Periodic Re-Homing During Prints
I just received my third Prusa i3. The previous two were Mk3S that I built from kits, this one is a Mk3S+ that I purchased fully assembled.
I did the first layer calibration and started a pretty large print. I noticed that periodically during the print, it would lift up the extruder by maybe 10-15mm, move all the way to the left on the X axis, and then move the table all the way back in the Y axis. Basically re-homing. Then it will go back and resume the print normally. Also when this happens "Crash Detected" appears on the display. I normally never see that error unless the extruder has actually crashed into something, like a hardened blob on a print.
After witnessing this I went in and performed a full X/Y/Z calibration. It said the X and Y axis were slightly skewed, but that it was good to go. This didn't alarm me, as I have witnessed it on my other printers and they still work fine. At the same time I lubricated the guide rods for the X and Y axis, in case they were dragging and triggering the crash warning.
This constant interruption of layers doesn't seem to have affected the print quality, except perhaps in one corner. The print is taking up almost the entire width of the bed.
One factor worth mentioning is that I am using a 10-pound roll of Matterhackers PET-G, on a spool holder overhead, fed through a wire shelf. So it takes a fair amount of force to pull filament off of the spool. However I used this exact same setup with my previous two printers and did not experience any problem. Maybe the 3S+ has a more sensitive detection that the 3S, or maybe just this particular printer does?
If it says crash detected then that's what has happened and you should investigate why and fix it, it should not happen under normal circumstances.
For example, perhaps the cable bundle to the extruder catches on something or the print itself gets lots of blobs so nozzle catches on it.
Also make sure the filament can be pulled freely from the spool without catching on something.
I have at times had problems with the prusament spools and the little slot where the filament is stuck in to when they are new.
When filament gets pulled over the edge (i got an MMU2) it can on rare occasions get stuck there and at least once it has caused a jam.
Another possible reason for crashes is running into infill pile-ups. Some infills—such as grid—cross inside a layer, causing material to pile up, which can (but does not have to) cause crashes.
Formerly known on this forum as @fuchsr -- until all hell broke loose with the forum software...
Y Axis Crash
Thanks for the replies!
I have established that it is not due to filament tension, using a different roll that pays out easily.The failure log on the machine tells me they are all Y axis failures, over a hundred of them since I fired up the machine for the first time yesterday.
There are no obstructions that I can see, and when I slide the carriage forward and back in the Y axis with the steppers disabled, I don't feel what I would think to be abnormal resistance at any point in the path. The belt tension feels normal.
As far as filament buildup... This is my third i3 printer, and I have done hundreds of prints, with both PLA and PET-G, virtually all of them with infill. While I have seen crashes due to filament buildup, it has always been obvious globs of material. I'm not seeing anything like that with this print.
By the way, I have printed this same file with one of my other printers, multiple times, with no crash, so I think if it were infill I would have seen it then.
I welcome any further suggestions. Thanks!
Are your Y bearings over-tightened?
Focusing on the Y aspect of your details. Do you hear a "klunk" when it crashes or is it silent? Have you used the belt tension gauge on Y to determine if the belt is too tight? One thing that can cause this is Y bearing holders are over-tightened. I can't speak to the bearings or whether they were adequately lubricated on assembly, as I don't use linear bearings that need lubricating.
RE: Y Drag?
There does seem to be a clunk, but I haven't witnessed an event directly enough to see exactly when it is happening; there's also a clunk when it gets to the far left X stop, and when the table gets to the back Y stop, so I could be confusing it with that.
I think I have ruled out hard interferences (obstructions) in the Y path, but like you I do still suspect drag in the Y axis, although it doesn't feel severe. I will check the bearings. I do believe that when I lubricated the rods it reduced the frequency of the failures.
I need to have both of my side-by-side printers not running so I can do a direct comparison of the Y drag, right now they are both printing jobs. That's the thing though, the recovery of these printers compared to other ones I have had is remarkable. The Y faults have certainly slowed down the process, but they haven't affected the ultimate print quality.
I don't have a belt tension gauge, but the tension does not seem unreasonable.
This all is a little disappointing, since this is my third i3, but the first one I didn't assemble myself.
Additional Info
I compared the Y belt tension to my other two i3's and it is definitely tighter, but the carriages seem to move equally freely.
Before I went in and tried to tweak the belt tension I decided to try something a little easier, I backed the idler pulley bolt out about 3/4 of a turn, in case it was binding. This did not make the problem go away.
Can belt tension alone cause this? I just would have thought that the factory would have better belt tensioning ability that me.
P.S., why does the forum software require me to apply a title to every single post? It doesn't appear to do that for everyone here.
Belt Test on Calibration Menu
Have you tried belt test on the calibration menu? This will give you a reading for the tension on the belt driven axis and you can lookup expected values online.
Hi Kel,
unfortunately, since the last forum update users have been required to insert a title.
In this instance, I inserted a single space, let's see what happens when I post the reply!
regards Joan
I try to make safe suggestions,You should understand the context and ensure you are happy that they are safe before attempting to apply my suggestions, what you do, is YOUR responsibility. Location Halifax UK
Looks like you get an invisible title, when you use a space...
Joan
I try to make safe suggestions,You should understand the context and ensure you are happy that they are safe before attempting to apply my suggestions, what you do, is YOUR responsibility. Location Halifax UK
Hello everyone, sorry for the delayed response, but I wanted to read up on the topic a little and try a tweak or two so as not to waste everyone's time.
Thanks for the information on the belt test (these printers are amazing). I ran the test four times successively. Here were the Y values, in order:
219 -> 245
245 -> 243
243 -> 244
244 -> 243
Then I flipped the printer over and loosened the belt by one full turn of the adjusting screw. Not much, I know, but I can do it again if necessary, it's a quick procedure.
New Y numbers:
243 -> 250
250 -> 251
251 -> 250
250 -> 250
So it looks like I bumped the average number up by about 6 or 7 points. I'll try some prints in the next week or so and see if the issue is resolved.
How's this work? My guess is that the total belt length is slightly longer when loosened, allowing a tiny number more microsteps per full carriage run. Effectively, the belt moves slightly more past the point where the carriage stops, and they count the extra microsteps before the motor stalls. This would explain why a larger number equals a looser belt. Am I guessing right?
Also thanks for the tip on the title line, I think it worked! I appreciate all of the help everyone.
hotend wiring?
I did have an issue like this, and from what I remember it was due to my wiring bundle position under the back of the hotend, it was knocking something and being detected by the TMC drivers as a crash. I aligned it correctly, cleaned my Y axes and readjusted the screws on the bed and Y axis, and the problem disappeared.
I have seen hotend wiring interference described as more of a threat to the X axis. There were a couple stray wires protruding from the control box on mine that could have interfered with the Y axis, but I discovered those early on and got them out of the way.
At the moment I am leaning towards excess Y belt tension as the culprit, I will have to do another long print job to find out if I have fixed it.
Thank you for the reply.
Get the belt tension gauge and print it
You can download it from the print library - print it in any color Prusament PETG as long as it's Prusament PETG (other brands/grades will not read correctly - the GCODE is for Prusament PETG and it's somewhat unique stiffness characteristic.
Your pre-adjustment readings indicate that there was too much resistance and your post adjustment readings are improved but still quite close to the limit. If it were me I would also consider the following actions:
(1) printing the Prusa belt tension test tool with Prusament PETG and checking the tension to see if there is scope to loosen the belt further
(2) checking the three bearing clamps are not too tight
(3) loosen the two bolts on each of the four clamps that hold the rods to the frame and then slide the bed all the way to the front, tighten the front screws, slide the bed all of the way to the back and then tighten the rear ones.
(4) Put some grease on the rods (or ideally remove the bearings and grease them properly)
Hopefully the above items would get you to a consistent reading of over 250.
Hello everyone, sorry for the delayed response, but I wanted to read up on the topic a little and try a tweak or two so as not to waste everyone's time.
Thanks for the information on the belt test (these printers are amazing). I ran the test four times successively. Here were the Y values, in order:
219 -> 245
245 -> 243
243 -> 244
244 -> 243
Then I flipped the printer over and loosened the belt by one full turn of the adjusting screw. Not much, I know, but I can do it again if necessary, it's a quick procedure.
New Y numbers:
243 -> 250
250 -> 251
251 -> 250
250 -> 250
So it looks like I bumped the average number up by about 6 or 7 points. I'll try some prints in the next week or so and see if the issue is resolved.
How's this work? My guess is that the total belt length is slightly longer when loosened, allowing a tiny number more microsteps per full carriage run. Effectively, the belt moves slightly more past the point where the carriage stops, and they count the extra microsteps before the motor stalls. This would explain why a larger number equals a looser belt. Am I guessing right?
Also thanks for the tip on the title line, I think it worked! I appreciate all of the help everyone.
That is all great advice, thank you Brad. I will be back using the thing this weekend so I can see what works. That belt tension gauge is simple and cool too. I appreciate the replies!