[not next revision] No wake up with TJA1442AT - 5v CAN power supply

[rusefillc]

canRxAdd(0x570)
function onCanRx(bus, id, dlc, data)
    print('got CAN id=' ..id ..' dlc=' ..dlc)
    id11 = id % 2048
    if id11 == 0x570 then
        mcu_standby()
    end
end

I've soldered TJA1442AT on one unit with stm32f407 MCU

I have CAN validated see https://github.com/rusefi/alphax-2chan/issues/16#issuecomment-1014172777

I issue "stm32_standby" via rusEFI console

i send CAN packets

expected result: wake up

actual result: no wake up :(

[rusefillc]

[rusefillc]

at the moment we are trying to wake up through op-amp, let's try bypassing manually?

@andreika-git where is a good source of +3v and good poking spot to attempt waking up more manually bypassing op-amps?

[rusefillc]

Actually code was not ready https://github.com/rusefi/rusefi/commit/54b70fbc3c9cf320f46783310aaea3ea52877c73

[mck1117]

Should work with opamp

[rusefillc]

https://github.com/rusefi/rusefi/commit/54b70fbc3c9cf320f46783310aaea3ea52877c73 did not help

@andreika-git please advice on test points for jumper wake up test

[rusefillc]

Just verified baseline: wake-up via PA0 works on F407 discovery

[rusefillc]

Hellen does not wake up with +3 or +5 applied to CAN pin 4 with CAN transceiver removed. I am stuck and out of ideas :(

[NMSTEC]

Either I'm misunderstanding, or you guys are trying to wakeup an F4 via CAN RX. F4 cannot wake via any pin other than PA0, and rising edge.

[rusefillc]

This board feeds can_rx into pa0 via op-smp

[NMSTEC]

My bad, apologies.

[rusefillc]

great news! discovery firmware on alphax hardware wake up via 5v on CAN RX pad with transceiver removed!

op-amp divider drama - same discovery firmware on alphax hardware does NOT wake up with 3v on CAN RX pad with transceiver removed!

[rusefillc]

with full erase alphax firmware wakes up on 5v, looks like we had some settings drama or maybe I did not flash latest properly

sounds like next step to address op-amp vs 3v drama.

i assume divider would not be easy to remove due to 4-resistor-arrays. do we want to replace 4-resistor-array with three individual resistors or is that a bad track overall?

[andreika-git]

Let's do a little fix: 1) Find a 510 Ohms 0603 resistor on the Hellen72a board https://rusefi.com/docs/ibom/hellen72a-ibom.html "510","R212, R213, R214, R215, R116, R119, R120, R124","R0603" 2) Solder it over the 0604-x4 here:

[rusefillc]

I've removed CAN transceiver for first round of experiments

I've soldered resistor on top of array as recommended

first tests stand by was rebooting right away, got that a couple of time

took stuff back for inspection / poking, solder looked good

did more tests and behavior changed, in a good way I guess

it goes to sleep into standby and wakes up with just an antenna - i touch CAN RX pad with jumper wire and that wakes MCU up, even with jumper wire floating

[rusefillc]

i went forward and re-install can transceiver

now it again reboots after standby right away :(

[andreika-git]

1) boardPrepareForStop() has palEnableLineEvent() but boardPrepareForStandby() does not 2) PAL_EVENT_MODE_RISING_EDGE --> PAL_EVENT_MODE_FALLING_EDGE ? Rx is "1" by default?

[rusefillc]

@mck1117 any ideas why does it start right after standby command with CAN transiever?

[mck1117]

TJA1051

yellow = CAN_TX pink = CAN_RX (via opamps, connected to PA0) blue = CANH

~when we shut down the STM32 drags CAN_TX down (ie, dominant), and after 1ms the transmitter times out (prevents hung bus), when it times out it goes back high and wakes the STM32~

edit: never mind, I got it to do it once with no edge on rx, with the stm asleep for about 300us before it woke up

[rusefillc]

@mck1117 I am super confused

do we have same or difference behavior? Do you have desired behavior? Because I believe i have desired behavior?

https://youtu.be/c19ICpoVFuE

[rusefillc]

forgot to mention that I had a mod to change op-amp divider ratio

@mck1117 says he had similar mod

[rusefillc]

just removed the mod and board no longer wakes up, the op-amp divider change is mandatory

[andreika-git]

We should test another option: use +5V VIO for CAN driver instead of 3.3V - maybe it's better than trying to bypass the opamp divider...

[mck1117]

That would work (CAN RX is of course 5v tolerant), or we can add a line on the MCU module that goes directly to PA0 instead of thru the opamp.

[rusefillc]

i believe the open question if we should mod one of alphax boards to triple-check or just fab with 5v CAN VIO without a HW test of rev A

@mck1117 any chance you would be available to test rev A of alpha with +5v VIO?

[rusefillc]

[rusefillc]

Complete failure of my experiment - standby always reboots right away unconditionally

op-amp resistor hack removed corner lead lifted red jumper added 5v confirmed in red jumper

same result with following tests: standby request via CAN standby request via USB with external CAN device connected standby request via USB with external CAN device disconnected

[andreika-git]

Does CAN work by itself with the red jumper?

[rusefillc]

standby request via CAN meaning I have a lua script "onCanRecieve of packet 0x570 sleep' so yes, RX worked

just tested can transmit as well just to be sure and it works

[andreika-git]

Proposal: connect a resistor (3..10k?) to CAN_TX and +5V.

[rusefillc]

10K between TX and +5v has changed things but did not solder the overall issue

good news there is no auto-reboot, i.e. "go to sleep" works via CAN now

still broken: no wake up on next CAN message

now I will solder test point for RX

[rusefillc]

I am failing to do these HW mods that's above my skillset, we need to go a different route.

while adding RX test point I've shorted the board somewhere. I had to remove the resistor and red jumper wire to get board back to life

i got board back to life

i've added red wire and TX pull-up

and now it reboots instead of sleep :(

let me see if @mck1117 can ship his alphax board to @NMSTEC maybe @NMSTEC would be better than me at HW hacks

[rusefillc]

@NMSTEC says https://sparkoslabs.com/product/dip-to-soic-op-amp-adapter/ https://sparkoslabs.com/discrete-op-amps/soic-to-dip-adapter/ https://www.sparkfun.com/products/13655

[NMSTEC]

Just makes prototyping easier with less chance of damaging pads or board. Can also run wires to a breadboard for prototyping. Used it often when I needed to constantly pull eeproms off OEM ecus.

[rusefillc]

@andreika-git please add resistor provisions

[andreika-git]

Done

[rusefillc]

See #38