Notifications
Clear all

Layer shifting problem solved (at least for me)  

  RSS
kyle.t2
(@kyle-t2)
New Member
Layer shifting problem solved (at least for me)

Folks - I wanted to share a solution I found that seems to be plaguing a lot of folks - random layer shifting. We have been using the Prusa MK2 for a year and a half and never had a layer shift problem, ever. Ever since we upgraded to the MK2.5, we have had this problem continually. And I'm not talking about a couple of millimeters. We are talking about 1 to 2 INCHES of layer shift - out of blue, and very repeatable (almost down to the layer). This made no sense to me, as if there was a belt tension issue or a pulley set screw issue (which is what Prusa suggests you look into), the problem should be far more random. You shouldn't get 100 perfect layers and then a 1 to 2 inch layer shift that alternates between the X or Y axis randomly. As I started researching this problem, it seems like it has been a problem for a lot of folks, regardless of which firmware version they are using. The only solutions have been to turn off Crash Detection or downgrading to a very old firmware version. But even turning off Crash Detection doesn't solve the problem 100% of the time.

I will be the first to admit that I am not an expert on 3D printers (my son holds that mantle), but I am an electrical engineer. After a lot of investigation and experimentation, I am convinced that the root cause of all this layer shift is a fundamental heat problem with the rambo control box. Let me give a little background as to why I have come to this conclusion.

We started printing with the MK2.5 in september of this year. At first we were printing very small simple parts, and we had no problems with those. It was only when we started to print larger, more complex parts did we start having layer shift.

The first instance of it came when the Y axis cable got caught up inside a part we were printing. The cable wrap that shipped with the upgrade was a little short, and so it didn't quite make it all the way to the rambo box. This meant that the cable didn't come in vertically into the rambo box, it was closer to a 90 degree angle. This in turn caused the cable to sag a bit over the print plate. Not a problem for short parts, but could cause a legimate jam once the part got tall enough, which in turn caused a 2 or 3 mm shift. I chaulked this up to the cables getting stuck in the part. What I didn't know was that this jam pulled out the cable stiffener from the extrusion unit, and lodged the pointy end of it right into the Y axis belt. This in turn caused the friction of the Y axis to slowly increase. As this happened, we started to see layer shift more and more often, just as you might expect.

Since the cables were not coming in vertically into the rambo box, when the part being printed got tall enough, the cables started to get pinched between the Y axis motor and the frame. In an effort to solve that, I put a piece of glass right along the left side of the rambo box to prevent the cables from moving too far to the left during printing. It worked great, but In the process of doing that, I basically covered up half of the rambo box vents. As the layer shift problem started to get worse, we changed the print mode to high current, thinking that the motors needed a bit more poop to do the job. And the problem got worse.

The last piece of the puzzle was that we switched filaments from PLA to something else, which required the bed temperature to go from 60 C to 85C. With this change, the layer shift that used to take 12-15 hours to occur, now happened in 2 hours.

My first step was to find out what was making the Y axis friction increase, which is when I discoved the cable stiffener was jammed up against the belt. So I fixed that and all of a sudden, the Y axis was smooth as silk again.

I restared a print and boom, layer shift again. Then I found someone somewhere who was having an identical problem with 1 to 2 inch layer shifts at almost the same layer over and over. He opened up his Rambo box and ran the part again, and got his first successful print after 4 tries.

So that all got me thinking about heat. Computers have massive amount of fans inside them because microprocessors become very unpredictable if they get too hot. Then I thought that everything we had been doing was increasing the heat load. We increased the current to the motors, we cut off part of the Rambo ventilation system, we ran a material that needed a higher bed temperature and every step of the way, the problem got worse.

So I did the following. I used a tie wrap to get the cables to enter the rambo vertically as they are supposed to. It's not a perfect solution, but far better than what I had. That allowed me to remove the glass plate next to the Rambo unit. I also changed the mode to Stealth, which reduced the current going to the stepper motors and lastly, I bought a small fan and I have it blow onto the Rambo unit (without blowing on the heat bed).

The net result - zero layer shift. Every part that I had trouble printing without layer shift has now beeing printed successfully without a single failure. We even have had parts with a 70% density (which shakes the printer a ton), print perfectly. Before, I couldn't get anything much over 12 hours to print without shift (2 hours with a bed temp of 85C), and now I'm running 30 hours prints without a hitch. I didn't upgrade or downgrade the firmware, at the end of the day, I just make sure the boards inside the rambo box stayed cool.

Now while this solution is working, It's the middle of winter here, so the ambient temperature where the printer is located is low. In the middle of summer, when it is 25 degrees warmer, I'm not sure if all of these steps will be enough to ensure successful printing. Hopefully Prusa will investigate this further.

Hope this helps.

Posted : 31/12/2018 7:09 am
Sembazuru
(@sembazuru)
Prominent Member
Re: Layer shifting problem solved (at least for me)

I'm still waiting on my Mk3 kit to arrive, but I've studied the assembly instructions. It did seem to me that the electronics enclosure is a bit cramped with the wiring and doesn't seem to be engineered for good air flow. I'll definitely pay a lot of attention here during assembly, and possibly design a larger enclosure with better cable management allowances some time in the future.

See my (limited) designs on:
Printables - https://www.printables.com/@Sembazuru
Thingiverse - https://www.thingiverse.com/Sembazuru/designs

Posted : 02/01/2019 3:11 am
kyle.t2
(@kyle-t2)
New Member
Topic starter answered:
Re: Layer shift problem solved (at least for me)

Yep - I just managed the cables a bit better and use that fan and we are 2 weeks in now and have run about 100 hours of parts with zero shift. Problem is basically solved for us.

Posted : 08/01/2019 11:23 pm
lohring
(@lohring)
Active Member
Re: Layer shifting problem solved (at least for me)

Thanks for your explanation. My layer shifting was more like 10 mm maximum and only on the X axis. I also made sure the wiring was better, but did not consider running in stealth mode. I thought that large or tall prints could be the issue, but heat buildup over time also makes sense. I have had very few issues since reworking the wiring. I'll also try some of your suggestions.

Posted : 09/01/2019 2:13 am
Joseph Craig
(@joseph-craig)
Active Member
RE: Layer shifting problem solved (at least for me)

So I've been having the exact same problem with shifts of about 1-2 mm but my brother keeps his prusa mk2 in a heated enclosure and doesn't have this problem, I wonder if if it is specific to the MK2.5 and maybe fixed in software? 

Posted : 17/11/2020 7:42 am
--
 --
(@)
Illustrious Member
RE: Layer shifting problem solved (at least for me)

no - it is 100% a hardware issue - sticky bearings; nozzle collisions with print curls; etc.

Posted : 17/11/2020 8:00 am
Joseph Craig
(@joseph-craig)
Active Member
RE: Layer shifting problem solved (at least for me)

@tim-m30

Im sorry but how would you explain it happening on identical layers over several different prints? do you have any suggestions to solve the problem? 

Thanks in advance

 

This post was modified 3 years ago by Joseph Craig
Posted : 17/11/2020 8:03 am
dickr@mac.com
(@dickrmac-com)
New Member
RE: Layer shifting problem solved (at least for me)

I’m having Y-axis layer shifts for one large print, but I’m sure that the hot-end cable is well above the print, so it can’t catch on a printed feature. I’m running an MK 2.5S with the latest firmware. The .gcode displays properly in the viewer, so I don’t think there’s anything wrong with the slicing by PrusaSlicer. Other large prints are OK. I printed something that went to the Y limits of the bed with no problems, but it was a geometrically simple part (just a large ring). I’m puzzled. If it were a temperature problem, I’d expect more than the one shift, and at random places.

 

Posted : 12/07/2022 3:03 pm
Michael Hartmann
(@michael-hartmann)
Member
RE:

Good afernoon everyone,

i read the news here and the whole thing seemed somehow familiar. I've been the proud owner of an MK4 for a few weeks now and have had few problems with it so far. However, for a few days now I have been experiencing more and more layer shifts, which mainly occur with overnight prints as I usually print the larger models then.

Do any of you have a solution for such layer shifts?I have attached 2 videos and a picture for better illustration.

83940dfb7101efa7ea51596c308c3d08 (2)

8f45ab9d3538833294d1ffef96158248

Greetings, Michael

This post was modified 4 months ago by Michael Hartmann
Posted : 15/11/2023 4:34 pm
Share: