3.5 B3 Toolboard stepper intermittently running backwards

Phaedrux

We are aware of and looking into this issue. I'll create a github issue for it today hopefully.

And many thanks to those doing the excellent testing and reporting here.

tas

@Phaedrux I would like to add some information to this issue which will hopefully be helpful.

I have a simple setup, Ender 5 Pro with Duet 2 Wifi in standalone mode, no toolboards or other attachments. Running 3.5.0-beta.3.

I just finished installing an LGX Lite extruder and custom fan setup. I calibrated esteps and had some strange issues.

From the PanelDue Extrude retracts and Retract extrudes. I thought I had the direction backwards for the LGX Lite but I did not.

Strangely prints seemed to work. However doing more testing this morning I noticed that the purge at the start of a print actually retracts the purge amount. But then the print proceeds normally.

The same thing happens with the Panel Due in that at some point (seems random) the direction corrects itself.

My point is it may be a more general issue than just toolboards.

If you need any testing done I will be calibrating this printer all day so happy to oblige.

ctilley79

@tas That's very interesting. Pretty much rules out issues with the can bus.

dc42

@tas thanks for the additional data points. I expect to investigate this issue tomorrow.

dc42

@tas please provide your config.g file.

Do you have pressure advance or input shaping configured?

dc42

@ctilley79 @tas @Argo I have found a possible cause of this issue. It will indeed give the appearance of randomness because whether it happens will depend on the history of moves done. It is most likely to occur either segmentation is enabled. It is unlikely to occur unless pressure advance is in use, although it might possibly occur on a delta printer even when pressure advance is not used.

Are you able to test a new build of the firmware that should correct this? If so, which boards do you need firmware for?

Exerqtor

@dc42 said in 3.5 B3 Toolboard stepper intermittently running backwards:

@ctilley79 @tas @Argo I have found a possible cause of this issue. It will indeed give the appearance of randomness because whether it happens will depend on the history of moves done. It is most likely to occur either segmentation is enabled. It is unlikely to occur unless pressure advance is in use, although it might possibly occur on a delta printer even when pressure advance is not used.

Are you able to test a new build of the firmware that should correct this? If so, which boards do you need firmware for?

I've had a much similar issue on my Duet 3 @ 1LC and would love to test out a possible fix ✌️

tas

Hi dc42,

I just need firmware for Duet2wifi and I can test immediately.

I do have pressure advance enabled in my config.

dc42

@tas please try the file at https://www.dropbox.com/sh/in0t2j68ic66oa8/AADv_6WzRy07H6MhBOrkDt23a?dl=0. Caution: this is untested because I no longer have a Duet 2 machine to test it on.

@Exerqtor I will test the new tool board firmware on my machine before I make it available.

dc42

@Exerqtor my test print was OK so I have put the new binaries for tool board and 6HC at https://www.dropbox.com/sh/in0t2j68ic66oa8/AADv_6WzRy07H6MhBOrkDt23a?dl=0. The tool board binary should be all you need to fix this issue.

ctilley79

@dc42 I’m using an SBC Duet 3 6HC and Toolboard. Is there a process to update without the package manager?

dc42

@ctilley79 if you are already running 3.5.0-beta.3 then it should be OK to just upload the new files via DWC.

tas

@dc42 Yes that has fixed the problem.

I tried extrude and retract from the paneldue three times rebooting the printer each time. All ok.

I then tried a print and the prime did push out filament. I see no issues with the print so far.

So you have fixed the issue and my Duet2Wifi machine did not blow up.

Many thanks.

Argo

@dc42

I can confirm the issues are gone with Duet 3 Mini 5 + Toolboard 1LC

dc42

@Argo @tas thanks for the feedback! This gives us more confidence to release 3.5.0-beta.4.

ctilley79

@dc42 Worked for me too. Thank you.

dc42

@ctilley79 thanks.

Exerqtor

@dc42
Seems to be working here too!

It's off-topic but happened after moving too b3+ so i'll drop it here and open another thread if it's more than a one line reply to fix it. Now I'm having some issues with the new neopixel assignement, it won't accept M950 E0 C"DO" T2 and returns:

M950 E0 C"DO" T2
Error: M950: Unknown pin name 'do'

This is on the Duet 3 Mini.

dc42

@Exerqtor the configuration of LED strips has changed for 3.5 beta4. See the new E parameter of the M950 command and the updated M150 description in the GCode dictionary.

Exerqtor

@dc42
Yeah I allready have the GCode dictionary for M150 & M950 up trying to figgure out what i'm doing wrong.

The reason i "had" to look into it now is that it threw Error: LED strip #0 has not been configured . And when i try to configure it it won't let me lol. (Or do i have to wait for b4 for it to work again?).