: help numbers open-pull=
Help for parameter "open-pull=" (e.g. "numbers numbers open-pull="):
- Hardware port mode
Current value: push-pull
Valid values: "push-pull" or "open-drain"
Type "help numbers numbers" for list of numbers configuration parameters
Type "help numbers" for command description
$1.50: help numbers limited=
Help for parameter "limited=" (e.g. "numbers numbers limited="):
- Command duration if "rate=limited"
Current value: infinite
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", or "ns" suffix in range [13.8889ns ... 8124.2y], or
"infinite"
Type "help numbers numbers" for list of numbers configuration parameters
Type "help numbers" for command description
$1.50: help numbers unlimited=
Help for parameter "unlimited=" (e.g. "numbers numbers unlimited="):
- Command duration if "rate=unlimited"
Current value: infinite
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", or "ns" suffix in range [910.222μs ... 59.6514s], or
"infinite"
Type "help numbers numbers" for list of numbers configuration parameters
Type "help numbers" for command description
gpio
command $1.50: help gpio
Help for command "gpio":
Output 8-bit parallel data on port PA4(lsb) through PA11(msb).
- Input line of ascii or numeric data ("gpio ascii-num=") then <ENTER> key to
send. See data entry instructions at command start.
- Line editing and history supported, type "help help" for more info.
- Enter "999" if "ascii-num=numeric" or end string ("end=" parameter) if
"ascii-num=ascii", followed by <ENTER> key to exit command.
Command usage:
gpio [<parameter>=<value> ...] [<action>] [<configuration>
<parameter>=<value ...] [<action>]
Configurations (primary="gpio"):
gpio, ipc
Type "help gpio <configuration>" for configuration description, parameters, and
actions
Type "help gpio <configuration> <parameter>" for parameter description
gpio
configuration $1.50: help gpio gpio
Help for configuration "gpio" (e.g. "gpio gpio"):
Configuration for Interactive 8-bit parallel data (ports PA4(lsb) through PA11(msb).
Command/configuration usage:
gpio [<parameter>=<value> ...]
Parameters:
speed= Output port slew rate
open-pull= Output port mode
rate= Byte output data rate. See "Time/Frequency Error...
ascii-num= Terminal input/output mode
end= Character sequence to exit command in "ascii-num...
Configuration "gpio" used by commands:
configure
gpio
Type "help gpio" for command description
Type "help gpio gpio <parameter>" for parameter description
gpio
parameters $1.50: help gpio speed=
Help for parameter "speed=" (e.g. "gpio gpio speed="):
- Output port slew rate
Current value: 2MHz
Valid values: "2MHz", "10MHz", or "50MHz"
Type "help gpio gpio" for list of gpio configuration parameters
Type "help gpio" for command description
$1.50: help gpio open-pull=
Help for parameter "open-pull=" (e.g. "gpio gpio open-pull="):
- Output port mode
Current value: push-pull
Valid values: "push-pull" or "open-drain"
Type "help gpio gpio" for list of gpio configuration parameters
Type "help gpio" for command description
$1.50: help gpio rate=
Help for parameter "rate=" (e.g. "gpio gpio rate="):
- Byte output data rate. See "Time/Frequency Errors" section in "help".
Current value: 10Hz=100ms
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", "ns", "Hz", "kHz", "MHz", or "GHz" suffix in range
[72MHz=13.8889ns ... 16.7638milliHz=59.6523s], or "unlimited"
Type "help gpio gpio" for list of gpio configuration parameters
Type "help gpio" for command description
$1.50: help gpio ascii-num=
Help for parameter "ascii-num=" (e.g. "gpio gpio ascii-num="):
- Terminal input/output mode
Current value: numeric
Valid values: "ascii" or "numeric"
Type "help gpio gpio" for list of gpio configuration parameters
Type "help gpio" for command description
$1.50: help gpio end=
Help for parameter "end=" (e.g. "gpio gpio end="):
- Character sequence to exit command in "ascii-num=ascii" mode
Current value: END
Valid values: <any string <= 16 chars including optional \C character, \xMN
hexadecimal, \ABC octal, and/ or \uABCD unicode escapes) (use
"\x20" or "\040" for space)>
Type "help gpio gpio" for list of gpio configuration parameters
Type "help gpio" for command description
usart
commandThe usart
command implements a bidirectional USB-to-serial converter. Input and output on the buck50.py
host side can be directed to/from either the buck50.py
terminal (see example usage above) or to UNIX pseudoterminals (/dev/pts/NN
) and/or UNIX sockets.
Note that if synchronous mode is enabled (see usart synchro=
, below), both connected devices must be in synchronous mode with the buck50 “Blue Pill” being the “master” (STM32F1xx chips do not support USART synchronous slave mode). Bytes are only received on the master device’s RX port in sync with those sent on TX — it is impossible for the master to receive without sending, or the slave to send without receiving. This is similar to the SPI protocol (see spi
command, below), and in fact compatible with it if usart datalen=8bits
and parity=none
(defaults both) are set.
Note also that when used as a usart-to-USB-to-UNIX-pseudoterminal bridge (see ipc
and pty
configurations, below), baud, parity, datalength, and stop bits, etc. must be set via the buck50.py
interface. They cannot be set via stty
, ioctl
, termios
, etc. APIs on the /dev/pts/XX
device.
$1.50: help usart
Help for command "usart":
Transmit and/or receive UART/USART serial data.
- Asynchronous/UART ("synchro=disabled") or synchronous/USART
("synchro=disabled" modes
- Uses one of two USART peripherals and sets of I/O ports:
"usart ports=pa8-10": TX=PA9 RX=PA10, CLK=PA8
"usart ports=0-30pa": TX=PA2 RX=PA3 CTS=PA0 RTS=PA1
- Synchronous clock output on port PA8 ("master" mode)
- Hardware does not support synchronous "slave" mode (clock input)
- Synchronous "master" mode with "datalen=8bits", "parity=none" can compatibly
connect to SPI slave device
- TX and RX data to and from terminal, or socket and/or pty (see "help usart
ipc term-ptysock=", "help usart socket", and "help usart pty")
- Usart xmit and/or recv data to socket and/or pty if either active, else to
terminal
Terminal I/O, asynchronous mode:
- If "recv=enabled" will print asynchronous recv data to terminal.
- <ENTER> key to pause recv data and enter line of ascii data and/or <\xXY\>
hexadecimal values, <ENTER> key to xmit.
Terminal I/O, synchronous mode:
- "master" mode: enter line of ascii data and/or <\xXY\> hexadecimal values,
<ENTER> key to xmit
- Will display recv data bytes (same number as xmitted) and prompt for next
line of xmit data
Terminal I/O, both modes:
- Input line of ascii or numeric data ("gpio ascii-num=") then <ENTER> key to
send. See data entry instructions at command start.
- Line editing and history supported, type "help help" for more info.
- Enter "999" if "ascii-num=numeric" or end string ("end=" parameter) if
"ascii-num=ascii", followed by <ENTER> key to exit command.
Command usage:
usart [<parameter>=<value> ...] [<action>] [<configuration>
<parameter>=<value ...] [<action>]
Configurations (primary="usart"):
usart, ipc, pty, socket
Type "help usart <configuration>" for configuration description, parameters,
and actions
Type "help usart <configuration> <parameter>" for parameter description
usart
configuration $1.50: help usart usart
Help for configuration "usart" (e.g. "usart usart"):
Configuration for async/sync (UART/USART) serial peripheral on ports PA9(TX),
PA10(RX) and PA8(clock) (if "ports=pa8-10") or ports PA2(TX), PA3(RX),
PA1(RTS) and PA0(CTS) (if "ports=pa0-3")
- Note "datalen=7bits" with "parity=none", or "datalen=9bits" with other than
"parity=none," not supported.
Command/configuration usage:
usart [<parameter>=<value> ...] [<action>]
Actions:
enable Set parameter "active=enabled"
disable Set parameter "active=disabled"
Parameters:
active= Used by monitor command
xmit= Transmit RX data on port PA9
recv= Receive RX data on port PA10
ports= Select one of two hardware USARTs (by PAx-y port...
baud= Baud rate for both xmit and recv if "synchro=ena...
datalen= Number of data bits. See "help usart usart" for ...
stoplen= Number of stop bits
parity= Parity. See "help usart usart" for restrictions ...
synchro= Async(UART)(no clock) or synchronous(USART) mode
idle= Clock output level at idle in "synchro=enabled" ...
phase= Clock edge to sample bit at in "synchro=enabled"...
lastclok= Last clock pulse in "synchro=enabled" mode
gpio= Ports output slew rate
rate= Byte output data rate. See "Time/Frequency Error...
tx-timeout= Wait for hardware TX ready. See "Time/Frequency ...
rx-wait= Wait for hardware RX ready. See "Time/Frequency ...
tx-data= Byte to xmit in "monitor" command if "synchro=en...
rx-len=
cts= CTS handshaking on port PA0 (only supported for ...
rts= RTS handshaking on port PA1 (only supported for ...
ascii-num= Interactive text input/output mode. See input fo...
end= Character sequence to exit command in "ascii-num...
snoop= Print socket traffic to terminal when "socket_co...
Configuration "usart" used by commands:
configure
monitor
usart
Type "help usart" for command description
Type "help usart usart <parameter>" for parameter description
Type "help usart usart <action>" for action description
usart
parameters and actions $1.50: help usart enable
Help for action "enable" (e.g. "usart usart enable"):
- Set parameter "active=enabled"
Type "help usart usart" for list of usart configuration actions
Type "help usart" for command description
$1.50: help usart disable
Help for action "disable" (e.g. "usart usart disable"):
- Set parameter "active=disabled"
Type "help usart usart" for list of usart configuration actions
Type "help usart" for command description
$1.50: help usart active=
Help for parameter "active=" (e.g. "usart usart active="):
- Used by monitor command
Current value: disabled
Valid values: "enabled" or "disabled"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart xmit=
Help for parameter "xmit=" (e.g. "usart usart xmit="):
- Transmit RX data on port PA9
Current value: enabled
Valid values: "enabled" or "disabled"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart recv=
Help for parameter "recv=" (e.g. "usart usart recv="):
- Receive RX data on port PA10
Current value: enabled
Valid values: "enabled" or "disabled"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart ports=
Help for parameter "ports=" (e.g. "usart usart ports="):
- Select one of two hardware USARTs (by PAx-y port numbers)
Current value: pa8-10
Valid values: "pa8-10" or "pa0-3"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart baud=
Help for parameter "baud=" (e.g. "usart usart baud="):
- Baud rate for both xmit and recv if "synchro=enabled", only xmit if
"disabled". See "Time/Frequency Errors" section in "help".
Current value: 9.6kHz=104.167μs
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", "ns", "Hz", "kHz", "MHz", or "GHz" suffix in range
[4.5MHz=222.222ns ... 1.09865kHz=910.208μs]
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart datalen=
Help for parameter "datalen=" (e.g. "usart usart datalen="):
- Number of data bits. See "help usart usart" for restrictions with "parity=".
Current value: 8bits
Valid values: "7bits", "8bits", or "9bits"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart stoplen=
Help for parameter "stoplen=" (e.g. "usart usart stoplen="):
- Number of stop bits
Current value: 1bit
Valid values: "1bit", "2bits", "0.5bits", or "1.5bits"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart parity=
Help for parameter "parity=" (e.g. "usart usart parity="):
- Parity. See "help usart usart" for restrictions with "datalen=".
Current value: none
Valid values: "none", "even", or "odd"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
usart synchro=
, see above
$1.50: help usart synchro=
Help for parameter "synchro=" (e.g. "usart usart synchro="):
- Async(UART)(no clock) or synchronous(USART) mode
Current value: disabled
Valid values: "enabled" or "disabled"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart idle=
Help for parameter "idle=" (e.g. "usart usart idle="):
- Clock output level at idle in "synchro=enabled" mode
Current value: low
Valid values: "low" or "high"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart phase=
Help for parameter "phase=" (e.g. "usart usart phase="):
- Clock edge to sample bit at in "synchro=enabled" mode
Current value: 1st
Valid values: "1st" or "2nd"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart lastclok=
Help for parameter "lastclok=" (e.g. "usart usart lastclok="):
- Last clock pulse in "synchro=enabled" mode
Current value: enabled
Valid values: "enabled" or "disabled"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart gpio=
Help for parameter "gpio=" (e.g. "usart usart gpio="):
- Ports output slew rate
Current value: 2MHz
Valid values: "2MHz", "10MHz", or "50MHz"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart rate=
Help for parameter "rate=" (e.g. "usart usart rate="):
- Byte output data rateSee "Time/Frequency Errors" section in "help".
Current value: unlimited
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", "ns", "Hz", "kHz", "MHz", or "GHz" suffix in range
[72MHz=13.8889ns ... 16.7638milliHz=59.6523s], or "unlimited"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart tx-timeout=
Help for parameter "tx-timeout=" (e.g. "usart usart tx-timeout="):
- Wait for hardware TX ready. See "Time/Frequency Errors" section in "help
help".
Current value: 10ms
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", or "ns" suffix in range [13.8889ns ... 59.6523s], or "none"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart rx-wait=
Help for parameter "rx-wait=" (e.g. "usart usart rx-wait="):
- Wait for hardware RX ready. See "Time/Frequency Errors" section in "help
help".
Current value: 10ms
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", or "ns" suffix in range [13.8889ns ... 59.6523s], or "none"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart tx-data=
Help for parameter "tx-data=" (e.g. "usart usart tx-data="):
- Byte to xmit in "monitor" command if "synchro=enabled"
Current value: 0
Valid values: decimal or hex integer in range [0 ... 255]
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart rx-len=
Help for parameter "rx-len=" (e.g. "usart usart rx-len="):
- (no help available for "usart usart rx-len")
Current value: 62
Valid values: decimal or hex integer in range [1 ... 62]
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart cts=
Help for parameter "cts=" (e.g. "usart usart cts="):
- CTS handshaking on port PA0 (only supported for "usart=pa0-3"
Current value: disabled
Valid values: "enabled" or "disabled"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart rts=
Help for parameter "rts=" (e.g. "usart usart rts="):
- RTS handshaking on port PA1 (only supported for "usart=pa0-3"
Current value: disabled
Valid values: "enabled" or "disabled"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart ascii-num=
Help for parameter "ascii-num=" (e.g. "usart usart ascii-num="):
- Interactive text input/output mode. See input format instructions printed at
command start.
Current value: ascii
Valid values: "ascii" or "numeric"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart end=
Help for parameter "end=" (e.g. "usart usart end="):
- Character sequence to exit command in "ascii-num=ascii" mode
Current value: END
Valid values: <any string <= 16 chars including optional \C character, \xMN
hexadecimal, \ABC octal, and/ or \uABCD unicode escapes) (use
"\x20" or "\040" for space)>
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
$1.50: help usart snoop=
Help for parameter "snoop=" (e.g. "usart usart snoop="):
- Print socket traffic to terminal when "socket_config[active]=enabled" and/or
"pty_config[active]=enabled"
Current value: disabled
Valid values: "disabled", "hexadecimal", or "ascii"
Type "help usart usart" for list of usart configuration parameters
Type "help usart" for command description
spi
commandSimilar to usart
command, above, the spi
command implements a bidirectional USB-to-SPI converter. The same host-side I/O capabilities apply.
A full description of the SPI protocol is outside the scope of this document, but similar restrictions as in the USART synchro discussion above apply to data send/recv in SPI master vs. slave configurations.
$1.50: help spi
Help for command "spi":
Send SPI MOSI and/or receive MISO data on ports PA7 and PA6.
- SPI master or slave mode
- SPI clock on port PA5 ( output if master, input if slave)
- SPI select ("NSS") on port PA4 (see "help spi select=" and "help spi nss=")
- MOSI and MISO data to and from terminal, or socket and/or pty (see "help spi
ipc term-ptysock=", "help spi socket", and "help spi pty")
Terminal I/O, master mode:
- Enter line of numeric or ascii data (format as per instructions printed at
command start, mode set with "ascii-num=" parameter), then <ENTER> key to
send on MOSI.
- Will send data on MOSI and print received MISO data (same number of bytes as
sent) to terminal (unless "xmit-only=enabled") and prompt for next line of
MOSI data
- As per SPI protocol, will receive MISO data regardless of whether slave
device connected (defaults to zero)
Terminal I/O, slave mode:
- Prints received MOSI data, format as per "ascii-num=" parameter.
- <ENTER> key to pause, then enter MISO data (format as per instructions
printed at command start, mode set with "ascii-num=" parameter), then<ENTER>
to queue MISO data and return to MOSI data printing
- MISO data will be sent one-to-one with subsequent received MOSI bytes
- If MISO data queue empty (e.g. at command start) default "tx-data=" (number
of bytes and byte values) sent on MISO
- MOSI bytes printed in groups, size determined by lesser of number of queued
or default MISO bytes and MOSI bytes received (unless "rx-wait=" time
elapsed)
Terminal I/O, both modes:
- Line editing and history supported, type "help help" for more info.
- Enter "999" if "ascii-num=numeric" or end string ("end=" parameter) if
"ascii-num=ascii", followed by <ENTER> key to exit command.
pty/socket I/O:
- See "help spi ipc term-ptysock=", "help spi pty" and "help spi socket"
pty/socket I/O, master mode:
- Data received from pty and/or socket sent via MOSI. Same number of bytes
received on MISO sent to pty and/or socket(s)
pty/socket I/O, slave mode:
- Data received from pty and/or socket enqueued for sent via MISO, one-to-one
with received MOSI data
- Data received from MOSI sent to pty and/or socket(s)
Command usage:
spi [<parameter>=<value> ...] [<action>] [<configuration>
<parameter>=<value ...] [<action>]
Configurations (primary="spi"):
spi, ipc, pty, socket
Type "help spi <configuration>" for configuration description, parameters, and
actions
Type "help spi <configuration> <parameter>" for parameter description
spi
configuration $1.50: help spi spi
Help for configuration "spi" (e.g. "spi spi"):
Configuration for SPI serial peripheral on ports PA4(NSS), PA5(clock/SCK),
PA6(MISO) and PA7(MOSI)
Slave select:
Slave ("mode=slave"):
- "select=software" slave enabled/active
- "select=hardware" slave controlled by level on NSS port (port PA4)
Master ("mode=master") NSS pin (port PA4):
- "nss=floating" inactive/hi-Z
- "nss=low" set low (any connected slave selected)
- "nss-active" set high at idle, goes low "nss-time=<value>" before
first clock (port PA5) edge and stays low until
goes high "nss-time=<value>" after MOSI data finished
Command/configuration usage:
spi [<parameter>=<value> ...] [<action>]
Actions:
disable Set parameter "active=disabled"
master Set parameter "active=master"
slave Set parameter "active=slave"
Parameters:
mode= SPI mode, or "disabled"
xmit-only= Ignore MISO data in "spi" command, "mode=master"
snoop= Print socket traffic to terminal when "socket_co...
select= See "help spi spi"
baud= One of fixed set of supported baud rates
endian= Bit order for MOSI and MISO data
idle= Clock port PA5 level at idle, "mode=master"
phase= Clock edge at which data bit sampled
miso= Slave ("mode=slave"): MISO port PA6 hardware mod...
pull= Master ("mode=master"): MISO (port PA6). Use "pu...
speed= Output ports (PA7=MOSI, PA5=clock) slew rate
nss= See "help spi spi"
tx-data= NEED UPDATE: "mode=slave" MISO data byte for "mo...
rate= MOSI data output rate. See "Time/Frequency Error...
busy-wait= Wait time for hardware TX not ready. See "Time/F...
rx-wait= Delay for concatenating MISO data into single me...
nss-time= See "help spi spi". See "Time/Frequency Errors" ...
ascii-num= Interactive text input/output mode. See input fo...
end= Character sequence to exit command in "ascii-num...
Configuration "spi" used by commands:
configure
monitor
spi
Type "help spi" for command description
Type "help spi spi <parameter>" for parameter description
Type "help spi spi <action>" for action description
spi
parameters $1.50: help spi mode=
Help for parameter "mode=" (e.g. "spi spi mode="):
- SPI mode, or "disabled"
Current value: disabled
Valid values: "disabled", "master", or "slave"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi xmit-only=
Help for parameter "xmit-only=" (e.g. "spi spi xmit-only="):
- Ignore MISO data in "spi" command, "mode=master"
Current value: disabled
Valid values: "enabled" or "disabled"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi snoop=
Help for parameter "snoop=" (e.g. "spi spi snoop="):
- Print socket traffic to terminal when "socket_config[active]=enabled" and/or
"pty_config[active]=enabled"
Current value: disabled
Valid values: "disabled", "hexadecimal", or "ascii"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi select=
Help for parameter "select=" (e.g. "spi spi select="):
- See "help spi spi"
Current value: software
Valid values: "software" or "hardware"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi baud=
Help for parameter "baud=" (e.g. "spi spi baud="):
- One of fixed set of supported baud rates
Current value: 281.25KHz
Valid values: "36MHz", "18MHz", "9MHz", "4.5MHz", "2.25MHz", "1.125MHz",
"562.5KHz", or "281.25KHz"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi endian=
Help for parameter "endian=" (e.g. "spi spi endian="):
- Bit order for MOSI and MISO data
Current value: msb
Valid values: "msb" or "lsb"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi idle=
Help for parameter "idle=" (e.g. "spi spi idle="):
- Clock port PA5 level at idle, "mode=master"
Current value: low
Valid values: "low" or "high"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi phase=
Help for parameter "phase=" (e.g. "spi spi phase="):
- Clock edge at which data bit sampled
Current value: 1st
Valid values: "1st" or "2nd"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi miso=
Help for parameter "miso=" (e.g. "spi spi miso="):
- Slave ("mode=slave"): MISO port PA6 hardware mode. Set "open-drain" with
external pull-up if multiple hardware slaves, else "push-pull".
Current value: open-drain
Valid values: "push-pull" or "open-drain"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi pull=
Help for parameter "pull=" (e.g. "spi spi pull="):
- Master ("mode=master"): MISO (port PA6). Use "pull=floating" if single
connected slave MISO is push-pull or external pull-up on MISO line, else
"pull=up". Do not use "pull=down".
Current value: floating
Valid values: "floating", "up", or "down"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi speed=
Help for parameter "speed=" (e.g. "spi spi speed="):
- Output ports (PA7=MOSI, PA5=clock) slew rate
Current value: 2MHz
Valid values: "2MHz", "10MHz", or "50MHz"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi nss=
Help for parameter "nss=" (e.g. "spi spi nss="):
- See "help spi spi"
Current value: floating
Valid values: "floating", "low", or "active"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi tx-data=
Help for parameter "tx-data=" (e.g. "spi spi tx-data="):
- NEED UPDATE: "mode=slave" MISO data byte for "monitor" command, and for "spi"
command if interactive input empty/exhausted
Current value: 00
Valid values: 1 to 32 "."-separated values, each 2 hexadecimal, 3 decimal, or 1
ascii character(s)
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi rate=
Help for parameter "rate=" (e.g. "spi spi rate="):
- MOSI data output rate. See "Time/Frequency Errors" section in "help".
Current value: unlimited
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", "ns", "Hz", "kHz", "MHz", or "GHz" suffix in range
[72MHz=13.8889ns ... 16.7638milliHz=59.6523s], or "unlimited"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi busy-wait=
Help for parameter "busy-wait=" (e.g. "spi spi busy-wait="):
- Wait time for hardware TX not ready. See "Time/Frequency Errors" section in
"help".
Current value: 1ms
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", or "ns" suffix in range [13.8889ns ... 59.6523s], or
"infinite"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi rx-wait=
Help for parameter "rx-wait=" (e.g. "spi spi rx-wait="):
- Delay for concatenating MISO data into single message to host, "spi" command.
See "Time/Frequency Errors" section in "help".
Current value: 10ms
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", or "ns" suffix in range [13.8889ns ... 59.6523s], or "none"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi nss-time=
Help for parameter "nss-time=" (e.g. "spi spi nss-time="):
- See "help spi spi". See "Time/Frequency Errors" section in "help".
Current value: zero
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", or "ns" suffix in range [13.8889ns ... 59.6523s], or "zero"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi ascii-num=
Help for parameter "ascii-num=" (e.g. "spi spi ascii-num="):
- Interactive text input/output mode. See input format instructions printed at
command start.
Current value: numeric
Valid values: "ascii" or "numeric"
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
$1.50: help spi end=
Help for parameter "end=" (e.g. "spi spi end="):
- Character sequence to exit command in "ascii-num=ascii" mode
Current value: END
Valid values: <any string <= 16 chars including optional \C character, \xMN
hexadecimal, \ABC octal, and/ or \uABCD unicode escapes) (use
"\x20" or "\040" for space)>
Type "help spi spi" for list of spi configuration parameters
Type "help spi" for command description
i2c
commandI2C semantics are even more complex than SPI. Again, a full tutorial on the subject is outside the scope of this document, and the help
below assumes general knowledge of the I2C protocols.
$1.50: help i2c
Help for command "i2c":
Send and/or receive I2C data
- I2C master or slave mode
- I2C clock (SCL) on port PB10, data (SDA) on port PB11
- Master send to, and optionally receives data from, I2C address
- Slave receive data from master at one or more I2C addresses, and optionally
send TX data to master.
- I2C data to and from terminal (only), input and output in hexadecimal,
decimal, and/or ascii values
Master ("i2c mode=master")
- Enter lines with slave address, size of RX data to receive from slave, and
optional data bytes to TX to slave, finished with <ENTER> key
- Data bytes formatted as space separated entries, each as either 2
hexadecimal digits, 3 decimal digits, or one ascii character
- Multiple "<address> <rx-size> [data ...]" commands separated by ";" allowed
- Prints peripheral status, number of bytes sent and received, and received
byte values (if any) in "ddd-hh-a" format ("ddd" 3 decimal digits, "hh" two
hexadecimal digits, "a" one ascii character or "." if unprintable)
- Example:
"0 0" to end: 8 2 097 61 a
stat= OK tx=3 rx=2 : 077-4d-M 078-4e-N
- Enter "0 0" to exit command
- See help text printed at command start
Slave ("i2c mode=slave")
- Prints data received from master with status, size of data, which address
was matched ("oar1=", "oar2=", "gen-call=") and size of TX data sent to
master (if any requested by master)
- Received data bytes printed in "ddd-hh-a" format ("ddd" 3 decimal digits,
"hh" two hexadecimal digits, "a" one ascii character or "." if unprintable)
- Examples:
stat= OK tx=2 rx=3 @A1 : 097-61-a 097-61-a 097-61-a
stat= OK tx=1 rx=4 @A2 : 100-64-d 101-65-e 102-66-f 103-67-g
stat= OK tx=0 rx=1 @GC : 120-78-x
- Note: TX slave-to-master not allowed for "gen-call" address 0
- Press <ENTER> key to set maximum number of data bytes to receive from master
(any additional will be discarded), and optional data bytes to queue to be
sent to master (format: space-separated triple decimal or double hexadecimal
digits, or single ascii characters), finished with <ENTER> key
- If no slave-to-master data enqueued, will cyclically send "i2c tx-data="
bytes for any number of slave TX bytes requested by master
- Enter "0" to exit command
- See help texts printed at command start and at <ENTER> key
Both modes:
- Line editing and history supported, type "help help" for more info.
- Note I2C multi-master, 10-bit addresses, SMBus mode, etc. not supported.
Command usage:
i2c [<parameter>=<value> ...] [<action>]
Configuration: i2c
Type "help i2c i2c" for configuration description, parameters, and actions
Type "help i2c <parameter>" for parameter description
Type "help i2c <action>" for action description
i2c
configuration $1.50: help i2c i2c
Help for configuration "i2c" (e.g. "i2c i2c"):
Configuration for I2C serial peripheral on ports PB10(clock/SCL) and
PB11(data/SDA)
Command/configuration usage:
i2c [<parameter>=<value> ...] [<action>]
Actions:
disable Set parameter "active=disabled"
master Set parameter "active=master"
slave Set parameter "active=slave"
Parameters:
mode= I2C mode, or "disabled"
flavor= I2C protocol mode
fast-duty= Duty cycle for I2C clock in "flavor=fast" mode
freq= I2C clock frequency (only used for "mode=master"...
addr= Master TX address for "monitor" command
gen-call= Slave accepts "general call" address (0) in addi...
oar1= Slave address
oar2= Optional second slave address, or "disabled"
tx-data= Data for master TX in "monitor" command, or slav...
rx-size= Maximum size of slave RX data (any additonal dis...
timeout= Timeout value to abort firmware-to-hardware comm...
gpio= I/O port output slew rate, PB11(SDA) (master or ...
Configuration "i2c" used by commands:
configure
monitor
i2c
Type "help i2c" for command description
Type "help i2c <parameter>" for parameter description
Type "help i2c <action>" for action description
i2c
parameters $1.50: help i2c mode=
Help for parameter "mode=" (e.g. "i2c i2c mode="):
- I2C mode, or "disabled"
Current value: disabled
Valid values: "disabled", "master", or "slave"
Type "help i2c i2c" for list of i2c configuration parameters
Type "help i2c" for command description
$1.50: help i2c flavor=
Help for parameter "flavor=" (e.g. "i2c i2c flavor="):
- I2C protocol mode
Current value: standard
Valid values: "standard" or "fast"
Type "help i2c i2c" for list of i2c configuration parameters
Type "help i2c" for command description
$1.50: help i2c fast-duty=
Help for parameter "fast-duty=" (e.g. "i2c i2c fast-duty="):
- Duty cycle for I2C clock in "flavor=fast" mode
Current value: 2:1
Valid values: "2:1" or "16:9"
Type "help i2c i2c" for list of i2c configuration parameters
Type "help i2c" for command description
$1.50: help i2c freq=
Help for parameter "freq=" (e.g. "i2c i2c freq="):
- I2C clock frequency (only used for "mode=master". See "Time/Frequency Errors"
section in "help".
Current value: 100kHz=10μs
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", "ns", "Hz", "kHz", "MHz", or "GHz" suffix in range
[1MHz=1μs ... 5kHz=200μs]
Type "help i2c i2c" for list of i2c configuration parameters
Type "help i2c" for command description
$1.50: help i2c addr=
Help for parameter "addr=" (e.g. "i2c i2c addr="):
- Master TX address for "monitor" command
Current value: 0
Valid values: decimal or hex integer in range [0 ... 127]
Type "help i2c i2c" for list of i2c configuration parameters
Type "help i2c" for command description
$1.50: help i2c gen-call=
Help for parameter "gen-call=" (e.g. "i2c i2c gen-call="):
- Slave accepts "general call" address (0) in addition to "oar1=" and "oar2="
Current value: disabled
Valid values: "enabled" or "disabled"
Type "help i2c i2c" for list of i2c configuration parameters
Type "help i2c" for command description
$1.50: help i2c oar1=
Help for parameter "oar1=" (e.g. "i2c i2c oar1="):
- Slave address
Current value: 8
Valid values: decimal or hex integer in range [8 ... 119]
Type "help i2c i2c" for list of i2c configuration parameters
Type "help i2c" for command description
$1.50: help i2c oar2=
Help for parameter "oar2=" (e.g. "i2c i2c oar2="):
- Optional second slave address, or "disabled"
Current value: disabled
Valid values: "disabled" or integer in range [8 ... 119]
Type "help i2c i2c" for list of i2c configuration parameters
Type "help i2c" for command description
$1.50: help i2c tx-data=
Help for parameter "tx-data=" (e.g. "i2c i2c tx-data="):
- Data for master TX in "monitor" command, or slave TX in "i2c" command if no
data enqueued interactively in "i2c" command
Current value: 00
Valid values: 1 to 16 "."-separated values, each 2 hexadecimal, 3 decimal, or 1
ascii character(s)
Type "help i2c i2c" for list of i2c configuration parameters
Type "help i2c" for command description
$1.50: help i2c rx-size=
Help for parameter "rx-size=" (e.g. "i2c i2c rx-size="):
- Maximum size of slave RX data (any additonal discarded) in "monitor" command
Current value: 0
Valid values: decimal or hex integer in range [0 ... 60]
Type "help i2c i2c" for list of i2c configuration parameters
Type "help i2c" for command description
$1.50: help i2c timeout=
Help for parameter "timeout=" (e.g. "i2c i2c timeout="):
- Timeout value to abort firmware-to-hardware commands if no response. See
"Time/Frequency Errors" section in "help".
Current value: 100ms
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", or "ns" suffix in range [13.8889ns ... 59.6523s], or
"infinite"
Type "help i2c i2c" for list of i2c configuration parameters
Type "help i2c" for command description
$1.50: help i2c gpio=
Help for parameter "gpio=" (e.g. "i2c i2c gpio="):
- I/O port output slew rate, PB11(SDA) (master or slave) and PB10(SCL) for
master
Current value: 2MHz
Valid values: "2MHz", "10MHz", or "50MHz"
Type "help i2c i2c" for list of i2c configuration parameters
Type "help i2c" for command description
pulse
commandThe poorly-name pulse
command (see above, should be “pwm” or “waveform”) outputs a one, two, or three phase digital pulse train on ports PA1 through PA3.
Note (as per help pulse
, below) that as opposed to all other buck50.py
commands which run until completion/exited/interrupted, the pulse waveform outputs continue indefinitely until pulse disable
or a different command which uses the same ports is executed. Using this capability a single buck50 “Blue Pill” can drive external systems while simultaneously observing them (monitor
, logic
, oscope
, usart
, spi
, etc commands), or, alternately, do self-test experiments by connecting the pulse
output ports to the various commands' input ports.
$1.50: help pulse
Help for command "pulse":
Output digital pulse waveforms on ports PA1, PA2, and/or PA3.
- Waveform output continues, even after program exit, until explicitly halted
with "pulse enable=disabled", or "gpio", "numbers", or "usart ports=pa0-3"
commands (which use same ports/pins).
- "time1", "time2", and "time3" specify event times less than or equal to master
"time" period
- At event times "time1"/"time2"/"time3" ports PA1/PA2/PA3 respectively change
level in various ways controlled by "mode1", "mode2", and "mode3"
- At end of master "time" period ports PA1/PA2/PA3 return to default states
Command usage:
pulse [<parameter>=<value> ...] [<action>]
Configuration: pulse
Type "help pulse pulse" for configuration description, parameters, and actions
Type "help pulse <parameter>" for parameter description
Type "help pulse <action>" for action description
pulse
configuration $1.50: help pulse pulse
Help for configuration "pulse" (e.g. "pulse pulse"):
Parameters "mode1", "mode2", "mode3" control output levels on PA1/PA2/PA3 ports
at "time1", "time2", "time3", all within master "time" period.
Command/configuration usage:
pulse [<parameter>=<value> ...] [<action>]
Actions:
enable Set parameter "active=enabled"
disable Set parameter "active=disabled"
Parameters:
active= If "enabled" runs (even after program exit) unt...
time= Master period/frequency for time1, time2, and ti...
time1= Time for "mode1" event type. If "%" suffix is ...
time2= See "help pulse time1"
time3= See "help pulse time1"
mode1= ort PA1 change at "time1":...
mode2= See "help pulse mode1"
mode3= See "help pulse mode1"
gpio1= Disable ("hi-z") or enable "push-pull" or "open-...
gpio2= See "help pulse gpio1"
gpio3= See "help pulse gpio1"
speed1= Port PA1 slew rate
speed2= See "help pulse speed1"
speed3= See "help pulse speed1"
Configuration "pulse" used by commands:
configure
pulse
Type "help pulse" for command description
Type "help pulse <parameter>" for parameter description
Type "help pulse <action>" for action description
pulse
parameters $1.50: help pulse active=
Help for parameter "active=" (e.g. "pulse pulse active="):
- If "enabled" runs (even after program exit) until halted with "disabled" or
by "gpio", "numbers" or "usart ports=pa0-3" command.
Current value: disabled
Valid values: "enabled" or "disabled"
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse time=
Help for parameter "time=" (e.g. "pulse pulse time="):
- Master period/frequency for time1, time2, and time3. See "Time/Frequency
Errors" section in "help".
Current value: 1kHz=1ms
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", "ns", "Hz", "kHz", "MHz", or "GHz" suffix in range
[36MHz=27.7778ns ... 16.7638milliHz=59.6523s]
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse time1=
Help for parameter "time1=" (e.g. "pulse pulse time1="):
- Time for "mode1" event type. If "%" suffix is percent of master "time"
parameter. If explicit time/freq must be less than or equal to "time" (greater
than or equal to frequency), and will be clamped if "time" changes. See
"Time/Frequency Errors" section in "help".
Current value: 25%=250.028μs(error:-6.94444ns)
Valid values: floating point number with "%", "s", "ms", "us", "μs", "ns",
"Hz", "kHz", "MHz", or "GHz" suffix in range [0ps ... 18.2042hr]
or <xxx>% (xxx floating point number in range [0 ... 100])
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse time2=
Help for parameter "time2=" (e.g. "pulse pulse time2="):
- See "help pulse time1"
Current value: 50%=500.028μs(error:-13.8889ns)
Valid values: floating point number with "%", "s", "ms", "us", "μs", "ns",
"Hz", "kHz", "MHz", or "GHz" suffix in range [0ps ... 18.2042hr]
or <xxx>% (xxx floating point number in range [0 ... 100])
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse time3=
Help for parameter "time3=" (e.g. "pulse pulse time3="):
- See "help pulse time1"
Current value: 75%=750μs(error:+6.94444ns)
Valid values: floating point number with "%", "s", "ms", "us", "μs", "ns",
"Hz", "kHz", "MHz", or "GHz" suffix in range [0ps ... 18.2042hr]
or <xxx>% (xxx floating point number in range [0 ... 100])
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse mode1=
Help for parameter "mode1=" (e.g. "pulse pulse mode1="):
- Port PA1 change at "time1":
frozen : no change
up : set high
down : set low
toggle : set high if low, or set low if high (no change at master "time")
high : continuously high
low : continuously low
fall : high-to-low (return high at master "time")
rise : low-to-high (return low at master "time")
Current value: rise
Valid values: "frozen", "up", "down", "toggle", "low", "high", "fall", or
"rise"
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse mode2=
Help for parameter "mode2=" (e.g. "pulse pulse mode2="):
- See "help pulse mode1"
Current value: rise
Valid values: "frozen", "up", "down", "toggle", "low", "high", "fall", or
"rise"
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse mode3=
Help for parameter "mode3=" (e.g. "pulse pulse mode3="):
- See "help pulse mode1"
Current value: rise
Valid values: "frozen", "up", "down", "toggle", "low", "high", "fall", or
"rise"
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse gpio1=
Help for parameter "gpio1=" (e.g. "pulse pulse gpio1="):
- Disable ("hi-z") or enable "push-pull" or "open-drain") on port PA1
Current value: push-pull
Valid values: "push-pull", "open-drain", or "hi-z"
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse gpio2=
Help for parameter "gpio2=" (e.g. "pulse pulse gpio2="):
- See "help pulse gpio1"
Current value: push-pull
Valid values: "push-pull", "open-drain", or "hi-z"
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse gpio3=
Help for parameter "gpio3=" (e.g. "pulse pulse gpio3="):
- See "help pulse gpio1"
Current value: push-pull
Valid values: "push-pull", "open-drain", or "hi-z"
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse speed1=
Help for parameter "speed1=" (e.g. "pulse pulse speed1="):
- Port PA1 slew rate
Current value: 2MHz
Valid values: "2MHz", "10MHz", or "50MHz"
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse speed2=
Help for parameter "speed2=" (e.g. "pulse pulse speed2="):
- See "help pulse speed1"
Current value: 2MHz
Valid values: "2MHz", "10MHz", or "50MHz"
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
$1.50: help pulse speed3=
Help for parameter "speed3=" (e.g. "pulse pulse speed3="):
- See "help pulse speed1"
Current value: 2MHz
Valid values: "2MHz", "10MHz", or "50MHz"
Type "help pulse pulse" for list of pulse configuration parameters
Type "help pulse" for command description
ipc
configurationThe ipc
configuration controls whether serial protocol (usart
, spi
, i2c
) I/O is directed to the buck50.py
terminal or to external UNIX devices.
$1.50: help configure ipc
Help for configuration "ipc" (e.g. "configure ipc"):
Enable/disable IPC (pty and/or socket) I/O
- Also see "help <command> pty" and "help <command> socket"
Command/configuration usage:
configure ipc [<parameter>=<value> ...] [<action>]
Actions:
terminal Set parameter "i/o=terminal"
external Set parameter "i/o=external"
Parameters:
i/o= I/O to terminal or to pty and/or socket (set "pt...
flush= Delay time to flush pending received device-to-h...
Configuration "ipc" used by commands:
configure
gpio
usart
spi
Type "help configure" for command description
Type "help configure ipc <parameter>" for parameter description
Type "help configure ipc <action>" for action description
ipc
parameters and actions $1.50: help configure ipc terminal
Help for action "terminal" (e.g. "configure ipc terminal"):
- Set parameter "i/o=terminal"
Type "help configure ipc" for list of ipc configuration actions
Type "help configure" for command description
$1.50: help configure ipc terminal external
Help for action "external" (e.g. "configure ipc external"):
- Set parameter "i/o=external"
Type "help configure ipc" for list of ipc configuration actions
Type "help configure" for command description
$1.50: help configure ipc terminal i/o=
Help for parameter "i/o=" (e.g. "configure ipc i/o="):
- I/O to terminal or to pty and/or socket (set "pty enable" and/or "socket
enable")
Current value: terminal
Valid values: "terminal" or "external"
Type "help configure ipc" for list of ipc configuration parameters
Type "help configure" for command description
$1.50: help configure ipc terminal flush=
Help for parameter "flush=" (e.g. "configure ipc flush="):
- Delay time to flush pending received device-to-host data at command exit
Current value: 1s
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", or "ns" suffix in range [1s ... 1min]
Type "help configure ipc" for list of ipc configuration parameters
Type "help configure" for command description
pty
configuration $1.50: help configure pty
Help for configuration "pty" (e.g. "configure pty"):
Connect USB I/O to /dev/pty/... special device file.
- Connect external program to /dev/pts/N ("N" printed to terminal at startup)
- Example: `miniterm /dev/pts/32`
- Pty stty/termios values not passed through to usart -- use usart config
parameters (see "help usart usart") to set baud, parity, etc.
- Must also set "ipc external"
- Set "snoop=enabled" parameter to view xmit/recv data
Command/configuration usage:
configure pty [<parameter>=<value> ...] [<action>]
Actions:
enable Set parameter "active=enabled"
disable Set parameter "active=disabled"
Parameters:
active= Create, and I/O to, /dev/pts/N
snoop= Print socket traffic to terminal
Configuration "pty" used by commands:
configure
usart
spi
Type "help configure" for command description
Type "help configure pty <parameter>" for parameter description
Type "help configure pty <action>" for action description
pty
parameters and actions $1.50: help configure pty enable
Help for action "enable" (e.g. "configure pty enable"):
- Set parameter "active=enabled"
Type "help configure pty" for list of pty configuration actions
Type "help configure" for command description
$1.50: help configure pty disable
Help for action "disable" (e.g. "configure pty disable"):
- Set parameter "active=disabled"
Type "help configure pty" for list of pty configuration actions
Type "help configure" for command description
$1.50: help configure pty active=
Help for parameter "active=" (e.g. "configure pty active="):
- Create, and I/O to, /dev/pts/N
Current value: disabled
Valid values: "enabled" or "disabled"
Type "help configure pty" for list of pty configuration parameters
Type "help configure" for command description
$1.50: help configure pty snoop=
Help for parameter "snoop=" (e.g. "configure pty snoop="):
- Print socket traffic to terminal
Current value: disabled
Valid values: "disabled", "hexadecimal", or "ascii"
Type "help configure pty" for list of pty configuration parameters
Type "help configure" for command description
socket
configuration $1.50: help configure socket
Help for configuration "socket" (e.g. "configure socket"):
Connect USB I/O to Unix socket.
- Socket is server, supports connections from multiple external clients
- Set "host" and "port" parameters
- Connect via external client program, e.g. `telnet localhost 1024`
- Must also set "ipc external"
- Set "snoop" parameter to view xmit/recv data
Command/configuration usage:
configure socket [<parameter>=<value> ...] [<action>]
Actions:
enable Set parameter "active=enabled"
disable Set parameter "active=disabled"
Parameters:
active= Enable/disable I/O to socket
host= Server socket hostname
port= Server socket port number
snoop= Print socket traffic to terminal
Configuration "socket" used by commands:
configure
usart
spi
Type "help configure" for command description
Type "help configure socket <parameter>" for parameter description
Type "help configure socket <action>" for action description
socket
parameters and actions $1.50: help configure socket enable
Help for action "enable" (e.g. "configure socket enable"):
- Set parameter "active=enabled"
Type "help configure socket" for list of socket configuration actions
Type "help configure" for command description
$1.50: help configure socket disable
Help for action "disable" (e.g. "configure socket disable"):
- Set parameter "active=disabled"
Type "help configure socket" for list of socket configuration actions
Type "help configure" for command description
$1.50: help configure socket active=
Help for parameter "active=" (e.g. "configure socket active="):
- Enable/disable I/O to socket
Current value: disabled
Valid values: "enabled" or "disabled"
Type "help configure socket" for list of socket configuration parameters
Type "help configure" for command description
$1.50: help configure socket host=
Help for parameter "host=" (e.g. "configure socket host="):
- Server socket hostname
Current value: localhost
Valid values: <any string <= 63 chars including optional \C character, \xMN
hexadecimal, \ABC octal, and/ or \uABCD unicode escapes) (use
"\x20" or "\040" for space)>
Type "help configure socket" for list of socket configuration parameters
Type "help configure" for command description
$1.50: help configure socket port=
Help for parameter "port=" (e.g. "configure socket port="):
- Server socket port number
Current value: 1024
Valid values: decimal or hex integer in range [1024 ... 65535]
Type "help configure socket" for list of socket configuration parameters
Type "help configure" for command description
$1.50: help configure socket snoop=
Help for parameter "snoop=" (e.g. "configure socket snoop="):
- Print socket traffic to terminal
Current value: disabled
Valid values: "disabled", "hexadecimal", or "ascii"
Type "help configure socket" for list of socket configuration parameters
Type "help configure" for command description
reset
commandThe reset
command implements a set of varied secondary actions. With the exception of the reset ext-trig
action and its associated reset ganged=
parameter, none of them should be necessary during normal buck50 operation. ext-trig
is important for “ganging” multiple buck50s together – see help reset ext-trig
and help reset ganged=
, below.
Most of the other actions are designed for recovery in the unlikely 21 event of a firmware fault. The somewhat redundant ability to disconnect a running buck50.py
process from one buck50 “Blue Pill” and reconnect to another is also supported. The reset serialnum
and reset blink
actions may be useful in identifying multiple simultaneously-connected buck50 “Blue Pill” boards – see buck50.py startup and device connection, above.
Changing the default ST-recommended reset wait=
and pre=
parameters with the reset flash
action is not advisable, but included is case the user wishes to live dangerously.
See above for why there is no reset defaults
command.
$1.50: help reset
Help for command "reset":
Reset various host, device, and USB subsystems:
- Enabling or disabling multiple connected devices external triggering and
sampling synchronization ("ganged=", "ext-trig")
- USB disconnect, (re-)connect, or port/driver change ("connect", "usb=")
- Stalled USB communications recovery ("flush")
- Firmware halt and/or reset ("halt", "init")
- Change flash memory (firmware code) wait states and/or enable/disable
prefetch queue ("flash" with "wait=" and/or "pre=")
Performing "halt", "flush", and/or "init" required only in case of
firmware/software bug or USB communication failure.
Similarly, "connect" with or without new "usb=/dev/<...>" for error or to
connect to different port/device.
Command usage:
reset [<parameter>=<value> ...] [<action>]
Configuration: reset
Type "help reset reset" for configuration description, parameters, and actions
Type "help reset <parameter>" for parameter description
Type "help reset <action>" for action description
reset
configuration $1.50: help reset reset
Help for configuration "reset" (e.g. "reset reset"):
- Set any/all of "halt=", "flush=", "reconnect=", and/or "init=" to "enabled"
- Will execute in above order
- All reset to "disabled" after command completion.
Command/configuration usage:
reset [<parameter>=<value> ...] [<action>]
Actions:
ext-trig Activates or de-activates external trigger sync ...
flash Set flash memory wait states and prefetch buffer...
halt Send halt command to firmware
flush Flush USB data pipe (/dev/tty/ACM0)
init Send initialization commands to firmware
connect Disconnect/re-connect to same or different USB p...
serialnum Get device serial number
blink Blink user LED to identify device
Parameters:
ganged= Synchronized digital ("logic") and/or analog ("o...
wait= Flash memory wait states. Stability unlikely if ...
pre= Flash memory prefetch buffer. Performance impact...
usb= USB port and device driver
timeout= Timeout value for reset USB operations (can inte...
Configuration "reset" used by commands:
configure
reset
Type "help reset" for command description
Type "help reset <parameter>" for parameter description
Type "help reset <action>" for action description
reset
actions and parameters $1.50: help reset ext-trig
Help for action "ext-trig" (e.g. "reset ext-trig"):
- Activates or de-activates external trigger sync after setting "ganged="
parameter ("enabled" or "disabled"). See "help reset ganged=". Also resets all
PAx and PBx ports to default.
Type "help reset reset" for list of reset configuration actions
Type "help reset" for command description
$1.50: help reset flash
Help for action "flash" (e.g. "reset flash"):
- Set flash memory wait states and prefetch buffering. Must be after "wait=" and
"pre=" parameters.
Type "help reset reset" for list of reset configuration actions
Type "help reset" for command description
$1.50: help reset halt
Help for action "halt" (e.g. "reset halt"):
- Send halt command to firmware
Type "help reset reset" for list of reset configuration actions
Type "help reset" for command description
$1.50: help reset flush
Help for action "flush" (e.g. "reset flush"):
- Flush USB data pipe (/dev/tty/ACM0)
Type "help reset reset" for list of reset configuration actions
Type "help reset" for command description
$1.50: help reset init
Help for action "init" (e.g. "reset init"):
- Send initialization commands to firmware
Type "help reset reset" for list of reset configuration actions
Type "help reset" for command description
$1.50: help reset connect
Help for action "connect" (e.g. "reset connect"):
- Disconnect/re-connect to same or different USB port / device driver (first set
"usb" parameter if changing)
Type "help reset reset" for list of reset configuration actions
Type "help reset" for command description
$1.50: help reset serialnum
Help for action "serialnum" (e.g. "reset serialnum"):
- Get device serial number
Type "help reset reset" for list of reset configuration actions
Type "help reset" for command description
$1.50: help reset blink
Help for action "blink" (e.g. "reset blink"):
- Blink user LED to identify device
Type "help reset reset" for list of reset configuration actions
Type "help reset" for command description
$1.50: help reset ganged=
Help for parameter "ganged=" (e.g. "reset reset ganged="):
- Synchronized digital ("logic") and/or analog ("oscope") command triggering
across multiple devices. Requires two open-drain bus lines, each with pull-up
resistor to Vdd. Connect all devices' PB14 ports to one open-drain bus, and
all PB15 ports to other. After enabling with "ganged=enabled" must be
activated via "reset ext-trig" action before "logic" or "oscope" commands.
Triggering on any device will trigger all.
Current value: disabled
Valid values: "enabled" or "disabled"
Type "help reset reset" for list of reset configuration parameters
Type "help reset" for command description
$1.50: help reset wait=
Help for parameter "wait=" (e.g. "reset reset wait="):
- Flash memory wait states. Stability unlikely if less than default value of 2.
Use "flash" action to set.
Current value: 2
Valid values: decimal or hex integer in range [0 ... 2]
Type "help reset reset" for list of reset configuration parameters
Type "help reset" for command description
$1.50: help reset pre=
Help for parameter "pre=" (e.g. "reset reset pre="):
- Flash memory prefetch buffer. Performance impacted if not enabled. Use "flash"
action to set
Current value: enabled
Valid values: "enabled" or "disabled"
Type "help reset reset" for list of reset configuration parameters
Type "help reset" for command description
$1.50: help reset usb=
Help for parameter "usb=" (e.g. "reset reset usb="):
- USB port and device driver
Current value: /dev/tty/ACM0
Valid values: <any valid file or path name> (no overwrite warning if filename
contains "tmp", "temp", or "scratch")
Type "help reset reset" for list of reset configuration parameters
Type "help reset" for command description
$1.50: help reset timeout=
Help for parameter "timeout=" (e.g. "reset reset timeout="):
- Timeout value for reset USB operations (can interrupt by <ENTER> key
Current value: 1s
Valid values: floating point number with "y", "d", "our", "min", "s", "ms",
"us", "μs", or "ns" suffix in range [0ps ... 1.82795e+10y], or
"infinite"
Type "help reset reset" for list of reset configuration parameters
Type "help reset" for command description
There are several open-source digital and analog waveform viewing programs that can be used to display digital and analog sample data captured by buck50.py
. Each has its pros and cons, but unfortunately none is ideal. In somewhat reverse order of usefulness …
As per above, if gnuplot
(gnuplot website) is installed on the buck50.py
host system, it is used as the default viewer for both digital and analog captures. The captures are automatically dumped (see dump auto-digital
and dump auto-analog
above) to a temporary CSV file and gnuplot
invoked with the requisite 9 commands to view it.
The intent is to provide a quick “sanity check” overview of the capture. gnuplot
’s interactive measurement and analysis tools are limited 22, and it certainly has no protocol decoding capabilities. But both pulseview
and gtkwave
are slower to start up, and function poorly when repeatedly invoked with commandline-specified data files. It is much more practical to start them independently, use their menus to load buck50.py
VCD output, and “reload” when those output files are changed. See dump digital-frmt=
and dump viewer-vcd=
, above/above.
gtkwave
looks to be a highly capable program with many features, but somewhat aimed at different use-cases (VLSI simulation?) than what buck50.py
requires. It’s also a classic example of non-intuitive graphical interface with non-useful default behavior. When a file (maybe just VCD files?) is loaded, nothing is displayed. The above image is the result of much post-load setup: Click “top”, then when the initially undisplayed list of captured signals appears select them all, and then “Append” 23. And if color-coded traces are desired, each signal name must be individually right-clicked, then “Color Format” in the pulldown menu, and a color chosen. Simple and obvious, right? Did the designers ever consider displaying everything by default with the ability to delete unwanted elements?
It also has no signal decoding capabilities 24, again likely because that falls outside the program’s intended use-cases.
The sigrok website “PulseView” software is among the best programs, open-source or commercial, I have ever used. I prefer its design, interface, usability, extensibility, and range of available binary and protocol decoders to even Saleae’s “Logic” software, below.
Unfortunately, pulseview
has one well-known, almost fatal flaw (along with several minor ones, see 25), particularly for use with “buck50” captures: It allocates memory on a per-tick/per-signal basis, regardless of whether there is any change in any signal at that tick’s time.
I believe this is the basis for the oft-cited complaint against pulseview
: That it’s slow. I don’t believe this is caused by any failing in the code’s algorithms or their implementations, but simply due to memory swapping. For example, do NOT load the following example VCD file into pulseview
(at least with the default VCD import settings):
$timescale 1 ns $end
$scope module top $end
$var wire 1 z blk $end
$upscope $end
$enddefinitions $end
#0
0z
#9999999999
1z
#10000000000
0z
That’s a 1ns pulse at the end of 10 seconds of inactivity. On my system it consumes approximately 10 gigabytes of memory. On systems with less than that amount of RAM (and with disk swap enabled) I’ve witnessed real, non-example captured data with similar periods of inactivity lock up the host in infinite swap activity requiring a hardware reboot to recover 26.
Yes, there’s the “compress idle periods” option of the VCD importer. (Is there any way to set that from the pulseview
commandline when loading a file at startup?) It’s a useful hack to avoid the “fatality” of the problem, but it’s a hack nonetheless – it requires a priori knowledge of the maximum length of actual important null periods, and it destroys real information about the interval between events (compress idle periods above 1ms, and is it 1ms or 10 minutes between two UART characters?)
The saddest part is that a casual examination of the source code for pulseview
and its libsigrok
library show them to be well designed and structured pieces of software engineering. In particular, pulseview
is written in C++, and it seems that, for example, class Segment
could be derived from, with an overloaded get_raw_sample()
method that handled sparse, memory-efficient, non-consecutive samples.
I might get desperate and try to implement this in pulseview
myself. But there’s an obvious learning curve to be breached, and I’ve already devoted far too much time developing “buck50” itself, so I’m hoping that a sigrok developer will someday address the issue (which has been been publicly requested by others). FWIW, the offending code in libsigrok/src/input/vcd.c
is:
/* Generate samples from prev_timestamp up to timestamp - 1. */
count = timestamp - inc->prev_timestamp;
sr_spew("Got a new timestamp, feeding %zu samples", count);
add_samples(in, count, FALSE);
inc->prev_timestamp = timestamp;
inc->data_after_timestamp = FALSE;
continue;
buck50.py
VCD/pulseview parameters The “buck50” firmware timestamps its sparse samples with 1/72MHz precision 27. Ideally, its VCD output files would specify their $timescale
header with this value, but not only does the VCD format not support frequency-based nor floating point period timescales, but loading such a file into pulseview
would almost certainly cause excessive memory consumption, possibly even with “compress idle periods” set. This is the purpose of buck50.py
’s dump tick-units=
and dump per-tick=
parameters. See the buck50.py
help system, above, regarding the topic.
The Logic
software by Saleae (company website, software download) is not open-source software but is distributed for use free of charge (license). Much appreciation to Saleae for doing so.
The following comments are strictly my own (possible mis-)understandings and opinions:
In at least their original products, Saleae leveraged the ability of the Cypress Semiconductor FX2 USB microcontroller to transfer data acquired on GPIO port inputs to USB at very high speed (full USB2.0 bandwidth?). This is in contrast to traditional logic analyzers (also “buck50”) which buffer data in memory and transfer it to a host computer (or even more traditionally use a built-in display) for viewing. The advantage of Saleae approach is that it provides effectively infinitely long captures as long as the analyzer-to-host communication bandwidth and host processing/storage speeds are not exceeded.
Consequently, the basic use of the Logic
software is to drive the Saleae hardware and capture its output in real time. The software can save and reload captured data, but only in Saleae’s own proprietary, undocumented format. The realtime communication protocol is publicly known (unsure if this is with Saleae’s permission or not) and in fact PulseView can be used with a realtime-attached Saleae analyzer.
The bottom line to all this is that Logic
is not usable for viewing “buck50”-captured data. I would prefer to spend any development effort improving PulseView’s failings as described above instead of simulating the Saleae protocol for non-realtime viewing. In addition, doing the latter would likely fall into the same “sparse data” hole that PulseView has, requiring the generation of the non-existent static/non-changing samples only to have them elided by Logic
(which does handle long idle periods without problems).
FWIW, note that the data in the above image was captured using an older, discontinued, but genuine Saleae product which was on temporary loan to me by its owner. Using it (my first hands-on experience with a logic analyzer although I was certainly aware of them and their capabilities) along with lack of funds to purchase one for permanent use was in some ways the impetus for creating “buck50”. Note also that the Saleae products (and similarly other commercially-available logic analyzers) “kick ass” on buck50’s capabilities 28.
Trying to optimally map peripherals to I/O pins on ST microcontrollers is very much like solving a multi-dimensional jigsaw puzzle. This is especially true with the primitive “alternate function” mapping on the STM32F1 series MCUs (which ST improved slightly in their subsequent chips by inverting the many-to-one, function-to-pin API into a pin-to-function scheme). I’ve complained vociferously about this in my other GitHub repositories, citing the huge advantages in capabilities and ease of use of NXP’s “switch matrix” crossbar approach (but note that NXP has their own brain damage, such as the lack of software SPI slave select which forces wasting two pins to achieve the same result).
An early version of buck50 used ports PA0 through PA7 for its digital sampling and monitoring lines. That was much cleaner from a conceptual (ports 0–7 mapped to bits 0–7) and hardware (PA0 through PA7 are contiguous on the “Blue Pill” board) standpoint, but had the fatal flaw that reading GPIOA also brought in the USB-
and USB+
bits from pins PA11 and PA12 respectively. As those are changing at random with USB activity (including activity not directed at the buck50 “Blue Pill” in use) it required masking the bits out with an extra instruction before the crnt==prev
test in the hyper-time-critical sampling loop (see Hacks^H^H^H^H^Htechniques, below). 29
Additionally, as buck50 grew from its beginnings as a pure logic analyzer into the multi-function mess^H^H^H^Hsuper tool it is now, switching over to the PB ports allowed better utilization of the additional functions (SPI, USART, etc.) Note that PB2 is not brought out as a usable pin on the “Blue Pill” boards and it was again going to be inefficient if the port bits were not contiguous. That left a choice between PB3-PB10 vs PB4-PB11, and it was felt that as long as the pins were not going to be contiguous on the boards it was better to split them 6-and-2 instead of 7-and-1 with the lone PB10 on the opposite side of the board from the others.
That’s the story, anyway. If I was laying out the board I’d do it differently (and use a more modern/powerful chip), but then it wouldn’t cost the US$1.50 which is a large part of the 30 charm of this project.
As mentioned, the buck50 firmware uses several somewhat unusual programming idioms to achieve its time-critical sampling performance given the limitations of the STM32F103 platform. Among them are …
The first hack 31 is the innermost sampling code loop, which in its simplest form consists of:
uniform:
ldr r3, [r0, #8]
ldr r5, [r1, #8]
orr.w r5, r5, r3, lsl #20
cmp r3, r4
it ne
strne.w r5, [r2], #4
mov r4, r3
b.n uniform
Note that this is written in ARM Thumb assembly, which isn’t strictly necessary for this part of the code but avoids fighting GCC’s limited control of register allocation when using embedded assembly from C/C++. That control is, in turn, required for other parts of the overall architecture, to be described shortly.
The fine details of the code, which is essentially a translation of the C-like pseudocode:
while (true) {
crnt = gpiob->idr; // read GPIO port B
time = systick->val; // read current time, lower 24 bits
smpl = time | (crnt << 20); // GPIO bits 4 thru 11 or'd with time
if (crnt == prev)
*samples++ = crnt;
prev = crnt;
}
are not as important as its obvious faults:
All of these are intentional in order to make the loop as fast as possible. Incrementing and checking a counter (or comparing samples
to an end address), checking a timer’s count or a timeout bit, and/or checking the USB communication layer’s state or message interrupt flag would all add extra cycles.
(Note that “buck50” includes optional variants of the assembly code (see logic mode=
, above) that eliminate the prev=current
copy by “ping-ponging” between two registers, and/or partially unroll the loop to amortize the time-costly branch instruction and pipleline stall it causes, but the principles remain the same.)
So how, you the hypothetical curious reader ask, does this ever work? The scheme was part of the original “buck50” conceptual design from its very beginning. As per the buck50.py
help system above:
$1.50: help logic
...
- sampling continues until first of:
. user interrupt (<ENTER> key)
. duration elapsed ("logic duration=")
. number of samples ("logic edges=") (incl. extra samples @ 233ms)
. memory full
This is implemented via the following hacks:
.ld
configuration. From low to high it’s .data
(initialized data), then .bss
(uninitialized data), then the .stack
, and finally sample memory from there until the end of RAM. (Usually the stack is at the top of RAM.)end_of_ram - number_of_samples
.logic duration=
, above), a timer is started with that time period which then generates an interrupt, also trapped, when it expires.longjmp
from interrupt handler The memory fault and timer interrupt handlers always, and the USB handler if sampling in progress, set a few global variables and then do a C-style longjmp
to the top of main()
’s main loop. The variables are read and handled, an appropriate USB message is sent to the buck50.py
host program, and the main loop continues waiting for another host command. Standard setjmp
/longjmp
methodology.
This again was part of the basic “buck50” design from the beginning. Alternately, in addition to the other time-consuming code designs described above, the interrupt handlers could set a flag (either a C/C++ variable or, for faster testing, directly in a register) and the sampling loop could check that. Either way (more so with the variable) would add precious cycles and slow down the sampling rate.
But there was a fly in the ointment. A fair amount of work went into initial “buck50” development, with unthinking confidence in the ancient C longjmp mechanism before it was discovered that the longjmp
distributed in the GNU Arm Embedded Toolchain’s C libraries doesn’t work.
Or to be more accurate 32, it doesn’t work if called from an ARM interrupt handler. Or more accurately still, it does work when called from an ARM interrupt handler – but it only works once. Or maybe more, but that’s impossible to determine, because … if called from an interrupt handler, the standard library’s longjmp
leaves the the ARM interrupt system in an illegal state (the corresponding interrupt number’s bit is set in the NVIC IABR register and cannot be cleared by software) and no further interrupts handlers will ever be executed again until after a hardware reboot.
Much pain, confusion, hair-pulling and teeth-gnashing was involved in discovering these facts. Once known there was more of the same, with many fine hours spent poring over the detailed and informative but complex and confusing ARM v7-M Architecture Reference Manual. What eventually resulted from this work was a hack-of-all-hacks: Code to manipulate the bizarre state of the call stack inside an interrupt handler, which inserts the longjmp
address in place of the normal return address so that upon arriving there more custom cleanup code can restore the registers and stack to their original setjmp
conditions. See .macro irq_handler_exit
in buck50_asm.s and the // setjmp
and longjmp_return:
sections in main()
in buck50.cxx for the gory details.
Side note: ARM seems to pride its interrupt handling architecture for both its speed and efficiency, and the fact that any normal C function can be used as an interrupt handler without special compiler support to insert special entry setup and exit teardown code. To accomplish this, the interrupt hardware saves all CPU registers on the stack before invoking the handler – “buck50”’s crazy return address manipulation and the failure of the standard library’s longjmp
is due to the CPU’s hacking the normal link register as part of the process. I personally would happily trade the ability to use normal functions as interrupt handlers for a faster interrupt entry (not saving all the registers, thus requiring the handler to save any it will use, or all if it is not a leaf function and is going to call others with unknown register usage) and, as a side benefit, allowing standard longjmp
to work from an interrupt handler.
Any thoughts, ideas, insights, corrections, or proposed improvements to the above explanation or the code it describes would be welcome.
One more hack, not as radical as the ones above …
As per the infinite loop code/pseudocode above and other documentation, the buck50 firmware stores logic analyzer samples only when one or more signal lines change level. If this was not the case, storing samples at 6 MHz into less than 5K of sample RAM would result in an unusable c. 1 ms maximum sampling time.
The samples are timestamped with the value of the 24-bit ARM “systick” timer running at full 72 MHz core clock speed 33. This presents a problem as the 24-bit counter rolls over every 0x1000000/72e6==0.233 seconds – if no signal changes within that period, the elapsed time between samples can only be computed as N*0.233+T where “T” is the delta between the systick values and N is an unknowable integer.
The solution is to internally connect one of the STM peripheral timers to an output port and have it toggle faster than 1.0/0.233 Hz. This introduces a fictitious level change causing at least one sample to be stored within the systick rollover period, which makes “N” uniquely 1 and the delta time between samples deterministic. The extra c. 4 Hz sample memory consumption is minimal, and extra timer port bit is shifted out in the same instruction that combines the systick value with the 8 input ports, i.e. at with no time penalty 34.
With all the above assembly code and custom implementations, why is “buck50” still limited to an approximately 6.25MHz sampling rate?
Good question, and one that I’d appreciate any help with in finding a conclusive answer. Once again, the original conceptual design was based on an estimate that 8MHz would be achievable with register ping-ponging and a 4 times unrolled loop.
This was to be implemented with the four “unrolls” looking something like:
6884 ldr r4, [r0, #8]
688d ldr r5, [r1, #8]
ea45 5504 orr.w r5, r5, r4, lsl #20
429c cmp r4, r3
bf18 it ne
f842 5b04 strne.w r5, [r2], #4
(with alternate unrolls swapping r3
and r4
) and a single amortized branch back to the beginning. (Disassembled binary instructions included for reference in discussion below.)
The ARM Cortex-M3 Processor (Revision: r2p1) Technical Reference Manual supposedly gives the number of cycles required for every instruction. There is some vagueness in its specifications (“An IT instruction can be folded onto a preceding 16-bit Thumb instruction, enabling execution in zero cycles.”), but basically most instructions are documented to take 1 cycle each, with loads and stores taking 2, and branches taking “1+P” where “P” is “The number of cycles required for a pipeline refill. This ranges from 1 to 3 depending on the alignment and width of the target instruction, and whether the processor manages to speculate the address early.”
So given the above, and assuming the it
instruction takes zero cycles due to the preceding 16-bit Thumb cmp
instruction (cmp
with two “low” r0
to r7
registers can fit into the 16-bit Thumb format), that’s 8 cycles for the loop. Add 4 cycles for the amortized-over-4-unrolls branch – the likely worst case scenario of 1+3 given the distance of the branch almost certainly causing a full pipeline flush – and that’s 9 total which at a 72MHz clock speed is 8MHz per sample. And that’s without the benefit of the TRM’s claim of “Neighboring load and store single instructions can pipeline their address and data phases. This enables these instructions to complete in a single execution cycle.” (“single instructions” in this context means load and store a single register, as opposed the the ldm
, stm
, push
, pop
, etc instructions which load or store multiple registers at once.)
Unfortunately, the TRM is work of complete fiction. I’ve done extensive testing, both inside buck50 and with standalone code, and every instruction takes more cycles than is documented. If nothing else, there is a difference between the same assembly opcode depending on whether its arguments (high-vs-low registers, addressing modes) require 16- and 32-bit instruction encoding – a difference never mentioned in the TRM. Also, branches typically take 5 cycles, not 4.
Note that the Cortex-M3 core in the STM32F103 series MCUs is a relatively simple CPU. It has a 3-stage pipeline (source: Cortex-M3 Devices Generic User Guide, © 2010), and limited branch prediction (“speculatively prefetches from branch target addresses”, ibid).
One possible explanation I’ve seen online is that the cycle counts listed in the TRM are for CPU execution only and do not include additional time required for memory accesses and other issues related to the non-ARM logic that ST and other licensees add to their chips. Firstly, if true, doesn’t that make the TRM documentation fairly useless, especially if ST doesn’t publish their own amended specs with the full numbers? Second, ST does document specs on their chip’s internal buses (AHB, APB1, APB2, etc). But the “buck50” code is accessing GPIOB which is on the APB2 bus operating at full 72 MHz speed, and I believe that the systick
timer is part of the core or directly connected to it and also should be accessible without slowdown.
Finally, the biggest mystery is this: Despite every piece of documentation and online forum/blog/etc advice I’ve seen stating that to run at the fastest possible speed code should execute from RAM, I found (to my great surprise) that “buck50” runs faster when executed directly in flash memory(!!). This is of course with flash prefetch enabled, but according to the RM0008 Reference manual STM32F101xx, STM32F102xx, STM32F103xx, STM32F105xx and STM32F107xx advanced Arm®-based 32-bit MCUs the prefetch buffer is only a “Read interface with prefetch buffer (2x64-bit words)”, so two of the 16-byte loop unrolls above will fit, but no more and that doesn’t include the branch instruction. Nevertheless, all of the logic mode=
variants, regardless of loop size, run faster in flash. Note also that standalone experiments show a speedup if flash wait states are reduced from the recommended value of 2 (with a 72MHz main clock) but doing so is of course ill-advised.
In any case, the logic code-mem=ram
option was left in buck50.py
for purposes of experimentation. Note that only small, sampling mode specific, amounts of executable code are copied from flash to RAM (dynamically, when the logic
command is run) so reduction of available sample memory is minimal. Any interesting reports comparing =ram
to =flash
would be greatly appreciated.
One more thing: Given the insanely low “street price” of the “Blue Pill” boards (far lower than even large quantity pricing for the STM32F103xb chip alone from reputable U.S. distributors, much less the added cost of the crystal, USB connector, LEDs, reset switch, passives, the PC board itself and assembly) there exists the possibility that these are clone/bootleg/counterfeit chips. (There is much information on the subject available online.) Regardless, my amateur speculation on the topic is that even if counterfeit and not up to ST’s quality standards, that would result in analog/silicon-level problems such as the chips not running stably at their full rated speed. It seems likely that a fake chip would use stolen VHDL IP or IC masks, compared to the cloner generating a design from scratch using API documentation but getting it “wrong” resulting in the fake design requiring more cycles per instruction than the genuine article.
Pardon my rancor on the subject, but I spent an inordinate amount of time trying to uncover these facts. I may have never embarked on this project had I known them in advance – 8MHz sampling was an original design goal that I had to abandon because of them. Once again, any definitive information, insights, corrections or other advice on the topic would be welcome.
Source code for the buck50 firmware and its associated dependencies is included in the directory tree rooted at build.
The top-level directory contains ready-to-flash pre-built binaries buck50.elf, buck50.bin, and buck50.hex (see Firmware installation, above), a Makefile for building them from source, and the buck50.py host driver program.
The buck50.cxx implementation was written using the regbits, regbits_stm, and the 35 papoon_usb libraries. All are included here in the build/include/thanks4opensource and build/util/stm32f10_12357xx subdirectories to avoid the need for linked or separate repository downloading. Additional dependencies are provided in those and other subdirectories.
The pre-made binaries have been built with the GNU Arm Embedded Toolchain, release “9.3.1 20200408”. Note that building with this compiler suite requires GCC options of -std=c++17
36 and -O1
or higher (the Makefile and pre-built binaries are -O2
). Also, the Makefile
requires a GCC_ARM_ROOT
environment variable pointing to the toolchain’s installation directory.
Building with other toolchains will likely require option and possibly source code changes, and the user may want to replace the stm32f103_flash_init.c startup code and low_stack_flash.ld with a more platform-appropriate implementations. (See above and the low_stack_flash.ld
source concerning the required memory layout.)
Reports about successes or failures building buck50 in different environments are welcome.
$ egrep 'VERSION.*{' build/src/buck50.cxx
VERSION [] = {0, 9, 2},
$ ./buck50.py --version
buck50.py 0.9.2
Copyright 2020 Mark R. Rubin aka "thanks4opensource"
set_pre_input_hook()
so ?
alternate help syntax can immediately retype line without requiring user to press the UP
key.spi
command.gpio
command. Dedicate one PA0–7 line for strobe, set data on remaining ports and pulse strobe high/low (configurable, and with configurable timing). Implement similar option for pb4-11
input in monitor
command.buck50.py
host driver program can flash firmware onto a second “Blue Pill” STM32F103 MCU via their respective serial ports (see Hardware connections, above, and “help usart pty” in the buck50.py
program), but … those damn turtles again.byte
and string
, and the disaster that is Unicode remains regrettable.gnuplot
doesn’t have command completion, and its hierarchical help system is, IMHO, lacking compared to buck50.py
’s. YMMV, and it also should be noted that gnuplot
’s capabilities are more extensive and complex in comparison.TAB
key, Luke!”buck50.py
code. Open source FTW.gnuplot
wasn’t designed to be a waveform viewer.buck50.py
’s configure adjust trim=
command.longjmp
, would it?-std=c++11
was needed for the release 7 and 8 versions. The full reason for this is unknown.