Odd vertical lines showing up

AK_Eric

New to Repetier, but I'm familiar with other firmwares like Marlin and Sailfish. I've recently changed the electronics on my custom core-xy bot from Rumba\DRV8825s\Marlin to RADDS\SD6128s\Repetier.

The old & new stepper drivers are both set to 1/32 microstepping.

I'm seeing new artifacts in my prints that didn't show up with the Rumba\Marlin combo: Tiny 'steps' in my layers, that when lined up form vertical bands. Measuring them, there's 3 per mm. I have two pictures showing this below.

I've done additional tests printing cubes lined up with the build platform, and rotated 45 deg on Z: The issues only show up on the 'lined up' cubes, not the '45 deg cubes': Which makes me think this is an issue with core-xy: When at 45 deg, only one stepper is driving the motion, and while 'lined up with the bed', both steppers are firing to drive the head straight.

Any thoughts as to what could be causing this? Thanks

ekaggrat

did you find a solution . i am also having the same problem on my corexy with ramps and sd6128

AK_Eric

Nope. No replies here, or anywhere else I asked. They're still showing up :(

Repetier

Please check the fast core xy implementation of repetier, see

https://github.com/repetier/Repetier-Firmware/issues/466 </div><div>

15. Feb onwards you find explainations on how to use. It is a different approach to satisfy core xy that leads to faster handling and is similar to math done by marlin.

AK_Eric

Thanks, I'll give that a shot. Feedback appreciated!

AK_Eric

I did a complete download\install of the latest firmware 92.9, added this line to my configuration.h:

#define FAST_COREXYZ

Halved my steps per mm on X&Y, and re-uploaded...

Right now, everything is traveling twice as far as it should. As if I shouldn't have halved my steps per mm.

Thoughts?

AK_Eric

Ok, my eeprom wasn't being updated, that's the issue with the steps per mm, got that fixed.

Ok, reprinted with those updated settings and.... same exact lines show up. Was worth a shot.

Also, whenever it homes or moves the nozzle to start the print, it's really jerky. Once the print starts it's ok.

ekaggrat

my lines are worse.. kind of stuck here..

i have the sd6128 running on the ramps.. i even have a4988 and drv8825.. the issue stays with all

AK_Eric

@ekaggrat : That looks like a different problem from what I'm seeing. My lines are super tiny and very regular. Yours almost look like problem with the mechanics of your machine, if I could theorize about it. The screen-left vertical edge of the R has regular waves as it goes up Z, like the leadscrews are tweaking the position of your bed or something. Just a guess.

Repetier

@ekaggrat I agree with AK_Eric regarding z wobble. What you also see is waves from vibration moving around corners. That is quite normal and can be reduced with good tension and not too high accelerations.

@AK_Eric your problem looks like you see every step. Could be that the motor do not lock into substeps. Have you tried increasing resolution and current? If the current is not high enough you do not get full substeps. And since it depends on 2 motors running in sync it could be one or the other falling into near full step position.

The jerks you see are normal and a result of not full planned paths, so it speeds down between 2 nonlinear moves. Try updates per second set to 10. You do not need so high values like for deltas as moves are not nonlinear. As you already saw it disappears once buffers are filled.

AK_Eric

Background: They're the same steppers I was using with my RUMBA board (with Marlin), with DRV8825's set to 1/32 microstepping:

http://openbuildspartstore.com/nema-17-stepper-motor/</div><div>These lines\dots never showed up in that config. The steppers are the one thing I'd expect to not be causing this. Could be wrong though. I used the same 1/32 microstepping on the new SD6128's to try to keep as much in parity.

I've checked with my meter, and both steppers are getting the exact same voltage applied. When I first did the new RADDS\SD6128\Repeteir setup, I had their voltage higher, but they'd get molten hot to the touch, so I dialed them back to a less "house-fiery" temp ;) but I should be able to put a bit more juice back into them without fear: I can try pushing that up a little though an seeing if the behavior changes.

Appreciate the troubleshooting!

AK_Eric

Changed the ref voltage from .58 a piece up to .75v: Printed again, exact same results.

Repetier

From raps128 docs:

Recommended voltage range: 0.8 to 1.6 V. This corresponds to a theoretical peak engine power from 0.73 to 1.46 A - See more at: http://doku.radds.org/dokumentation/raps128/#sthash.wp0ep6W0.dpuf </div><div>

Voltages are not the same as for DRV8825. 1.6V is for 2.2A motors. So you are running on the weak side. They get quite hot easily. I put a fan on the coolers to keep them cool.

AK_Eric

When I first got them I'd set them to .9v: This caused them to nearly turn to molten slag. But like you said maybe they need to be cooled down in the process. I'll try some more tests, thanks.

AK_Eric

So I tried two more tests: .9v, and just over 1.0v (while holding two fans blowing directly at them) : In both instances, the behavior did not change: Exact same lines\dots popping up in no greater or lesser quantities. The issue thus appears to be independent to the voltage going to the steppers.

Repetier

Your pattern looks bigge rthen only a substep and even if firmware would lag one substep I doublt it would look like this. And the RAPS128 are normally known to be better. Anyway have you tried with DRV8825 on RADDS? Just set STEPPER_HIGH_DELAY and double step delay 1 for these or you will loose steps.

AK_Eric

Sounds like a good next step: To do that, I'll have to completely unwire my board to remove it from the case since all the dip switches I'd need to update to support this are on the back side of the board. Sigh, wish they would have put them under the drivers a-la the Rumba instead of a completely inaccessible place.

Not something I want to tackle right now, but if I ever get around to it I'll post my findings.

Repetier

Aren't 1/32 steps the same switch setting?

AK_Eric

Unfortunately not. For 1/32 microsteps:

The SD1628 the dips are on,off,on.

The DRV8825's, it's on,on,on

http://doku.radds.org/dokumentation/radds/ </div><div>So I have to basically take everything apart to access that. Not impossible, but not something I want to deal with right now.

While the flaws bug me, they are really small and hard to notice.

AK_Eric

To follow up: I installed RepRap Firmware on my bot: Same exact lines are showing up with that firmware.... so this isn't a firmware issue. At this point, I'm leaning towards it being something with the RADDS board itself, or the SD6128 stepper drivers. Next test will be to change from 1/32 microstepping to 1/16th, and see if that fixes anything.