Vida CEM swapping

[dikidera]

I did one of the things you just suggested. I twisted the CANH CANL wires together, this actually made the same SBL produce more duplicates.

I have a 2 channel scope I can use to take a look at the signals but no logic analyzer. And a 12v battery I could use to power the ECU and remove PSU noise temporarily.

[vtl]

I have a 2 channel scope I can use to take a look at the signals but no logic analyzer. And a 12v battery I could use to power the ECU and remove PSU noise temporarily.

Where do you get 3.3v from? How's the transceiver powered?

[dikidera]

I have a 2 channel scope I can use to take a look at the signals but no logic analyzer. And a 12v battery I could use to power the ECU and remove PSU noise temporarily.

Where do you get 3.3v from? How's the transceiver powered?

3.3v and 5v all come from the RPI's pins. The transceiver tja1050 works at 5v levels and as such I had to cut the power trace as it shared them with the controller which supports both 3.3v and 5v. As such I had to solder a wire to the power pin of the tja1050 for a separate 5v source which is the 5v RPI pin I just mentioned.

https://forums.raspberrypi.com/viewtopic.php?t=141052

[vtl]

3.3v and 5v all come from the RPI's pins. The transceiver tja1050 works at 5v levels and as such I had to cut the power trace as it shared them with the controller which supports both 3.3v and 5v. As such I had to solder a wire to the power pin of the tja1050 for a separate 5v source which is the 5v RPI pin I just mentioned.

That would be my first suspect - the onboard voltage regulator. Often they are not great.

[dikidera]

Spot on. I just checked with a multimeter, the 5v line does indeed drop to as low as 4.5v during the tranmissions. Luckily I have a little toy for just this occasion.

[rkam]

The signal looks pretty clean.
Frame takes about 550microseconds, and pause in between the same.

denso-signal.png (26.89 KiB) Viewed 1066 times

Edit:

The signal did not look well at first while the Denso was streaming normally today.
After wriggling the cable that looked bad, the signal was good.

And the log file made while checking the signal shown above contains the full 32768 lines.
So it is quite possible that the signal was poor yesterday when I was missing frames, and is now working.

[dikidera]

Thank you rkam, so it looks like we both have had signal interference, what luck! So this means my code(SBL) works just fine.

This is great as it means we can conduct work faster, dumping memory in seconds is much better than waiting 30 full minutes.

Here is an updated version of the code with a reduced sleep of just 10 microseconds.

Ideally I would've attached the .c and .bin files, but they are forbidden attachments types, so I have zipped them.

[vtl]

Thank you rkam, so it looks like we both have had signal interference, what luck! So this means my code(SBL) works just fine.

This is great as it means we can conduct work faster, dumping memory in seconds is much better than waiting 30 full minutes.

Here is an updated version of the code with a reduced sleep of just 10 microseconds.

Ideally I would've attached the .c and .bin files, but they are forbidden attachments types, so I have zipped them.

Thanks.

Create a github repo?

[dikidera]

I have found that..Github repos are not indexed by google very well and all that information is "lost" there. I had to manually search Github to find a lot of interesting Volvo and CAN related repositories that google didnt yield.

[rkam]

sbl2 works great. Approx. 50 microseconds between frames.
Full 256kB dump in 19.5 seconds.