RE: Prints getting squished at the left edge of the print bed
Gnat -this is a head scratcher. There is nothing I can think of in the drive system than can deliver a non-linear feed except a belt that is stretching or is bad from the factory. And it has to be a change in belt tooth pitch. Nothing else I can think of will cause the extruder to move linearly for 200 mm then start compressing over the last 10 mm.
The easiest reason for this belt defect is problematic, and makes it even weirder (belt was over stretched during motor installation). Because the bad length should be shorter, not longer than the rest. Though, if the belt chords were pulled loose from the rubber, the rubber might be compressing the section. lol - a lot of hypotheticals ...
I agree, weird one. If it were a stretched belt I would expect to see consistent results no matter where the print was placed. I suppose it would be an issue with the teeth right near the left side of the extruder though. That would only present when they met the gear which would only be at the near-0 position...
@david-j49
I couldn't really get it in a photo, but I've noticed the top of the belt is closer to the from of the printer than the bottom. As if I pushed it too far into the extruder body.
Is there any warping/twisting of the belt as it comes out of the extruder on that side? Is the belt inserted evenly on that side of the extruder? If the belt is not in uniformly I suppose it could cause some twisting that could theoretically cause it to act like a stretched belt or skipping teeth as it got close to the drive gear...
RE: Prints getting squished at the left edge of the print bed
Is there any warping/twisting of the belt as it comes out of the extruder on that side? Is the belt inserted evenly on that side of the extruder? If the belt is not in uniformly I suppose it could cause some twisting that could theoretically cause it to act like a stretched belt or skipping teeth as it got close to the drive gear...
I thought so, too. There was a bit of twisting, so I pulled the end of the belt out from the extruder body until it lined up with the bottom part of the belt. Now the belt sits in the middle of the pulley (not pushed up against the front side as in the photos) and doesn't twist at all. No change.
Also I've noticed the printer sounds different when it's in the problem area. Kinda like it's straining, but I have no idea against what. When I move the extruder manually, I feel no change in resistance anywhere.
RE: Prints getting squished at the left edge of the print bed
Whatever is causing the straining will be the issue.
Try removing and reversing the belt and see if the issue stays on the left or if it moves to the right. If it does move that would point to a bad belt.
Try removing the belt and turn the motor a quarter or half turn, reinstall the belt, and see if that effects the print. If it does that may mean you have a motor problem. I doubt this one.
RE: Prints getting squished at the left edge of the print bed
I'll give that a try. Unfortunately I don't have time for that right now. Maybe tomorrow, but probably Sunday. I'll post the results.
RE: Prints getting squished at the left edge of the print bed
Straining near EINSY case, but not obvious interference....
That could be the cable bundle and its nylon producing extra strong spring effect as carriage nears EINSY box. I ran into this issue while sorting out X-homing with 0.9 degree motors. You can reduce this effect with a couple maneuvers.
1. Reduce tightness and twisting of fabric wrap. This reduces intra-bundle stiffness and lengthens the bundled portion.
2. Don't tighten cable camp at EINSY with x-carriage in middle of its range. Instead, position carriage max left while tightening cable clamp. Doing so sets the bundle to be less stressed when carriage is max left.
RE: Prints getting squished at the left edge of the print bed
Thinking about this some more... if it were a spring effect problem cause missed steps and compression, it would NOT correct as print head moves rightward. So, no. This can't be spring issue.
However, you can induce a non-linearity at end of axis if motor pulley is above or below level of belt connection to x-carriage. As carriage approaches motor, the belt would pull with increasing vertical motion and less horizontal. That would non-linearly compress prints near carriage travel end AND correct itself as carriage travels rightward.
Is the TOP of belt level when x-carriage is near left?
RE: Prints getting squished at the left edge of the print bed
Thinking about this some more... if it were a spring effect problem cause missed steps and compression, it would NOT correct as print head moves rightward. So, no. This can't be spring issue.
I think it depends on what is causing the springing actually. If there is something physically in the way or there is a localized belt issue (e.g. bad teeth) that could cause it just at that side. I think you have another good idea as to a possible cause though with the motor position.
Another idea I had last night was to move the extruder to the right side and clean off the X rods with a paper towel and IPA. Then feel the rods at the left side and see it it feels like there is any imperfections. An extreme test would be to reverse the rods.
You might also check the height of the Z carriers and make sure they are close to matching. If one side was high that might provide some binding on the X axis. I would expect calibration issues if it was that bad though.
I mentioned doing a small print that utilized most of the left side to see if the warping is localized, we're you able to do that? Another thought is to do one (not full Y length) that uses a good portion of the Z plane to see if it is localized in that direction.
RE: Prints getting squished at the left edge of the print bed
Gnat - springy - as in giant spring between extruder and support - won't cause what's happening. Stepper motors down 'slow down' when they are over loaded. They stall, and that leads to shifting, not distortion. Try this your self: try holding the X-Axis motor gear or belt while printing.
A small section of belt is damaged. You can see signs of damage in the photos. The inner radius of the belt appears worn, and a smaller belt radius will move the extruder a smaller distance over the gear. Of course, this is all hypothetical because I can't measure the belt... lol.
Guy - the entire motor carriage assembly is fixed. It the belt vector were a problem, it's something we'd all see, and it'd be a variation over the full length of the belt - side to side (non-linear, but visible).
RE: Prints getting squished at the left edge of the print bed
Straining near EINSY case, but not obvious interference....
That could be the cable bundle and its nylon producing extra strong spring effect as carriage nears EINSY box.
Actually quite the opposite — the cable bundle is causing a bit of resistance towards the right end of the X axis (but it's obviously not causing any issues). After moving the zip tie, the extruder now moves freely all the way to the left until the fan screws hit the Z axis screw.
However, you can induce a non-linearity at end of axis if motor pulley is above or below level of belt connection to x-carriage. As carriage approaches motor, the belt would pull with increasing vertical motion and less horizontal. That would non-linearly compress prints near carriage travel end AND correct itself as carriage travels rightward.
Is the TOP of belt level when x-carriage is near left?
Yup, top of the belt stays level all the way. I did have a bit of twisting as I'd pushed the belt too deep into the extruder body, so it was pushing against the front of the pulley as the extruder moved left, but I've fixed that and it's made no difference.
I think it depends on what is causing the springing actually. If there is something physically in the way or there is a localized belt issue (e.g. bad teeth) that could cause it just at that side. I think you have another good idea as to a possible cause though with the motor position.
Another idea I had last night was to move the extruder to the right side and clean off the X rods with a paper towel and IPA. Then feel the rods at the left side and see it it feels like there is any imperfections. An extreme test would be to reverse the rods.
You might also check the height of the Z carriers and make sure they are close to matching. If one side was high that might provide some binding on the X axis. I would expect calibration issues if it was that bad though.
I mentioned doing a small print that utilized most of the left side to see if the warping is localized, we're you able to do that? Another thought is to do one (not full Y length) that uses a good portion of the Z plane to see if it is localized in that direction.
I will try rotating the motor (I'm assuming you guys mean just loosening and giving the shaft a quarter turn, not rotating the whole motor, right?) and feeling the rods when I get home from work. I don't think I'll have time to reverse the belt tonight though (and I'll be out of town till Sunday evening).
As for the test print, I was thinking of printing a 45° rotated checkerboard across the whole bed and then putting the whole thing in a scanner. To make sure the issue does indeed only occur at the left edge (I think I'd have noticed, but you never know). As for Z, how about a tall cuboid with pyramids pointing to the edge? Models attached (the checkerboard may need work).
RE: Prints getting squished at the left edge of the print bed
A small section of belt is damaged. You can see signs of damage in the photos. The inner radius of the belt appears worn, and a smaller belt radius will move the extruder a smaller distance over the gear. Of course, this is all hypothetical because I can't measure the belt... lol.
If calipers are enough, I can measure the belt. Just tell me what to measure.
RE: Prints getting squished at the left edge of the print bed
The thickness of the belt at the teeth and thickness of the belt in the notches. Compare the section that is on the pulley when the extruder is over the "bad" area to a mid-sections of belt. The differences may be small. Normally the belt should be exactly the same thickness everywhere.
And this may not show the issue if the belt rubber has broken free of the cord. You can lock the motor gear (non-marring pliers) and try moving the extruder to see if you can feel any play then reposition the extruder midway and compare.
Guy's idea is intriguing, but doesn't explain an almost immediate change in scale you are seeing. A pulley height alignment will be non-linear over the entire length of travel.
Anyway - this issue is fringe territory.
A test of the belt theory would be to do a print of the diamonds in the four corners. If the two left corners are identical to the left middle, and show the compression, there isn't much else that would cause this than the belt. Though, you may also want to do a general linearity test, lots of diamonds across the bed, just to be certain it's only happening at the one edge.
RE: Prints getting squished at the left edge of the print bed
Guy's idea is intriguing, but doesn't explain an almost immediate change in scale you are seeing. A pulley height alignment will be non-linear over the entire length of travel.
I think his idea has potential (though david has noted that the belt appears level). The further the X carriage is from zero the more room the belt has to naturally adjust to the angle being dictated by the rods and allow for smooth normal movement. As the carriage approaches zero, however, the belt has less room to adjust and would start attempting to move the carriage away from the dictated path. Since there is no way the belt/motor is going to win that battle, it would simply manifest as an increased drag close to zero. At least that is my impression of his theory.
I would expect calibration/crash issues if it was significant enough to cause a print issue though.
RE: Prints getting squished at the left edge of the print bed
A pulley height alignment will be non-linear over the entire length of travel.
Yes, but it will greatly steepen as carriage gets close to pulley. However, I agree, this is very weird effect and damnedly hard to explain with usual defects.
Could this be Marlin modifying travel distance that it computes as out of bounds? Maybe test piece with piece 5 mm shy of max left to avoid crossing a possible threshold.
RE: Prints getting squished at the left edge of the print bed
The thickness of the belt at the teeth and thickness of the belt in the notches. Compare the section that is on the pulley when the extruder is over the "bad" area to a mid-sections of belt. The differences may be small. Normally the belt should be exactly the same thickness everywhere.
And this may not show the issue if the belt rubber has broken free of the cord. You can lock the motor gear (non-marring pliers) and try moving the extruder to see if you can feel any play then reposition the extruder midway and compare.
Guy's idea is intriguing, but doesn't explain an almost immediate change in scale you are seeing. A pulley height alignment will be non-linear over the entire length of travel.
Anyway - this issue is fringe territory.
A test of the belt theory would be to do a print of the diamonds in the four corners. If the two left corners are identical to the left middle, and show the compression, there isn't much else that would cause this than the belt. Though, you may also want to do a general linearity test, lots of diamonds across the bed, just to be certain it's only happening at the one edge.
How's this (see attachment) for a full-bed linearity test? I could put the whole bed into a scanner afterwards.
I was also thinking I'd go back to the filament and bed (smooth PEI) used for the model where the problem first manifested itself, just to remove as many variables as possible. I've also changed the infill pattern to concentric, because the original part was pretty much all perimeters, so it should be a better representation. Any suggestions? I'll be heading home shortly, so I'll start the print in about an hour.
RE: Prints getting squished at the left edge of the print bed
A really really simply full bed linearity test is a 250 x 210 box - set for 20% grid infill, no bottom layers.
ps: and infill alignment set to 0 degrees.
pps: but those triangles should do it.
ppps: and one or two layers should be enough ...
RE: Prints getting squished at the left edge of the print bed
Mystery solved. The print is simply going outside allowed print area. Marlin is scaling the x motion to keep things inside legal bounds.
There is even a warning in PrusaSlicer when you slice the file.
RE: Prints getting squished at the left edge of the print bed
Mystery solved. The print is simply going outside allowed print area. Marlin is scaling the x motion to keep things inside legal bounds.
There is even a warning in PrusaSlicer when you slice the file.
I'm getting no such warning, have you turned off the skirt?
Also if this were the case, I wouldn't be the only person having this problem. Or are you saying I'm the first person to try to use the entire advertised printing area?
RE: Prints getting squished at the left edge of the print bed
I must have nudged it. You're right at the limit of the build area. Even a 1 mm nudge puses out of bounds.
That is so close that Marlin and PrusaSlicer may not exactly agree as to legal build area. It could even be a < vs <= level difference.
At any rate, the model is making Marlin scale motion during moves printer firmware defines as out of bounds.
RE: Prints getting squished at the left edge of the print bed
I've seen prints that were out of bounds due to an improper bed offset. It clips, it does not compress.
And the test model has no issues:
RE: Prints getting squished at the left edge of the print bed
OK I don't see what I am seeing.