Optical endstops can provide very high precision homing: https://drmrehorst.blogspot.com/2020/03/testing-ummds-xy-optical-endstops.html

@oliof said in Use accelerometer to measure backlash:

whether it would be possible to measure backlash with an accelerometer

I wonder if any of these approaches would be practical for the tiny backlash we expect to find:

Move in one direction and stop. Then move a small distance in the other direction and see if it registered an acceleration bump.

Compare the accelerations (second integral?) in these two scenarios:
a. Move and stop. Then move distance x in the same direction.
b. Move and stop. Them move distance x in reversed direction.

Move back and force in small direction x. Increase x until some acceleration is registered.

I'll add this to the firmware wish list. What is the preferred communication method between the filament width measuring system and the Duet: async serial, or something else?

I don't think there's anything to worry about. If and when MCU temp probe calibration becomes available for expansions it would be an easy fix.

@dc42 some manuals how to calibrate S value? I have best results with M309 S0.03 (lowest temperature fluctuation by extruding the same amount of filament (PETG 235 degrees) on same speed G1 E60 F200).

@fcwilt I don't know. There is underlying housekeeping that the firmware must be doing. The drive mapping could be changed but I don't know whether that would be sufficient.

@owend, it may be helpful in the use case below, splinting the terms and converting them to numbers

https://forum.duet3d.com/topic/27201/macro-parameter-letters

@diamondback

I was not happy with the limitations of homing with a probe or leveling the bed.

I solved it all with multiple Z endstops. Homing is as fast as your printer allows and the result is good enough the final leveling with a probe always works without any change in settings.

Here is a video showing how it works in practice:

Multiple Z Endstops

Frederick

@ccs86

Sounds like a good idea.

In the meantime you can edit those points in the CSV file as needed. I've done that in the past when it was just one point that was way off.

Frederick

@konk you can use triggers on a GPIO pin to run a macro. Think its M581 to define the pin as a trigger. This page details how to do it:
https://duet3d.dozuki.com/Wiki/Using_M581_-_External_Triggers_and_Building_a_Control_Panel

@dc42 You know better than I do. I thought the processor in the PanelDue and the amount of RAM available made implementing sorting there impractical

@cncmodeller That's a good tip, I have actually resorted to something similar!
I created a "M2001.g" file in the sys folder and added a M2001 command on each layer change...

The issue is that I am using a custom slicer that I can't easily change myself and that doesn't have this capability (that at least adds ";LAYER: ##" markers at the beginning of a new layer). I was hoping to not have to manually change the gcode file, but thank god for the "change all occurrences" function in VSCode 🙂

@jay_s_uk Ah perfect, thanks!

@phaedrux Thanks. Perhaps we can hope that this feature will eventually come in the majority of slicers.

I noted that this had also been proposed for Klipper, but never implemented.

I suppose you could remap the toolfan to a new fan number and control it manually. The slicer fan commands wouldn't do anything anymore. But that would only get you a fixed manually set fan speed, not one that scales with slicer commands.

DWC used to have (still does maybe?) a pin next to the fan speed slider for the tool that would let you fix the speed, but it was only active while the DWC window was open. I'm not sure if it's still there or not. I rarely look at DWC at all in my workflow.

@owend there isn't a public one yet. We had planned to move issues and feature requests to github by now, but for various reasons that hasn't happened yet. This is a list of the main features we currently plan for RRF 3.5:

Combine Move and DDA debug in M111 using the bit map
Support for displaying thumbnail images (if not completed in 3.4)
Support two concurrent GCode streams
Axis backlash compensation
M571 when using a remote extruder
Support stall homing on expansion boards
M291 support multiple choice
Support input shaping on EXP3HC
Support input shaping on EXP1XD if possible (may hit performance issues)
Support input shaping on EXP1HCL
Support filament runout sensors (not filament monitors) attached to main board for extruders on tool/expansion boards

Other feature requests will be considered too, such as array-values variables.