3.5 B3 Toolboard stepper intermittently running backwards

ctilley79

@gloomyandy said in 3.5 B3 Toolboard stepper intermittently running backwards:

@ctilley79 Does the distance make any difference? So for instance if you move just 10mm at F300 does it move in the correct direction?

Ok. The threshold is F254 which makes sense because thats the last integer in an 8 bit number. Extrusion distance does not matter.

oliof

I cannot reproduce this on my Pico based toolboards. I am at 16x microstepping to and 631 microsteps.

Exerqtor

@ctilley79 said in 3.5 B3 Toolboard stepper intermittently running backwards:

@Exerqtor The problem persists even with input shaping and pressure advance disabled.

It might be the case, since i'm short on time atm. I had to revert back to beta2 this time. Hope David has time to crunch this when he gets back to the UK.

ctilley79

@oliof is your pico toolboard running 3.5 beta 3? Does the pico even support RRF? I’m confused.

oliof

@ctilley79 Yes, 3.5-beta.3 on the Duet3 Mini 5+, and 3.5-beta.3 on the RP2040 toolboard, which has experimental support in 3.5b-eta.

ctilley79

@oliof SBC or Standalone?

oliof

@ctilley79 standalone

oliof

My test may have been invalid since I tested with F360; I sae a report that direction corrects itself again at F327 and higher. I am now traveling so wont be able to test before Sunday.

ctilley79

@argo I'm experiencing similar behavior here

Argo

@ctilley79

Yes, I also had this issue and the extruder was overall extruding a bit too much filament whilst printing. My normally calibrated extrusion multiplier set in the slicer would not work anymore.

What helped:

Duet 3 Mini5+ reset + Toolboard 1.2 reset (via button)
formatted the SD card and copied just the config files
flashed the Duet 3 Mini 5+ directly over the boot loader as a mounted drive to a PC and then flashed the Toolboard over the web interface

Now the reverse running extruder issue is gone and also my set extrusion multiplier works again.

I don't know what caused the issue tbh. Maybe the countless flashings over the web interface with all the beta versions (I also flashed Klipper to it fort testing purpose) causes issues over time but that's just a wild guess on my side.

Exerqtor

@Argo that's interesting, maybe i'll give this problem child another go with a "clean instal" 🤔

ctilley79

After more testing I've found that the threshold isn't always consistent. I thought that I was onto something but it was just chance that F254 was the threshold. At some point the threshold changed again.

One thing to note that when I downgraded back to stable 3.4.5, I had to invert the motor direction one more time, but I no longer had the problem. I most likely inverted the motor during initial troubleshooting while on 3.5b3 and lost track of changes I made. This may indicate that at low speeds the motor ran backwards not the other way around. Good luck with this one @dc42.

oliof

Trying to actually print something I am running into this issue as well, just not entirely clear what the thresholds are as testing with F254 on a freshly started machine works for me. But a gcode file I have is actually starting to stall at some time early in the print job.

Argo

@Exerqtor

It seems it only temporarily fixed the issue. I just startet my printer for the first time since I was able to successfully print a Voron test cube and when inserting filament via PanelDue the extruder runs backwards at 5mm/s and forward at 10 mm/s when commanding it to extrude. With other words the bug came back

Exerqtor

@Argo
Ah ok, lets hope @dc42 has time to look into this soon.

oliof

and another 12h print completed without issue. Really weird.

Exerqtor

@oliof
Yeah like i said in my thread, it seems totally sporadic.

oliof

@Exerqtor well, I have one gcode file now that seems to stop at a certain point to just extrude, and another that runs for 12 hours. On monday I will have a couple hours to try and find a reproducer.

Out of interest, are you using slicer or firmware retract?

Exerqtor

@oliof
I'm using FW ✌

mule

I am having the same issue withe B3: When trying to print a test for measuring pressure advance the filament gets retracted so far that it even leaves the extruder gears after printing some lines. It is always happening at the same position when trying to print the following gcode:pa_pattern.gcode