That is an excellent idea, thank you I will go that route. Your explanation was exactly my fear with the board(totally destroyed). To the display cabinet of mistakes it goes.

So we have been building a big printer using Clearpaths for a while now and in the process we have had the misfortune of shorting of our 2 duet boards. The first one is likely complete toast, Looking at the pins that make contact on the semi flat metal surface it was set down on while powered by 24V it likely shorted 24v to 3.3v at the paneldue pins. I imagine this board gets a 10/10 on the "how did you destroy your Duet scale" so I'll call it the 10/10 board.

This brings me to our second board. This one is still functioning, however the x driver not only let the smoke out it let about a 1 inch jet of fire out the top when I accidentally shorted 2 of the stepper driver pins... yeah I know. As far as I can tell the damage was limited to that channel everything else works well as I am just using E1 to control X instead and it is printing beautifully. This is however a dual extruder printer and I would like that channel back. Might it be worth my time to swap the driver from the 10/10 board to the bad X channel? I imagine there may be other components on the X channel that were damaged or possibly the 24v short damaged all of the 10/10 boards drivers too. I am capable of doing the soldering I just don't want to bother if this is a fool's errand.

Excellent, seems to me the sinking method would be the simplest solution. We will give that a try.

We don't have a single channel that tests over 4v.

Have you tried other files? I print alot of the same parts and have separate print files ready for each quantity 1-20 and each file worked fine except the 18 parts file which did this, no idea why to this day but it was only this file.

Otherwise it could be a backlash setting in your config somehow. Or one of the axis is moving farther than the drive thinks it is which I would think with a delta would cause it so gradually drift to one side like that. Recalibrate the steps per mm?

The question is how much current can your drivers handle supplying?

We have swapped cables, motors, driver channels and get the same performance out of each channel regardless of any other physical changes. The performance is also not necessarily repeatable exactly. If we move the axis one direction and then back to zero sometimes it is off by a little or alot.

It seems that the board is not supposed to actually shift the voltage of every channel like it says in the pdf so we will have to accomplish that another way. Hopefully this is the issue, seems to explain it if the 3.5v is only sometimes activating the opto-isolators.

If you read over the first page I linked it seems to suggest that the 5 motor channels are not necessarily level shifted to 5v but the 2 heater channels are. However if you click the link on the page to the documentation which takes you here https://duet3d.dozuki.com/Wiki/Expansion_breakout_board
It says very clearly "all 5 driver channels are level shifted to 5v", did we get a bad board, is this a typo, am I missing something?

This is the breakout board in question https://www.duet3d.com/Expansion_Breakout

I am working on this project with Adam and wanted to chime into the troubleshooting directly here.

My question is does the expansion breakout board boost the signal voltage of each pin to 5v? We were under the impression it did but after looking at this problem we checked it and found that each pin is producing between 3.1-3.6v which may be sporadically triggering the isolators(they have a 5v minimum) resulting in different speeds for each different voltage. dc42's comment about using 5v instead makes me think we were wrong in our thought that this board boosted the signals.