Layer shift mid layer

Need a nudge in the right direction to figure this out. The last few prints I've started are begining well, but either after a few layers, the next layer is about 10mm off. Or sometimes within the same print it changes. I first thought it was something wierd going on with the slicing of the models, but I doubt it at this point. After getting the same results from a few different prints, I'm thinking its more mechanical. Last night I noticed that its always in the same direction, along the Y Axis. Thought maybe I needed to tighten the YBelt. I cant seem to get it as tight as a drum, but its moderately taught I suppose. This morning I was hoping I'd be back in business, but same results again. I was smart enough to take a picture of this one. This is first layer. Printed perimeters, started left hand fill and when it moved the head to the right hand side, it mysterisouly oriented about 10 mm off in the Y. X still looks to be spot on.

So now what?

Thanks!
JW

Couple of things I have run into giving Y shift

- pulley occasionally slipping on stepper shaft.
I eventually printed a 3-grub-screw pulley.

- marginal Y stepper current, so it appears to run OK most of the time, but a chance combination of resistance (maybe combo of tight or dry spot on rods, small eccentricity in pulley) causes it to skip.
Increase stepper current.
(I do a simple / brutal test - grab the bed, increase the current until it doesn't skip on moderately firm resistance.)

- pololus heating up and briefly hiccuping. - Too *much* current -
( they're like Goldilocks & the 3 bears - they like it just right )

nb99 Wrote:
-------------------------------------------------------

I have yet to put grub screws into the pulley. Although I should. I had to gently tap the pulley on with a hammer when I assembled, so I dont think the pulley is slipping. Although I've added a sharpie mark across the end to see if indexing changes.
> - pulley occasionally slipping on stepper shaft.
>
> I eventually printed a 3-grub-screw pulley.

I lubed the rods last night with some WD40. "Eccentricities" seem to be inherent in the entire model, right?
> - marginal Y stepper current, so it appears to
> run OK most of the time, but a chance combination
> of resistance (maybe combo of tight or dry spot on
> rods, small eccentricity in pulley) causes it to
> skip.
> Increase stepper current.
> (I do a simple / brutal test - grab the bed,
> increase the current until it doesn't skip on
> moderately firm resistance.)

The pololus themselves never seem to get past cool. I've got heatsinks on them, the motors however start to get screaming hot if I crank it much past where they are now. I havent put a probe on the motors, but from all I've read, they should be getting plenty of current. Although I completely agree with the goldilocks.
> - pololus heating up and briefly hiccuping. -
> Too *much* current -
> ( they're like Goldilocks & the 3 bears - they
> like it just right )


At this point I've had the most luck with dialing back acceleration in Marlin. Seems to have made a good difference. I also went back to some previously printed GCode and all went well. I have yet to get the printer back up to the speeds I was trying to print at earlier when it started failing, but I'm hopeful that the change will make the difference.

Anybody find the sweet spot for RAMPS and Marlin acceleration numbers?


Now if I could get the %^#%^ Z coupler to stay put! (Right now I'm trying to eek out a new coupler)

Thanks for the insight!

---

Josh

//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
//Prusa / RAMPS_1.4 & ATMega_2560 / Slic3r > Pronterface > Marlin / Luck
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

So I was rooting through GCode that I was printing with and it would appear that the biggest contributing factor was I somewhere along the line decided to start slicing things with a 180mm/s travel speed. My little guy cant move that quick that fast so it was skipping steps on the axis. Just so happened that most of the really long travel moves were in the Y axis.

Final answer: Slowed down acceleration in Marlin, bumped the amps on the motors (tiny little bit), dont slice at stupid settings.

---

Josh

//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
//Prusa / RAMPS_1.4 & ATMega_2560 / Slic3r > Pronterface > Marlin / Luck
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

I got the same problem.
I had read this issue and tried everything I can do. Y Axis is shifting

My machine: Mendel-A buy from Lulzbot.com, works fine
My machine: Mendel-B. I assembly it myself. Parts are:
Stepper driver: A4983
Board: Sanguinololu V1.3A
Firmware: Sprinter V1.3.11T


one of my checking can proof the problem came from firmware.
I exchange the motor wire of X and Y.then the shift is show on Logic Y axis, the physic X axis.
At the end, I think this is the Sprinter firmware bug.

anybody have met this problem and solved it at the end? I am wishing help.

Hi All,

Just the other day I had a print suddenly skew ~10mm in the +X direction for just one layer midprint. It turns out that my X axis was binding at higher travel speeds (130mm/s on my prusa setup) occasionally. I hadn't seen this previously because I had always run at ~90mm/s.

Might be worth checking? It's quite simply to test - I simply incremented the speed in pronterface in 1000mm/min increments until it was really evident - and then to fix I loosened one of the x-axis smooth rods, which seems to have done the trick so far. When I get the photos off my camera I'll upload a photo and description to the Art of Print Failure pool.

Cheers,
Gary

---

------------------------------------------
garyhodgson.com/reprap | reprap.development-tracker.info | thingtracker.net

Ultimately my solution was to reduce acceleration in the firmware and not slice a model at exceptionally high speeds.

#define DEFAULT_MAX_ACCELERATION {900,900,80,10000}


The motors were missing steps and causing the shift. I was able to narrow it down and saw it happening at the end of one layer, when the non print move was allowed to move at a faster than normal printing speed as it moved to the starting point of the next layer.

The answer is the most common of all problems with these printers; slow it down!

HTH
JW

---

Josh

//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
//Prusa / RAMPS_1.4 & ATMega_2560 / Slic3r > Pronterface > Marlin / Luck
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

mydogjustice Wrote:
> I lubed the rods last night with some WD40.


WD40 is a cleaner, not a lube. Try using white lithium grease as a lube after cleaning the rods with WD40.

-Rob A>

RobA Wrote:
-------------------------------------------------------

> WD40 is a cleaner, not a lube.

[www.wd40.com]

---

Josh

//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
//Prusa / RAMPS_1.4 & ATMega_2560 / Slic3r > Pronterface > Marlin / Luck
//++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

WD-40 is classed as a "light lubricant", and is great on small items like hinges, locks, and toys.

On heavy or fast-moving parts, the lubricating power of WD-40 is minimal. WD-40 will actually strip away any existing lubricant and leave your parts dry– metal on metal. (The linked page at WD-40 states "WD-40 gets under dirt, grime and grease to clean" which means it degreases!)

[www.popularmechanics.com]

from [diy.stackexchange.com]

Quote

1. Firearms - you should never use WD-40 to clean or maintain firearms. It's hygroscopic and will attract moisture to the firearm which will result in rust.
2. Drive chains of any type - because WD-40 isn't a lubricant it really will not work well on Drive chains of any type.
3. Gears of any type - WD-40 doesn't have enough lubrication affect to be useful on any type of gearing. only specified gear lubricants should be used.
4. Locks. Use powdered graphite or silicon spray instead.

-Rob A.

Thank you mydogjustice

1) I reduced acceleration to [200,200,50,300], but the shift still on Y-axis only. works fine on X-axis.
2) I swapped the wire of motor X and Y, the shift is on the logic Y-axis (mechanic X-axis) only, logic X-axis(mechanic Y-axis) works fine.
3) I swapped the stepper driver board, the shift is on Y-axis. X-axis works fine.
4) I swapped the stepper driver board, at the same time, swapped the wire, the shift is on logic Y-axis,X-axis works fine.

So, I think the reason point to in front of stepper driver, I mean they work good from the input of stepper driver ,to motor, to pulley, to belt, to bed, to glass. But maybe the firmware or my A4983 cause the problem. I checked the Sprinter and Marlin code, I was trying to find "Step Pulse Output" codes or functions, and I still not found it. I will post the my checking result in few days.

Sorry to bring this back to life, but...
Having the very same problem. My shift is significantly large, like 10 mm, always after the 4th infil layer, always when moving to the start coordinate of the next layer.
I already slowed down marlin, checked for mechanical issues, read all threads I could find, but m problem persists.
voicevon, did you find the results?
thanks.

i have same problem,sometimes i find it shift,i will home x and y,then continue print ,it can print right and continue print

If you are not using Polulu/Stepstick but some other drivers - especially those including OPTO, check the rise/fall time of the opto/driver. You might need to add delay accordingly, to both DIR and STEP output.

Always the 4th layer? of every STL? If it's a particular STL, might be a particular pattern is causing the problem. Try lowering the acceleration and jerk number, the stepper might be not powerful enough.