RRF3.4 Dual Z-Axis with 1XD expansion - Height Drift/Mismatch

GHJENG

Hi,

Thanks for developing an amazing product!

I have a 3HC main board connected to RPi4 SBC. I’ve just ‘updated’ and ‘upgraded’ the main firmware, plus all the expansion board firmware to 3.4, via the stable link. The update ran flawlessly on all the boards.

I have the SBC and 3HC connected to four (4) 1XD expansion boards and a 1LC tool board via CANBus. That are physically daisy-chained in this order… X-Y-Z1-Z2… then into the 1LC with termination resistor.

I am using 48V leadshine external stepper drives (the ubiquitous black ones). I have set all the timing to 25;25;25;25 for fault-finding purposes… still the same issue with height drift on Z2. Strangely the Z1 axis doesn’t drift.

The machine I’ve built is a big 2000mm x 1200mm x 1200mm (X-Y-Z) Core-XY printer, with a Hemera hot end and super volcano nozzle.

The CANBus connections in the panel are unsheilded, but twisted pair runs, less than 300mm long. The main CANBus cable to the tool head is an industrial spec profibus cable with terminated shield.

I know you’ll want to see my config file, but I don’t have the machine on WiFi yet and haven’t had a gap to grab a flash drive to save it out. However, everything about the machine works great, other than Z2 progressively drifting lower during G29 bed mapping.

I have checked all mechanical joints and connections of pulleys to shafts, etc. And everything is solid. I have also set the currents high enough and speeds low enough that there are no skipped steps due to under-torque. I read in a previous post that RRF3.3 had a drift issue due to CANBus latency, but that RRF3.4 had sorted this out… so I’m a bit clueless now as to why the drift is still there after the update?

Any help or advice would be much appreciated.

Thanks!

dc42

@ghjeng have you verified using M115 B# (where # is the board CAN address) that all the 1XD boards really are running RRF 3.4.0 ?

Under what conditions does the height drift on Z2 occur? Is it during normal printing; or during bed probing; or something else?

If you swap the Z1 and Z1 motor connectors over, and also change the M584 command in config.g to swap the driver addresses over, is it still Z2 that drifts?

GHJENG

@dc42

Hi Dave, that was a ridiculously fast reply

I have verified each of the boards with M115, yes.

I’m still busy commissioning so haven’t managed to run a first print yet. I’ve got the bed to within 0.5mm corner to corner… so I have only seen the drift during a G29 so far. If I run G29, then look at the height map on the Z2 side of the bed, then run another G29 immediately after, I can see on the subsequent height map that it’s dropped about 0.5mm at Z2. (Verified this with a physical measurement too.)

I haven’t tried swapping the motors and their addresses yet… I can possibly go to the factory to try this a bit later this afternoon. I see what you’re getting at there. I have checked the mechanics quite conclusively, but I’ll do what you suggest in order to rule a line through it.

Anything else that I should be checking at the same time?

Kind regards

dc42

@ghjeng said in RRF3.4 Dual Z-Axis with 1XD expansion - Height Drift/Mismatch:

Anything else that I should be checking at the same time?

Yes. Also try swapping the motor connectors, but using the original M584 command. So the positions of Z1 and Z2 will be swapped. If you are using separate endstop switches for them, you will need to swap the endstop ports; and if you are using true bed levelling you will need to change the order of leadscrew coordinates n your M671 command.

GHJENG

@dc42

Thanks very much. I ran out of time to get back to the factory today but will do all of the relevant swaps tomorrow and feed back on this thread

Excited to resolve this and try to print some large carbon fibre layup tooling