CAN Failure after bootloader update

dc42

This is an interesting one! It could be either the bootloader or the firmware not starting.

If you temporarily downgrade the 3HC firmware to 3.1.1, does it start up and flash the LED with the tool board connected?
If it still doesn't start up with the tool board connected, then I suggest you downgrade the bootloader to version 1.3. When that is working and booting the firmware, put firmware 3.2.2 back on the 3HC.

psyblue

As a result, The worst has happened. All three HC3 boards have been ICE-state(red diagnostic LED off, No response).
No response for M122 B1. No response to the reset button

I tested several things using the HC3 board I had.
RRF3.1 + Bootloader1.3 (estimated, No ID on diagnostics) -> good
RRF3.2.2 + Bootloader1.3 -> good
RRF3.2.2 + Bootloader2.2 -> Abnormal
RRF3.2.2 + Bootloader1.3 Down grade -> HC3 ICE-state

No problem on 1LC board.

I need the ATMEL ICE tool.

dc42

Please try the following:

Upload EXP3HC firmware version 3.1.1 to the Duet
Do a factory reset of one of the 3HC boards (set all 4 switches off and then turn on the power).

psyblue

@dc42

Dear dc42,

First of all, thank you very much for the always quick and kind reply.

I've tried following your advice, One of three boards could be made with firmware v3.1.1 + Bootloader v1.3.

I repeated the test with this.

I thought it was OK with firmware v3.1.1 + Bootloader v1.3, but there was a problem.
When 3HC is disconnected from every cable (include CAN) and only power is connected, does not the DIAG red LED flash at all, just off state.
Repeated plugging in and unplugging the CAN cable, the red LED starts blinking.

The other 2 3HC are still ICE"ICE-state" (I don't know what to call it in English. I call it'ICE'. power-on, no response, no blinking, do nothing state)
Disconnected the 3HC board from CAN and set the address to zero (all switches off), does not the DIAG red LED flash at all, just off state.
The 3HC is not connected to CAN in any way.

In my experience, I suspect the bootloader 2.2. the bootloader procedure seems to be stuck in the CAN init loop.
Even when it worked normally in the past, sometimes (30% probability) there was a situation where it was suspected that booting was delayed.
Because I found a case where the contents of config.g were not passed to 3HC.
So, I put "G4 S2" (delay 2 seconds) at the beginning of config.g.

I hope you try to reproduce the same situation as me.

And 1LC looks fine.

dc42

@psyblue, which country are you in, and which reseller did you buy the boards from?

psyblue

@dc42

I am Korean in Korea, bought at filastruder.com

Isn't it a known problem?
It could be an important source of potential bugs in the future.
I can send you one, if you want. Of course, I want to get it back with the improved one.

However, I have anxiety that seems to work normally in your hand as always

Or, I ordered ATMEL ICE. I will reprogram it and report the result.
It will take about 2 weeks.

And short question,
Is there any changed or added pull-up/ pull-down resistor on 3HC board?

psyblue

@dc42

All of the above symptoms are considered a problem with Bootloader 2.2.

After updating to Bootloader 2.3 via ATMEL-ICE, everything works fine.

Thank you for the bootloader 2.3.

dc42

@psyblue, I'm glad you solved it with bootloader 2.3. I'm sorry that it was necessary for you to get an ICE to fix it.

psyblue

@dc42, I am also glad that the problem disappeared with the bootloader 2.3, but I am curious if the above symptoms match the potential problem with Bootloader 2.2

dc42

@psyblue the main change between bootloader 2.2 and 2.3 is that in 2.2 if it didn't get a satisfactory frequency reading within a certain time, its reset the processor to start again. In 2.3 it just keeps on taking readings, in effect allowing the crystal oscillator as long as it needs to stabilise.