DCC++ intermittent communications failure

[wilwahabri]

G'day, I have assembled a Geekcreit (Chinese Clone of a UNO) and an Arduino motor shield.
I have downloaded base station V 1 from Github and loaded it into the Uno.
I have downloaded the latest version of JMRI from their website and installed it.
I have Engine Driver running on my Android phone. There is never a communications problem between JMRI and Engine Driver.
JMRI is running on a Lenovo ideapad 100S with Intel Pentium N3710 processor running at 1.6GHz, 4Mb RAM, under Windows 10 Home, build 1909, Screen saver disabled.

It all runs perfectly well for about half an hour to 45 minutes and then the JMRI software seems to stop communicating with the UNO and the trains carry on running a per the last instruction sent. There might be a correlation between computer idle time and failure but that is not consistent.

The JMRI log indicates that there has been a timeout error when the failure occurs

I am using a single loco with a Lenz Standard + V2 decoder.

Current drain is around 130-150mA at 12 volt supply to the motor board when driving the train and around 40mA idle, no components seem to be getting hot - not even warm in use. Current drain drops to 0 when the track output is switched off.

The baud rate is set in JMRI to 115200 and greyed out and I have set my COM4 port (USB connection) to the same baud rate, 8 data bits, no parity, 1 stop bit.

I suspected my cheap USB cable but changing that out made no difference. Anyone have any ideas?

[WIMorrison]

Should this be moved to the Computer Control part of the forum?

[Nigelcliffe]

Prob should be moved as Iain says. 

 

However, on laptop check that ALL the relevant power saving settings are switched off (may require some digging in control panels).   On Windows there is an annoyance that it will power-off things (silently) when it thinks they are not being used, including USB ports, etc..   Windows idea of "not being used" and everyone elses might differ. 

 

- Nigel

[wilwahabri]

OK good news, after trying on another computer with the same results I changed out the Geekcreit UNO board for a spare and now it is working fine. It appears there may be a faulty UART on the bad board which is failing during the constant transmission of data to and from the JMRI software.

 

Happy bunny!

[wilwahabri]

Unfortunately it appears to have lulled me into a false sense of security and it failed again. I am now using a Geekcreit Mega 2560 board and that is working fine. Seems that the Geekcreit Uno is not compatible.