Picochess + Pi + DGT 3000 standard

[Dirk Sommerfeld]

Hi folks,

I'm a owner of a smartboard and DGT 3000 (Firmware2.02). I had also installed picochess on a Pi. The clock is normal connected to the board (standard cable in mode25). 

If I powerup the pi, I got after a few seconds a sound message "picochess". But nothing happens in the display of my clock. Only after I plug OFF and IN the USB cable from the board, I got a beep and the I can start. Is there any hint to start picochess with out reconnect the board from the scretch?

Thanks Dirk

[Kiko Gomez]

Hi Dirk,

that happens to me also, and I also have a Smart Board, I think that the problem is the initialization of the Board and Clock connection, probably because our board support I2C comm. through clock connector ... may be, anyway, what I do is switch on the clock after picochess sound, and work 99% of times.

BR

Kiko

[RandyR]

Hi Dirk and Kiko,

I don't see this with my USB DGT board. I's sure you have made the required changes for stand-alone setup. If you set the log level to debug in picochess.ini, the log may show something helpful. Also, does anything show on the webserver while the clock is not showing correctly? Have you tried older images?

Randy

[RandyR]

Ha ha. That should have said "I'm sure...".

[Kiko Gomez]

Hi Randy,

I dont see any strange in my log when that happen, and I try all versions from 0.9n, that is the reason I discard Picochess, and start to think about new firmware 4.00 and new USB C in the Smart Board. Also webserver is normal, and issue is solved just reconnecting as Dirk say or restarting without switch off the clock.

Anyway here is my log just in case you are able to see something. Of course the log was taken when the problem happen.

BR

Kiko

[Dirk Sommerfeld]

Hi Randy, I will look tomorrow for the logfile....  I have this Problem in V 0.9 and also V2. 

After replugin, it works fine

[RandyR]

Kiko,

I finally had a chance to test the RPi4 stand-alone and compare my picochess.log to yours. I does look like there is a problem with the initialization of the clock - board communication using the new smartboards (v4.00) compared to the wood USB E-board (v1.8) that I have. Also, my board appears on /dev/ttyUSB0 whereas yours shows /dev/ttyACM0. That probably has to do with the new interface. One thing I noticed was that I am able to get the clock version through the interface (v2.20) at startup whereas I don't see that in your log. It's only afterwards that the display on the clock happens:

DEBUG         hw - _display_on_dgt_3000: [pico 3  ]

I don't know if Lucas from DGT is monitoring this thread - maybe he could provide more insight.

Perhaps you and Dirk need a small delay at startup. I'll attach my log if you want to compare.

Randy

[Dirk Sommerfeld]

I found this in my logfile....

2021-02-28 16:33:51.365 WARNING      board - write_command: (ser) clock is already locked. Maybe a "resend"?

2021-02-28 16:34:07.984 WARNING      board - _process_board_message: (ser) clock strange old time received (0, 17, 72, 0, 17, 64, 18) l:(0, 11, 30) r:(0, 11, 46)

2021-02-28 16:39:05.587 WARNING      board - _process_board_message: (ser) clock strange old time received (0, 16, 8, 0, 16, 54, 18) l:(0, 10, 36) r:(0, 10, 7)

(END)                                                         

[Dirk Sommerfeld]

Now with debug level:

2021-03-02 07:38:23.320   DEBUG      iface - _process_message: (web) handle DgtApi: DGT_CLOCK_SET started

2021-03-02 07:38:23.322   DEBUG dispatcher - run: (ser) max timer not running => processing command: DGT_CLOCK_START

2021-03-02 07:38:23.323   DEBUG      iface - _process_message: (web) handle DgtApi: DGT_CLOCK_SET ended

2021-03-02 07:38:23.324   DEBUG dispatcher - _process_message: (ser) handle DgtApi: DGT_CLOCK_START

2021-03-02 07:38:23.328   DEBUG dispatcher - run: (web) max timer not running => processing command: DGT_CLOCK_START

2021-03-02 07:38:23.330   DEBUG dispatcher - _process_message: (web) handle DgtApi: DGT_CLOCK_START

2021-03-02 07:38:23.334   DEBUG      iface - _process_message: (web) handle DgtApi: DGT_CLOCK_START started

2021-03-02 07:38:23.335   DEBUG      iface - _process_message: (web) handle DgtApi: DGT_CLOCK_START ended

2021-03-02 07:38:24.638 WARNING      board - _watchdog: (ser) clock is locked over 2secs

2021-03-02 07:38:24.644   DEBUG      board - _watchdog: resending locked (ser) clock message [[<DgtCmd.DGT_CLOCK_MESSAGE: 43>, 12, <DgtClk.DGT_CMD_CLOCK_END: 3>, <DgtClk.DGT_CMD_CLOCK_ASCII: 12>, 85, 83, 66, 98, 111, 97, 114, 100, 0, <DgtClk.DGT_CMD_CLOCK_END_MESSAGE: 0>]]

2021-03-02 07:38:24.645   DEBUG      board - write_command: (ser) board put [DgtClk.DGT_CMD_CLOCK_ASCII] length: 14

2021-03-02 07:38:24.646   DEBUG      board - write_command: sending text [USBboard] to (ser) clock

2021-03-02 07:38:24.648   DEBUG      board - write_command: (ser) clock is locked now

2021-03-02 07:38:26.103 WARNING      board - _wait_for_clock: (ser) clock is locked over 3secs

2021-03-02 07:38:27.641 WARNING      board - _watchdog: (ser) clock is locked over 2secs

2021-03-02 07:38:27.643   DEBUG      board - _watchdog: resending locked (ser) clock message [[<DgtCmd.DGT_CLOCK_MESSAGE: 43>, 12, <DgtClk.DGT_CMD_CLOCK_END: 3>, <DgtClk.DGT_CMD_CLOCK_ASCII: 12>, 85, 83, 66, 98, 111, 97, 114, 100, 0, <DgtClk.DGT_CMD_CLOCK_END_MESSAGE: 0>]]

[RandyR]

Hi Dirk,

Can you attach your picochess.log showing the startup? You can do so from the terminal or ssh:

sudo service picochess stop

sudo rm /opt/picochess/logs/picochess*

sudo shutdown -h now

Give it some time to complete the shutdown (10 sec?). Then power ON. After picochess starts, wait 10 sec, then:

sudo service picochess stop

Attach your picochess.log file.

Also, is it only reconnecting the USB that gets it to work? What happens when you reconnect the clock cable? What about when rebooting without shutting down?

Randy

[Dirk Sommerfeld]

Randy, will do...

The Board is connected and I can make moves with the voice from the PI... Only the DGT3000 ist of.

The Clock will available after a reconnect of the USB Port from the smartboard. After this, I got a beep and can read the message on the display.

Dirk

[RandyR]

Ok, Dirk, but just unplugging/reconnecting the clock cable does not fix the issue. Is this correct?

Also, what about just rebooting without shutting down? Does the clock work correctly after the second boot without touching the cables?

Randy

[Dirk Sommerfeld]

Hi Rob... this is correct.... Reconnect of the clock has no effect... Yesterday I made a warm reboot and when I come back, the clock was available.

Dirk

[Dirk Sommerfeld]

Hi Randy,

logfile attached. Bix thx for help !!!

[RandyR]

Hi Dirk,

Yes, same thing as Kiko. It looks like you get the board version, but not the clock version after starting up here:

2021-03-02 17:28:58.752   DEBUG  picochess - main: #################### PicoChess v200 ####################

I'm just guessing but I suspect there is not enough time for the board to set up communications with the clock before picochess starts. You could confirm this by starting picochess manually:

sudo service picochess stop

sudo systemctl disable picochess

sudo shutdown -h now

Power OFF the Pi and turn OFF the clock. Then, turn ON the clock and set mode 25. Power ON the Pi and issue the following commands:

sudo service picochess start

Does the clock work? If so, maybe we can delay the start of picochess.

To go back to having picochess start automatically:

sudo service picochess stop

sudo systemctl enable picochess

One other suggestion - make sure the clock battery is good.

Randy

[RandyR]

Also, you might want to try picochess v3. :^)

[Dirk Sommerfeld]

Randy... first of all... I have an other SDcard with Pico 3.0... same issue....

I stop the service... 

I run the systemctl to revoe it from autostart

shutdown everything..... and start the pi with the clock on int25  (battary's full)

after i start the service... nothing 

I stop the service...

I start again and now.... Pic0 3.0 is in the display....

I need two times.. stop and start have the same effect wie replugin the USB cable.....

Dirk

[RandyR]

Ok. That's unfortunate.

Maybe we can see something in the kernel log. Go through the same startup from power OFF. Once picochess is running, remove/reinstall the USB connection to get the clock working. Then run the following:

dmesg > dmesg.txt

Attach it to your reply.

Randy

[RandyR]

Oops. By 'attach it' I mean attach dmesg.txt. 

[Dirk Sommerfeld]

here I'm back with the kernel message

hth

[RandyR]

It looks like you need a stronger power supply. I can see a few "Under-voltage detected!" entries. I'm not sure if that will affect your issue, but it's possible. Other than that, I don't see anything obvious. I wonder if you used a powered USB hub if it would make a difference.

Sorry I can't be of more help.

Randy

[Dirk Sommerfeld]

oki, thanks.... I have the same issue on a pi zero with an other power adapter....

thanks for your time....

[Kyle Washington]

I've already shut everything off downstairs for tonight, but I will try some of this tomorrow.  This sounds so familiar!  Maybe my problem is something with timing and not with my homemade cable!  :)

This is good information.  I will try to find some log files and see if I can at least help shed some light.

Kyle

[Kyle Washington]

To recap what I wrote in a different thread - I got Picochess running on my RPi 4 with a homemade cable in an Argon One M.2 case.  I wanted it to be running on my regular Pi case with a different Pi4, so I moved the cable. Suddenly I started getting my old errors as are described here.  I ended up trying both Pi4s that I have, and both give the error. I moved everything back to the M.2 case, and it works again.  The case shouldn't change the pinouts at all, so I don't get it... but I wondered if it had something to do with the timing issues described here?

So to avoid all of the moving back and forth, I made a new cable.  Now I know more about where it needs to be stronger and my crimping is becoming more natural, so I thought I'd just make a better (and longer) cable.  It works.  Both Pi4's work. That's the good news!  So I tried the old cable, and it only works in the M.2, which I cannot explain.  All of the wires show connectivity on my meter, and the clock initially responds... but it defaults to --:----.-- for the time and doesn't work after that.  I haven't looked at this log file yet, but I will attach it as I'm about to go into a meeting.. and I'll read it during that time while you guys are all solving my problem - lol.

The good news is that my other cable will let me start playing, but I'd like to learn more about why this error happens.  If it's just connectivity, maybe I could replace a wire even though it's ok for the meter...  At the very least, maybe it will help others with the error in comparison.

Kyle

[RandyR]

Well, the log isn't very helpful. It basically says you have an issue with communication with the clock. There could be a couple possible issues with the bad cable. There could be an intermittent connection that you don't see when using a meter. There could be some resistance in the wires that degrade the signal (connectivity does not imply 0 ohms). Also, if the wires aren't shielded you could be picking up noise from the Pi (or another source) depending on how the wires are routed. And, low battery level in the clock or insufficient power in the Pi could screw with the communication.

If I were you I'd be happy I now have a serviceable cable to carry on with and enjoy the fruits of Dirk's labour. :^)

Randy 

[Kyle Washington]

That's what I'm thinking too.  I

[Kyle Washington]

(No idea why it sent before I finished typing).

That's what I'm thinking too.  It may be the new case helping to shield from the interference... If it works now, it doesn't matter!  :)

Kyle

[Dirk Sommerfeld]

Randy, one last question.. 

What is the setup for pc without any DGT clock... I use will use the webinterface as standard and got erromessage that the clock would not be found...

Dirk

[RandyR]

Hi Dirk,

Although I always use a clock myself, just don't plug it in to the board. You can ignore the error messages in the log. Just make sure you are set up for standalone mode. You can also have a look at this post.

Randy

[Dirk Sommerfeld]

Randy, I love you.. thanks ;-)

[Randy Reade]

 🤣🤣 ☺

--
You received this message because you are subscribed to a topic in the Google Groups "PicoChess" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/picochess/020YSNjipqo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to picochess+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/picochess/14741f6a-f803-4945-8b2a-bc41c6df60c3n%40googlegroups.com.

[Dirk Sommerfeld]

Hi Randy, a small update from my side. I made a new test with an other DGT Board. I try it with a Bluetooth/USB Board. With both connection types, the clock work without problems.

I think there must be something different in the Firmware of my Smartboard. 

Best Dirk

[RandyR]

Hi Dirk.

I agree. I think you should talk to DGT support. They are very helpful. According to their web page, it is supposed to be compatible with the DGTPi, which to me means compatible with PicoChess. Maybe there needs to be a software update for the new smart boards.

Randy

[Kiko Gomez]

Hi Dirk,

the connection problem between the DGT Smart Board and DGT3000 is gone with the PicoChess Hat, also the "clock is locked ...." entries in the picochess.log ;)

Best Regards

Kiko

[Dirk Sommerfeld]

Hi Kiko,

what you mean? The Problem with the smartboard isn't fixed yet. I wait for the DGT Support....

LG

Dirk

[Kiko Gomez]

Hi,

with normal connection between board & clock I need also restart the PI or disconnect and connect the USB cable to the RPi before get communication with the clock. But with the Picochess Hat the connection is always sucessfully, even if the clock is off, will switch on when you switch on the RPi and switch off automatically when you switch off the RPi. Because the connection is directly to the RPi and not through the board.

BR

Kiko

[Dirk Sommerfeld]

Hi Kiko,

all right. The problem exists only with a newer smartboard and an external clock. DGT is working on it. I will wait if I can update my board.

Dirk

[Kiko Gomez]

Good to know that DGT are working on that, please keep us updated because this problem will happen on Smartboard and new DGT wood boards with USB C. (same hardware I think)

BR

Kiko

[mvanthoor]

Hi. Today I discovered that I have this exact problem with a wooden DGT e-board with USB-c connector. I have an older DGT e-board with Mini-USB which works fine. I can get Picochess to work IF it starts and beeps, and then says "USBboard"; then I can power the clock off and on again and then it'll work.

Has there been any news from DGT; CAN a wooden e-board actually be updated? (The BT boards have a firmware uploader... but I see nothing for the normal wooden boards.)