Consistent Y-Axis Drift can tell you what it ain't can't tell you what it is!!!

OK,

I have a very large (18" X 18" x 18") print volume printer. Custom built, RepRap based.

I see shockingly consistent drift on the Y axis. What I have tried so far:

1) Tighten all the mechanical components and clean, including and especially the screw backlash arrangement. All axes are driven by mechanical screws posts, there are no belts on my printer, so it can't be belts.
2) Tried driving the printer faster, tried driving the printer slower, can't be speed
3) Tried adjusting the voltage on pololu 4988 motor drivers up until the driver was super hot to the touch, can be too little current
4) Tried adjusting the voltage on the motor drivers down until the stepper started skipping, can't be too much current
5) Tried putting an acceleration constraint on printer movements, can't be acceleration
6) Tried hand writing custom G-Code to feed the the printer with little / no extrusion, the thing still drifts!!!! Can't be hitting bumps of extruded material
7) Tried using different parts of the printer, different parts of the bed etc. Can't be the specific part of the bed. Can't be bed related.
8) Tried moving the bed quickly back and forth with custom G-Code, did not get much drift on short back and forth, did get lots of drift on long back and forth - it can't be acceleration related.

I wrote custom G code that would extrude a line of PLA then move back and forth on the troublesome Y axis 50 times then print another line of PLA so that I could see how much drift there is. The drift remains completely consistent through all the above adjustments. It never changed, I always get the same amount of drift. The funny thing is that it clearly is related exclusively the the amount of travel of the Y axis. It is **not** acceleration related. With the custom G code I can move the bed back and forth by 10mm 50 times and I get much less drift than if I move the bed back and forth 90mm 50 times. If I move the bed back and forth 200mm 50 times I get exactly twice as much drift as when I move it back and forth 100mm 50 times. The drift is always in the positive Y axis and none of the above things change the amount of drift I get. It looks like there is always 4mm of drift for every 100 X 2 x 50 = 10,000 mm of linear travel or 0.04%

It almost feels like its a rounding error in the arduino software or a software bug somewhere.

Any ideas on what I could try next. I am kinda out.

David

Hi David,

did you fasten the pulley on the motor shaft?

Klaus

It sure sounds wierd..

Have you tried switching the connections to the X and Y motors? Remember to also switch the endstops..

Edit: What about jerk settings? There is only one parameter for X and Y, so in theroy you should see drift on both axis, but it may be worth a try to lower it a bit.

Edited 1 time(s). Last edit at 03/16/2014 01:05PM by Ralf.

It's hard to deduce the cause without some pics of your setup.

Do you get drift when you turn your acceleration down to 1000, your jerk to 5 and print at 10mm? If it goes away start increasing your settings incrimentally until it appears.

---

WWW.ZATOPA.COM - Your Place for high quality 3D Printing Filament and accessories

Flash new firmware. I have heard of firmware causing consistent drift. I use Marlin, so I can recommend it. You could also try replacing the driver that is driving the drifting axis.

Quote
kkberg
Hi David,

did you fasten the pulley on the motor shaft?

Klaus

+1 on KKberg comment.
I had consistent Y shift and it was a loose grub screw on the flat portion of the motor shaft. It introduced a repeatable amount of 'play' on the pulley. It was noticeable in both directions.

what firmware are you using? im using repetier host on a 20x20x20 mendel style with no problems. I'm using pololu drivers (not stepstick) on high torque nema 17s.