Duet3/pi DCS is unavailable after wiring 24V

fractalengineer

@bearer no I think I got a refurb; there was only the board and SD card in it

chrishamm

@Danal That error occurred because SPI wasn't enabled at that point, hence no /dev/spidev.0.0. It's always possible to define module options in /etc/modprobe.d, that doesn't require the driver(s) to be loaded.

It must be the connection from the Pi to the Duet 3 (ribbon cable?). The SPI link uses pins 19-24 (see https://pinout.xyz/pinout/pin22_gpio25#) so I'd once more check continuity between these pins on both sides. The error you see in the log without any extra comment (more verbose in v2.1.0) means that no signal could be transferred over pin 22 (BCM25). DSF expects this pin to be high when it starts.

A Former User

@chrishamm said in Duet3/pi flashing Diag LED after wiring 24V DCS is unavailable:

It must be the connection from the Pi to the Duet 3

of curiosity what does DCS say if it gets a response, but the wrong due to say wrong firmware or spi clock?

chrishamm

It's also worth starting DCS with -l debug to see a few more details. If the SPI clock is too high, you usually see warnings about failed transfers and/or checksum errors.

A Former User

@chrishamm said in Duet3/pi flashing Diag LED after wiring 24V DCS is unavailable:

It's also worth starting DCS with -l debug to see a few more details. If the SPI clock is too high, you usually see warnings about failed transfers and/or checksum errors.

-l debug shows the difference, no difference without it.

#with 22 low
pi@raspberrypi:~ $ /opt/dsf/bin/DuetControlServer -l debug
[debug] Lost connection to Duet
   System.OperationCanceledException: Timeout while waiting for transfer ready pin ...
[error] Duet is not available

#with 22 high
pi@raspberrypi:~ $ /opt/dsf/bin/DuetControlServer -l debug
[debug] Lost connection to Duet
   System.OperationCanceledException: Board is not available (no header) ...
[error] Duet is not available

Maybe change the first to Duet is not ready?

edit: sudo wasn't needed (at that point in starting up at least)
edit2: whops, wasn't testing the latest version, error changed to fatal, but same string.

chrishamm

Yes I thought it already showed more details so I'm going to improve that in v2.1.1.

fractalengineer

@chrishamm I just checked continuity once more; ever pin beeps from one board to the other.

Also checked for any short to ground; it’s all clear with the multimeter.

I’ll try wiring pin by pin using bridges; really can’t figure out what else to do

Unless the previous duet fried the pi4 too...

A Former User

@fractalengineer said in Duet3/pi flashing Diag LED after wiring 24V DCS is unavailable:

really can’t figure out what else to do

if you've got a steady hand you could measure the voltage on pin 22 with the Duet running (before and after starting DCS) and then cross refrene with the output of running DSC with -l debug to at least verify if the Pi is reading pin 22 correctly.

see next post instead:

A Former User

Hold off on that; not sure why loopback testing didn't occur to me a week ago ... but:

https://github.com/rm-hull/spidev-test

combine with wiringpi's gpiocommand and two jumpers to do full (or 80%? as SCK isn't tested) diagnostic on the rPi.

place jumper over 22-24 and 19-21, then run commands as shown, compare to output below.

(Quick copy/paste to get the spidev-test tool and wiringpi):

sudo apt install -y git wiringpi
git clone https://github.com/rm-hull/spidev-test ~/spidev-test
cd ~/spidev-test
gcc spidev_test.c -o spidev_test

Commands to run:

RDY=22 CS=24 ; { 
gpio -1 mode $CS out
gpio -1 mode $RDY in
gpio -1 write $CS 1 && echo "(Pin RDY/$RDY) `gpio -1 read $RDY` should equal `gpio -1 read $CS` (Pin CS/$CS)"
gpio -1 write $CS 0 && echo "(Pin RDY/$RDY) `gpio -1 read $RDY` should equal `gpio -1 read $CS` (Pin CS/$CS)"
{ ~/spidev-test/spidev_test -v -s 8000000 -D /dev/spidev0.0 && echo RX should equal TX. ;} | tail -n3 | cut -b-100 ;}

What output should look like:

(Pin RDY/22) 1 should equal 1 (Pin CS/24)
(Pin RDY/22) 0 should equal 0 (Pin CS/24)
TX | FF FF FF FF FF FF 40 00 00 00 00 95 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF F0 0D
RX | FF FF FF FF FF FF 40 00 00 00 00 95 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF F0 0D
RX should equal TX.

Repeat with wires jumpering the far side of the ribbon cable if test with jumpers on the rPi is passed.
(Print quality comments can be addressed to Tiertime pending full Duet conversion of the Up Mini 2)

Danal

I don't know about the "W" part... but I'll take a look.

I can also spiff it up so that its output is very clear for an "end user", in the Duet environment.

A Former User

@Danal wrong topic?

Danal

I thought a post said: "If a C Wizard happens along"... and then got edited... the request being to make spidev_test more duet specific, easier to interpret. ?

Hopefully, because I did take a look.

I am not a Wizard, but:

./DuetTestSPI
spi mode: 0x0
bits per word: 8
max speed: 8000000 Hz (8000 KHz)
TX | FF FF FF FF FF FF 40 00 00 00 00 95 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF F0 0D  | ......@....�..................�.
RX | FF FF FF FF FF FF 40 00 00 00 00 95 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF F0 0D  | ......@....�..................�.
Loopback on pins 19-21, SPI MOSI MISO tested successfully.
All tests passed, SPI should work for Duet 3 <> Raspberry Pi.

(removed jumper)

./DuetTestSPI
spi mode: 0x0
bits per word: 8
max speed: 8000000 Hz (8000 KHz)
TX | FF FF FF FF FF FF 40 00 00 00 00 95 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF F0 0D  | ......@....�..................�.
RX | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  | ................................
Send and receive buffers do not match! 
Check jumpers on pins 19-21; if they are correct this machine has an SPI problem. 
Aborted

Much clearer messages, etc. This is 1/2 of your suggestion above, I'll do the 22-24 part in a few min, and then post a link.

A Former User

@Danal said in Duet3/pi flashing Diag LED after wiring 24V DCS is unavailable:

I thought a post said: "If a C Wizard happens along"... and then got edited...

it did, but "W" didn't make much sense; anyways, as DSF depends on wiringpi anyways figured less effort than forking the spi tools. (and easier to change if other boards have different gpio mapping)

Danal

Understood... it was obscure. Perhaps should have said "not a Wiz, but"

Danal

OK, here you go, a 'tester' for SPI on a Pi, to see if it is hardware OK to run with a Duet.

Installation:  Either:
  git clone https://github.com/DanalEstes/DuetTestSPI
or, if you want just the binary:
  wget https://github.com/DanalEstes/DuetTestSPI/raw/master/DuetTestSPI
  chmod 744 DuetTestSPI

Run
Place a jumper on Pi GPIO pins 22-24 and 19-2, then:

./DuetTestSPI

All messages should be self explanatory.

Danal

@bearer

I invoked the great architectural pattern of "reuse" and put your jumper pic on the github for this. I will pull it if you'd rather I not use it.

A Former User

Everything on the forum is CC-BY-SA AFAIK.

I think I'll more or less just add some curly bracets to clean up the output and stick with it as is; simple to understand, simple to change if someone uses a different user configurable ready pin or spi device, stays up to date with past, present and future (physical) pin mappings. It also provides the techies with all the jucy details without dumming it down.

If the gloas was to make it shorter, it'll all fit on one line, but defeats the simple to undertand and simple to change part. Could even tee and grep it for humanreadable output, but as the SCK signal isn't covered saying "all good" feels wrong, miss matched data will suffice as a failed test.

pi@raspberrypi:~ $ RDY=22 CS=24 ; { gpio -1 mode $CS out ; gpio -1 mode $RDY in ; echo "Pin RDY/$RDY Pin CS/$CS" ; gpio -1 write $CS 1 && echo "         `gpio -1 read $RDY`         `gpio -1 read $CS` <-should be equal" ; gpio -1 write $CS 0 && echo "         `gpio -1 read $RDY`         `gpio -1 read $CS` <-should be equal" ; echo TX should equal RX: ; ~/spidev-test/spidev_test -v -s 8000000 -D /dev/spidev0.0 | tail -n2 | cut -b-100;}
Pin RDY/22 Pin CS/24
         1         1 <-should be equal
         0         0 <-should be equal
TX should equal RX:
TX | FF FF FF FF FF FF 40 00 00 00 00 95 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF F0 0D
RX | FF FF FF FF FF FF 40 00 00 00 00 95 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF F0 0D

A Former User

@fractalengineer the lovely Duet people released new updates RFF3.01-RC10 and DSF2.1.1.

While neither will solve your problem, it'll help diagnose the SPI issues with clearer error messages printed without needing the -l debug parameter.

Between that and the loopback test we should have done ages ago I think we should pinpoint the problem at last.

Fingers crossed!

fractalengineer

@bearer @Danal wow thanks for the overwhelming help guys; really grateful for the support

Even though most of it goes above my head as I'm not a codehead at all, I managed to install and run the SPI test on my own pi3

@bearer I bought the dupont cables and bridged the pin as per your picture; here's the results of the SPI test:

pi@duet3:~/spidev-test $ RDY=22 CS=24 ; {
> gpio -1 mode $CS out
> gpio -1 mode $RDY in
> gpio -1 write $CS 1 && echo "(Pin RDY/$RDY) `gpio -1 read $RDY` should equal `gpio -1 read $CS` (Pin CS/$CS)"
> gpio -1 write $CS 0 && echo "(Pin RDY/$RDY) `gpio -1 read $RDY` should equal `gpio -1 read $CS` (Pin CS/$CS)"
> { ~/spidev-test/spidev_test -v -s 8000000 -D /dev/spidev0.0 && echo RX should equal TX. ;} | tail -n3 | cut -b-100 ;}
-bash: gpio: command not found
-bash: gpio: command not found
-bash: gpio: command not found
-bash: gpio: command not found
TX | FF FF FF FF FF FF 40 00 00 00 00 95 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF F0 0D
RX | FF FF FF FF FF FF 40 00 00 00 00 95 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF F0 0D
RX should equal TX.

IF I REMOVE THE JUMPERS:

pi@duet3:~/spidev-test $ RDY=22 CS=24 ; {
> gpio -1 mode $CS out
> gpio -1 mode $RDY in
> gpio -1 write $CS 1 && echo "(Pin RDY/$RDY) `gpio -1 read $RDY` should equal `gpio -1 read $CS` (Pin CS/$CS)"
> gpio -1 write $CS 0 && echo "(Pin RDY/$RDY) `gpio -1 read $RDY` should equal `gpio -1 read $CS` (Pin CS/$CS)"
> { ~/spidev-test/spidev_test -v -s 8000000 -D /dev/spidev0.0 && echo RX should equal TX. ;} | tail -n3 | cut -b-100 ;}
-bash: gpio: command not found
-bash: gpio: command not found
-bash: gpio: command not found
-bash: gpio: command not found
TX | FF FF FF FF FF FF 40 00 00 00 00 95 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF F0 0D
RX | 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
RX should equal TX.

Looks like the Pi is fine?

Onto updating everything then.

If the ribbon still gives us trouble I'll be able to wire using the duponts

A Former User

@fractalengineer said in Duet3/pi flashing Diag LED after wiring 24V DCS is unavailable:

Looks like the Pi is fine?

40% fine, seems duet have changed a dependency so you're missing the GPIO utility. but the data part of the SPI bus is fine.

run sudo apt install -y wiringpi and do it again.

and you can also test the ribbon by making some male jumpers (or use male to male duponts for that matter).

It'd be interesting to see what DuetControlServer says after the update, because its looking less likely to be a rPi fault.