Babystepping and 12864 goes crazy

izeman

@pkos sorry. Not sure what you mean?!
Do you want to see code? Or photo of the screen? Or video of me altering the value?

pkos

@izeman In order for the screen to work, you have to first configure it with M918 gcode.
I'd like to see how you have the parameters set for it

izeman

@pkos The screen is working fine. And babystepping is working fine as well - AS LONG as the printer is not in state PRINTING.

This is the output of M918:

M918
Direct connect display: 128x64 mono graphics with ST7567 controller, 2.00MHz, 2 encoder pulses per click

pkos

@izeman Yes, I asked for your specific line configuring the screen in config.g

Reading from the output - change encoder pulses per click from 2 to 4. This should help you, although it will take more rotations of the wheel to make it work properly - not sure if this is something for @dc42 to take a look at - but I had the exact same issue as you did on two different screens and two different duet 3 mini 5+'s (one wifi, one ethernet).

izeman

@pkos Ok. Got it. It's "M918 P2 E2"

I will try, but I highly doubt that it will help. As mentioned: I workes perfectly fine as long as the printer is not printing. I can go inside the bedleveling "menu" and alter the nozzle just as you would expect it to work.

But as soon as it's printing, and this is when you really need the feature as it's laying down the first layer and you want to adjust the squish: You go to babystepping, it shows 0.00, and you turn the encoder BY A SINGLE CLICK and the nozzle moves like 5(!!)mm not 0.01mm as set in configuration. So like 50(!) times as much as it should move. Turning it on step back makes it move like 2mm, 3mm, 5mm ... whatever. It's not consistent.
So completely unusable.

Using the webfrontend and clicking the icons to alter the BS works perfectly fine all the time of course.

pkos

@izeman I know. I am saying I had exactly the same thing and messed up a couple PEI sheets because of it The change from 2 to 4 fixed it for me.

izeman

@pkos said in Babystepping and 12864 goes crazy:

@izeman I know. I am saying I had exactly the same thing and messed up a couple PEI sheets because of it The change from 2 to 4 fixed it for me.

Good to note that I'm not alone. I destroyed THREE perfectly fine PEI sheets by playing with babystepping. Not idea why I refused to learn ....
I now ordered a FR4 epoxy plate. Just need to find a way to attach it properly to a stainless steel plate.

dc42

@izeman which firmware version are you using?

izeman

@dc42 said in Babystepping and 12864 goes crazy:

@izeman which firmware version are you using?

RepRapFirmware for Duet 2 WiFi/Ethernet 3.4.0beta6 (2021-11-06)

dc42

@izeman please check whether 3.4.0rc1 has the same issue. You can print several mm above the bed to avoid risk of damage. I recall doing a fix to the 12864 but I can't remember the details.

izeman

@dc42 Sure. Will do so when back home. Thanks

dc42

@izeman also, please check that the cables to the 12864 display don't go anywhere near any stepper motor wires. One possibility is that it is interference to the signals from the rotary encoder that is causing the problem.

I will also run some tests here.

EDIT: I've tested in on a simple non-extruding print job on a Duet Maestro, and babystepping is working correctly for me. The menu entry I am using is:

; babystepping
text V4 R52 C0 "Baby step "
alter V4 N521 D2

izeman

@dc42 is there a difference between:

alter C88 D2 N521
alter V4 N521 D2

Well apart from the obvious positioning I guess it shouldn't make a difference if D2 comes before or after N521 ?!

dc42

@izeman correct, it should not make a difference.

Interference from stepper motor wires seems a likely cause to me.

pkos

@dc42 That can't be it. In the case of two of my printers - ribbons going to the screens are nowhere near any steppers or any other wiring for that matter.

izeman

@pkos @dc42: I agree. I did some testing now, and as you @pkos said: Setting "M918 P2 E4" solved the issue. Somehow at least.
Yes, it now need two clicks to move the "cursor" one step further, and it needs two clicks to make the BS value change - and it changes by 2 steps btw (!!), so if you got BS step set to 0.01, the smallest value for up/down is now 0.02(!!).
BUT: It is not behaving eratically anymore. The value is controllable just fine, printing or non-printing!

I tried this with firmware from November as posted before, and I tested with RC1. Both behave exactly the same.

And as it's working with E4 I also highly doubt it to be some interference.

gloomyandy

@izeman @izeman When I was doing some testing of these displays I found it was relatively easy to end up with the rotary switch positioned "bewrween" two of the steps which seemed to result with some odd combination of the switch output that resulted in constant change of the value. I wonder if you might be seeing this?

izeman

@gloomyandy said in Babystepping and 12864 goes crazy:

@izeman @izeman When I was doing some testing of these displays I found it was relatively easy to end up with the rotary switch positioned "bewrween" two of the steps which seemed to result with some odd combination of the switch output that resulted in constant change of the value. I wonder if you might be seeing this?

I thought so too, but it's working perfectly fine for EVERYTHING else. And once you set it to E4 it works as expected as well. But of course I have no idea how the encoder and debouncing is handled by the firmware ...