Firmware developent status

[puff]

5ms timeout?

[AndreyB]

5ms timeout?

Please use more words

[puff]

there seems to be a 5 ms pause between these two bytes - could be a timeout

[AndreyB]

there seems to be a 5 ms pause between these two bytes - could be a timeout

So we both see two bytes on the wire yet my code fails to read at least one byte. At the moment I would be happy to just read one byte. Not sure how is that 5ms relevant for my current cry for help to read one byte

[puff]

what's the schematics? which exactly wire is that? kline usually works at much higher levels (12v).
can it read bytes sent from usb-uart adapter?

[AndreyB]

Thank you!!! That's actually an amazing idea for me to try next.

It's https://github.com/rusefi/rusefi_documentation/raw/master/Hardware/Hellen/hellen125honda-a-schematic.pdf it already after the LIN transceiver so I show RX exactly how it goes into stm32. had to add a 1K pull-up since that's missing in the schematics at the moment

[puff]

"No LIN transceiver is fully compliant to the K-LINE, but compatible. You can see the difference from the document. So I think TJA1021 can work"
https://community.nxp.com/t5/MPC5xxx/TJA1021-K-LINE-compatibilty/td-p/550725
unfortunatelly, it's 403 for me.

[AndreyB]

rusEFI discovery, adafruit, secondary serial feature with perfect timing https://github.com/rusefi/rusefi/issues/5105 fix

9600 and 14400 works while 10400 does not work. Smells like uart configuration magic?!

image.png (20.03 KiB) Viewed 58622 times

image.png (19.96 KiB) Viewed 58622 times

image.png (19.73 KiB) Viewed 58622 times

[puff]

10400 is not supported by some usb-uart converters... (that's why people prefer ft232)

[AndreyB]

I got a chance to try my fake cheap eBay 16 channel logic analyzer! Looks like CP210x has performed as expected at 10400?

10400

image.png (39.83 KiB) Viewed 58610 times

9600

image.png (41.27 KiB) Viewed 58610 times

[puff]

you can tell for sure by measuring the length of the start bit probably.
still, didn't get: it doesn't work even when bytes are sent from the usb-uart dongle? so, it's a firmware bug?

[AndreyB]

you can tell for sure by measuring the length of the start bit probably.
still, didn't get: it doesn't work even when bytes are sent from the usb-uart dongle? so, it's a firmware bug?

As a test step I've decided to run proven and known to work TS protocol at 10400 not k-line connector. And at the moment TS protocol does NOT run at 10400 which makes me stretch my head and post https://forum.chibios.org/viewtopic.php?f=16&t=6194 wondering if I can blame stm32 or ChibiOS

[mck1117]

please use a dedicated thread instead of clogging a pinned announcement thread

[AndreyB]

pinned announcement thread

What pinned thread? What announcement thread?

[AndreyB]

Wow just wow.

The fix for https://github.com/rusefi/rusefi/issues/5354 turned out to be https://github.com/rusefi/rusefi/commit/1e43299eab2b2af033cb79d03446dbf07e4bd2c3

F4 does not seem to care. F7 cares deeply but compiled only throws us under the train at -Os optimization level.

Kudos @dron_gus for fixing.

[AndreyB]

Daihatsu 4 cylinder cam trigger added https://github.com/rusefi/rusefi/wiki/All-Supported-Triggers#daihatsu

[AndreyB]

Anyone with a Bluetooth board interested to test https://github.com/rusefi/rusefi/issues/5927?

image.png (17.7 KiB) Viewed 33636 times

[AndreyB]

Time flies - it has been over three years since @mck1117 has contributed that amazing performance tracing!
@andreika has just fixed a defect reviving it and I've added a touch of usability.

I have some questions around high frequency CANbus processing with Lua let's see if PTrace would be the magic tool.

image_2024-02-22_21-29-01.png (20.5 KiB) Viewed 31163 times

image.png (89.76 KiB) Viewed 31163 times

[AndreyB]

@dbh97 has revamped the build scripts! I expect way less broken local proteus builds now.

[AndreyB]

ADC sample value just a bit above 12 bits, that's not what one would expect from stm32f7?

Do we have a wild memory overrun which just happens to touch ADC region? Are we not good at reading that ADC app note? Else?

That's https://github.com/rusefi/rusefi/issues/6376

image.png (2.09 MiB) Viewed 23179 times

[AndreyB]

ADC sample value just a bit above 12 bits, that's not what one would expect from stm32f7?

Do we have a wild memory overrun which just happens to touch ADC region? Are we not good at reading that ADC app note? Else?

That's https://github.com/rusefi/rusefi/issues/6376

image.png

Status update:
* finally I am able to reproduce on the bench, vehicle not required
* looks like SD card has to be logging for the issue to occur. Wow, we have some sort of a connection between SD card and ADC?!
* so far I can reproduce only on one type of PnP board

image.png (122.19 KiB) Viewed 15959 times

image.png (22.34 KiB) Viewed 15959 times

[AndreyB]

Another surprise - USB CDC ACM and 64 bytes transaction - https://forum.chibios.org/viewtopic.php?f=35&t=6395