Duet 3 mini5+ nightly restarts 3.5.2

oliof

@Chriss does it also happen when simulating a print?

Chriss

@oliof No idea to be honest. I saw it only when the printer was actually printing. (I grounded my hotend and my e-stepper some month ago for a other problem) I have only one printer with this firmware and I need it at the moment to finish some projects. And the problem is not easy to reproduce. It seems to me like it was wore at the beginning, like every morning a rebooted printer, but this is not the case anymore. I guess that this was a false observation but is not that I experiance a reboot at every print or every 2nd print.

I have the uneducated feeling that the failure happens only on higher prints too. I printed many pars which were below 100mm in Z. And two higher parts and the 2nd failed.

Not saying that any of them are causing the issue, I'm listing my observations only.

@droftarts Is there any kind of debug firmware which we can run to log the data to the serial interface? I would be more than happy to test more and capture as much as I can? This topic is getting older and older and I have the impression that we make no progress at all.
I have 4 of this boards currently, 1 or two of them should be from the first batch, the others are a bit younger. I'm not sure which one is in which printer. Would it make sense to check that? (I guess that my bench board is the youngest one.) I can place the bench board into the working printer if this make any sene? Just to see whether it happens in a real printer with the bench board not. The bench board looks like at so far:

=== Platform ===
Last reset 1667:38:29 ago, cause: power up
Last software reset details not available
Error status: 0x00

I have plenty of steppers and hotends and all the stuff in my lab too. I can build a "mock printer" with this parts and the board which runs into the issue if needed. Just ask for it.

droftarts

@Chriss I asked @dc42 if he'd made any progress on this issue. He said:

My previous investigations suggested that the crashing was caused by a hardware error in the MCU when a DMA transfer finished and the SPI peripheral was disabled. So not a software bug but a hardware defect that I need to find a workaround for.
It's possible that the RRF 3.6 builds are not affected by this issue, so worth getting the affected users to try 3.6.0-alpha.4 at https://www.dropbox.com/scl/fo/cckwiq91gn16hvl1zdjnp/AF0SMEtkVfiArSPeYaBDGPY?rlkey=kqkknk9q1kiq684u4s55ce8d4&dl=0

Could you try the 3.6.0-alpha.4 build?

Ian

Chriss

@droftarts said in Duet 3 mini5+ nightly restarts 3.5.2:

Could you try the 3.6.0-alpha.4 build?

Your wish is my command, but not before mid of the week. I'm installing a new hardware feature at the moment. I will do the upgrade asap. (Any new traps you want to tell me?)

Cheers, Chriss

droftarts

@Chriss There's a few things not fully working, check the release notes: https://github.com/Duet3D/RepRapFirmware/wiki/Changelog-RRF-3.x-Beta

Ian

Chriss

@droftarts said in Duet 3 mini5+ nightly restarts 3.5.2:

@Chriss There's a few things not fully working, check the release notes: https://github.com/Duet3D/RepRapFirmware/wiki/Changelog-RRF-3.x-Beta

I just upgraded the printer, including the DWC, some error message popped up in the DWC that some versions are not compatible. Does the DWC version look OK to you?

Cheers, Chriss

jay_s_uk

@Chriss it's the mismatch between DWC (alpha2) and the main firmware (alpha4).
I'm not sure there has been an alpha 4 release of DWC.
You can ignore the error as there's nothing breaking between those version mismatches

Chriss

@droftarts I used the same object I always had a problem with and it finished the first time successfully.
I'm not sure how much this will tell us, a lucky punch maybe? Problem solved?
The only problem I see now is the quality:

I'm not sure where this came from. I do not see that on objects which are not as high as this object is. What shall I do now? Bring back the old version and be unable to print objects which this dimension? Or drop the bench board in and go back to the latest stable version and test again to see whether the quality is better than?

Cheers, Chriss

droftarts

@Chriss You'll need to retune Pressure Advance with the 3.6 release. It should work much better!

Ian

Chriss

@droftarts I saw that in the release notes. My PA dial in process came out with very much the same value as it was before.
I will perform the test again with a material I know better, this grey filament (ABS+) is new for me, the temperature may not be ideal for this material too. Maybe to much rush this time.
But I'm confused how the gab at the seams piled up over the layer hight, I never experienced that before.

Phaedrux

@Chriss said in Duet 3 mini5+ nightly restarts 3.5.2:

But I'm confused how the gab at the seams piled up over the layer hight, I never experienced that before.

That is common when there is under extrusion on unretract at layer change.

Chriss

This post is deleted!

Chriss

@Phaedrux

The top line is without any PA. I have the feeling that I do not need any PA at all. Can that be possible with the alpha?

Or I print to slow.

Cheers, Chriss

Chriss

That is the result with no PA at all.

oliof

@Chriss what are your retraction values?

Chriss

@oliof I have to admit that I do not know that anymore at this print. I shift here some profiles in SuSi back and forth. The current print with 5.8mm and 35 and 25mm/s look far better.
I know that 5.8mm are far to much for a direct dive, but I feel a bit helpless so I just tried the Voron 2.4 standard profile and some others.

oliof

@Chriss try 2mm @40mm/sec, if that's not better, give it a quarter to a half millimeter extra unretract.

Chriss

@oliof I do not why but the "avoid crossing perimeter = true" solved it so far with the standard profile. Only the first layer looks a bit strange now:
file:///home/chriss/Downloads/3.6.0-alpha.4/Pics/IMG_0612.JPG
I can not stop thinking that there is something strange ongoing with PA, which is off with my setup at the moment. The top side is overextruded at the touchpoints with the walls. That is more than strange with ABS+ and a EM of 0.88.

Chriss

@droftarts Back to topic..... I tuned a bit and solved my other problem with the missing extrusion. The printer just finished the 2nd 160mm print. So it is a 100% success rate now, I would not put a guarantee on it but it seems to me that the current alpha solved the reboots. But please to not count me on that. I did two prints only and I can not do more of them for now. The prints still look horrible in may ways, I need to tune the printer a bit further before I can do more tests. This thing eats to much filament at the moment.

Cheers, Chriss

Chriss

@droftarts

I'm afraid that I have to move back to the stable FW. I have soo many unexpected behaviours with the alpha since I print with it. I can not stand than anymore. The last pei sheet I sacrificed was to much. I saw steppers tuning into the wrong direction for some mm which ended in a bend gantry etc. Sorry, but I'm out of this testing, for a while now. My printer has to finish some prints and I can live without the hight but not with this strange behaviours I have seen in the last days.

I think that we can come to the conclusion that it is very likely that the next release will fix the reboot problems at least on my end. I'm not fully satisfied with this conclusion but it is how it is for now.

Cheers, Chriss