Layer shifts on prints
I have had a Original Prusa i3 MS3S for a month now. Most prints have come out good very good.
I sa Thomas Sanladerer video on YouTube about "Making STRONG shelves with Topology Optimization". I liked the design. I downloaded it and set it to print with Prusa Galaxy Black PLA. Using the slicer default settings for filament.
I got these layer shifts. And some bad bridging. From what I could see from the video, Thomas didn't use any support. So I didn't either.
After printing the shelf supports I printed a mobile phone holder. And i got even worse layer shifts.
I have uploaded som images to Google Photo: https://photos.app.goo.gl/RunHh4FzT9iPsNvd9
Any help to understand what is going on would be appreciated. I haven't experienced layer shifts on previous prints.
Shifts are most likely bearings seizing or other friction related issue.
Bridging looks like normal artifacts when printing with bridge detection off.
Recently I have been getting about a 1 to 2mm Y axis shift at about the 10mm level. If I read you right then I may need to replace my Y axis bearings and/or rods. Ideas? The printer has been running fairly constantly for over 2 years with only greasing of rods and no disassembly except for the 2.5, sensor and extruder upgrades.
Did another test print today and do not think it is bearings or rods as I get the shift(s) at exactly the same place and it is a shift first to the rear and then about 5mm higher a shift back to the front on the Y axis. A problem with slicing and the G code maybe or a problem in the stl file of the object that does not show up as a problem?????
Or it could be a problem with your wires, if there is an internal break it could happen when the cable bundle gets to certain positions. Difficult to track down though without a multimeter to buzz through as you move the bundle around.
got a multi meter. It would have to be a Y axis motor wire since that is where the problem appears to be (shift is in Y axis only) . Which wires do you suggest? The Y motor wires do not move but who knows.
I did some basic cube and cylinder combination objects and they printed fine. I just did new download and a reflash of the FW (latest) and a reload of the SW (latest). My last test prints of the objects I had the problem start with and it actually got worse.
If the single object print I am doing now does the same thing I will try a proven object (printed OK before) and if that also fails then I will tear into the bed and bed rails/bearings. New set of rails arrive today and have spare bearings. The rails do show some wear but are original Prusa Mk2 rails and have been operating for over 2 years.
Will advise of what I find. Thanks.
Keep in mind "greasing the rods" is not correct.
Pack the bearings with appropriate grease - and keep the rods clean.
Rods are funny - they only wear where the balls hit. Rotate the rod a bit - and you have a new surface for the bearings.
I have rotated the rods twice before so now, last night, tore down the bed and replaced rods and bearings. Printer much quieter now and prints OK. I guess this is something we all should do as part of periodic maintenance. Note that the bearings sounded very gritty.
So where is best source for new (good) bearings and rods?
Has anyone tried the plastic bearings and if so, how are they? Seems they are much quieter. Tried on old rods but no installed.
Another rule of thumb:
If you rotate the rods due to wear - replace the bearings. They are done. And grease them when you do.
I don think that this is an issue on bearings, rods and friction. I had the same problem when I changed my firmware. I have a MK2.5 and updated from 3.8.0 to 3.8.1. Going back to 3.8.0 using the same GCode File for print solved the issue.
I printed this file twice on fw 3.8.1 and the layer shifting started allways on same z-level. Going back to fw 3.8.0 with same rods, bearings, friction and so on solved the problem.
Best regards, Clemens
And an other question (I couldn't Edit my Post, because it is under status "Awaiting moderation"):
Where is the correct place to report an issue on firmware and to send the GCode File to give the software engineers the oportunity to check / solve such an issue?
I think there is a time out on editing posts rather than a restriction whilst awaiting moderation,.
I believe the correct place for raising issues is on the Prusa GITHUB pages.
I did the same thing. Now running 3.8.0 again and no shifts. 21 hours of flawless printing.
If the corners of the print lift off the bed due to warping from cooling, it will leave ridges on upper Z heights. Also using Octoprint with a Raspberry Pi 1 can give horrendous layer shifts.
None of those problems exist. No PI and no lift. Just did a 16 hour print and no problem. still using FW 3.8.0 ven though LCD tells me to update to 3.8.1, just press the button and go on and print.
I wonder if anyone at PRUSA has seen this problem or at GITHUB...whoever writes the firmware.
I just experienced a 2mm shift on the y-axis on a print after about 8 hours of printing. I am using the bondtech upgrade kit for the mosquito hotend on a Prusa i3 mk3s. I was using the bondtech firmware version 3.8.1 which is I believe the same as the Prusa firmware with modified values for extruder steps. I am printing directly from the SD card. I have had several successful prints with version 3.8.1 some of which were 12 or more hours long with no layer shifts. This time everything was working flawlessly when the layer suddenly shifted by about 2 mm. I have reloaded firmware version 3.8.0 and will report back if this solved the problem.
For those having shifts and believe it isn't bearings, a single bearing rod combination can and will occasionally stumble at the same location. Think of what happens when there is a microscopic notch on the rod. Things will work fine then when the bearing hits the notch things seize and the motor stalls. And the seizure can be affected by speeds, such that at high speeds it doesn't happen, at low speeds it doesn't happen: things have to be just right. And that can be at a spot 48 mm high into a build, and it will seem like it's not hardware.
It may also be firmware: disabling the filament sensor usually fixes those issues.
I switched to firmware 3.8.0 and this DID NOT fix the layer shift issue. I tried to reprint the same model and this time observed a layer shift and eventually a crash that stopped the printer. In my case, I believe the nozzle was crashing into the print due to excess build up of material on and around the nozzle. It appears as if the filament (carbon fiber infused PETG) was stringing excessively and building blobs that were hitting the nozzle causing crashes and layer shifts. I dried the filament for 8 hours and started the print again. So far I have not seen the layer shift after dying the filament.
Sounds like your synopsis is correct. Unless you have a serious processing issue, it's mechanical force that will cause a stepper to miss steps and then its game over.