RE: HOMING ISSUES
When I do autohome in the menu even when heating to 170C, it homes correctly. It is almost as if the homing through GCODE from the slicer is messed up somehow.
RE: HOMING ISSUES
Is it possible that this depends on the position of the printer bed? I had succsessfull homings when bed is on bottom. Firmware version 6.3.3, Core one conversation kit.
RE: HOMING ISSUES
I went through this issue and I have to say that at the end I solved it repeating well the config of the X Y axis and belt tension.
So disable motors, bring the extruder to the front completely and check if both left and right are exactly touching the frame. If left or right have few mm of gap try to push a little bit to align as much as possible. Then bring the extruder to the back and start increasing belt tension. Keep the belt in which you had the gap a little bit more tight (maybe 3 or 4 Hz more vs the other belt). Now my homing bumbs about 10 times and then start printing with original gcode.
had same issue as described by OP and other people.
I did the CoreXY alignment but still had issues with banging then I tried it carefully again but I changed step 4. a little bit. in stead of loosening belts little by little I loosened them all the way and started to tighten both of them little by little making sure both gantry corners are still touching corexy. when one corner stopped touching I loosened one belt until correct and then continued to tighten them synchronously.
from what I understand is that if you tighten one belt before you tighten the other it can squeeze the gantry and it won't be square anymore.
RE: HOMING ISSUES
This is basically what i did too. I started doing this after i noticed that the "squares" was lost after using a higher belt tension. I think loosening the belts completely and then aligning the gantry is the way, assuming all the gantry construction is solid but only the angle brakets are a bit off.
Then using the multitool key to jam one side ( there is a picture/video somewhere ) and bend on the other. But IMHO bending the angle brakets without a machinists square is suboptimal. Only with a square one can check if the are correct. Just bending for the gap gone can lead to alignment where the gap is gone but the brakets are not at ninety degrees.
I went through this issue and I have to say that at the end I solved it repeating well the config of the X Y axis and belt tension.
So disable motors, bring the extruder to the front completely and check if both left and right are exactly touching the frame. If left or right have few mm of gap try to push a little bit to align as much as possible. Then bring the extruder to the back and start increasing belt tension. Keep the belt in which you had the gap a little bit more tight (maybe 3 or 4 Hz more vs the other belt). Now my homing bumbs about 10 times and then start printing with original gcode.
had same issue as described by OP and other people.
I did the CoreXY alignment but still had issues with banging then I tried it carefully again but I changed step 4. a little bit. in stead of loosening belts little by little I loosened them all the way and started to tighten both of them little by little making sure both gantry corners are still touching corexy. when one corner stopped touching I loosened one belt until correct and then continued to tighten them synchronously.
from what I understand is that if you tighten one belt before you tighten the other it can squeeze the gantry and it won't be square anymore.
RE:
...BUT before each print the Home Calibration takes minutes!!
Here is a simple woraround for the Home Calibration problem before the print, in your Prusa Slicer, in Printers -> Custom G-Code, change the command
G28
To the command:
G28 I
(that is, add a capital I flag), and save. Now when you will print, the print head will bump twice on X, twice and Y, and will continue with printing.
This is a known issue and Prusa confirmed that they are working on a proper fix.
RE: HOMING ISSUES
I assembled a kit and had the same problem. I managed to completely resolve my head banging issues with the following steps:
1. I followed the first 3 steps instructions from this comment by SgtCaffran:
Don't just tighten the belts. Do the following:
1. Make sure the belts are completely loose (disconnect the tensioners)
2. Make sure both belts have the same number of teeth pulled through the Nextruder carriage. Aim for 5 teeth, should be possible with the belt tensioners disconnected.
3. Make sure each belt has the same number of teeth pulled through left and right. (Again: 5 teeth)
(I didn't follow the steps below!)
4. Grab the gantry in the middle and carefully bring it towards the front. Check which side hits the end stop first.
5. Put one of the flat tools between the end stop and gantry on the side that hit first. Now, carefully pull the other side of the gantry a little bit towards the front.
6. Repeat steps 4 and 5 until both sides of the gantry hit the end stop at the same time.
7. Reattach the belt tensioners and tension the belts according to the new strategy: https://belt.connect.prusa3d.com/
8. Check if both sides of the gantry still hit the end stop at the same time after tensioning.
9. Should be good to continue!
2. I noticed that the left belt tensioner pulley had a wrong screw (M3x12 and not M3x18). So the pulley axis perhaps was not very stable. I replaced it with the correct screw, see section 5.24, Assembling the belt tensioner left II. In fact I would recommend also using M3X18 for the door sensor adjustment, M3x12 is to short in my case to calibrate the door sensor.
3. I followed the belt adjustment procedure. But instead of tuning the top belt fully and then the bottom one I carefully tensioned both bit by bit while checking that there's no play between the left and the right (you have to move the head to the middle for that and then back to the right for the tension).
RE: HOMING ISSUES
I'm active here as well, happy to read that my comment helped you out! I think the mounting of the belts to the Nextruder carriage has much bigger effects than we think and deserves more attention in the Prusa guides as well.
RE: HOMING ISSUES
Thanks both, I’m a newbie to 3D and my core one kit is my first printer … in my head I’ve been wondering about the randomness of the setting up of the belts and mostly the fact that as I just pushed them ‘through a bit’ and certainly they were not identical for x & y … I’m going to look at how to start afresh with connecting the belts to the carriage to get them even … any hints on how to approach this task? Funny that I’m thinking “why didn’t I just buy a built one”, but strangely confident that as I built it I should be able to sort it out, just a bit upset that this wasn’t really explained in the build process! I agree more detail should be in the build instructions… as I said any shortcuts to carrying out redoing the connections would be most appreciated!
Petr
RE: HOMING ISSUES
I can't give you any shortcuts for the partial disassembly. However, when redoing the belts in the Nextruder carriage, make sure to completely disconnect the belt tensioners. makes it so much easier.
RE: HOMING ISSUES
in my head I’ve been wondering about the randomness of the setting up of the belts and mostly the fact that as I just pushed them ‘through a bit’ and certainly they were not identical for x & y … I’m going to look at how to start afresh with connecting the belts to the carriage to get them even … any hints on how to approach this task? Funny that I’m thinking “why didn’t I just buy a built one”, but strangely confident that as I built it I should be able to sort it out, just a bit upset that this wasn’t really explained in the build process! I agree more detail should be in the build instructions… as I said any shortcuts to carrying out redoing the connections would be most appreciated!
The build instructions are specific enough, I'd say. They tell you to "leave 4 to 5 teeth sticking out" on either side: https://help.prusa3d.com/guide/5-corexy-assembly_835522#864764
But I can't see how symmetry could be critical here. It's more about having a reasonable free belt length to get a good adjustment range via the belt tensioners, and of course having enough teeth clamping the belt to hold it securely. But if you have 5 teeth in the clamping loop on one side, 3 on the opposite side -- how can that make any difference for the behavior of the belt and the CoreXY motion?! No need to overthink this.
RE:
any shortcuts to carrying out redoing the connections would be most appreciated!
See section extensive belt adjustment. But first make sure to completely loosen the belts (disconnect the tensioners, as recommended by @SgtCaffran, see above).
While you're at it, check that screws in the tensioners, the ones going the pulley are the correct ones (going through the plastic part fully).
RE: HOMING ISSUES
Thanks for the link, that’s exactly what I needed to see to understand how best to get to the belts …
Pete - the Idle tinkerer
RE: HOMING ISSUES
But I can't see how symmetry could be critical here. It's more about having a reasonable free belt length to get a good adjustment range via the belt tensioners, and of course having enough teeth clamping the belt to hold it securely. But if you have 5 teeth in the clamping loop on one side, 3 on the opposite side -- how can that make any difference for the behavior of the belt and the CoreXY motion?! No need to overthink this.
Matching tension by matching frequencies assumes that the length of the vibrating 'strings' is identical, that is the length between the pulleys where the belt is free to vibrate. If you leave out on one belt more teeth than the other, the string lengths will be different and 85Hz on both sides will result in a different tension. That's theoretical, I don't know the actual numbers.
The root cause though is the poor homing refinement algorithm that Prusa invented and added to Marlin, which is causing more harm than benefit. Tensioning belts on other CORE X/Y printers with standard firmware is a non issue.
RE: HOMING ISSUES
Matching tension by matching frequencies assumes that the length of the vibrating 'strings' is identical, that is the length between the pulleys where the belt is free to vibrate. If you leave out on one belt more teeth than the other, the string lengths will be different and 85Hz on both sides will result in a different tension. That's theoretical, I don't know the actual numbers.
Nah, sorry -- I don't buy that. With the old "tuning" procedure, where the belts were plucked on the sides, there was a theoretical impact because the plucked belt length did change with the tensioner position. But we are talking about a 1% effect here for belt length and frequency, and about 2% belt tension; that's hardly significant.
With the revised procedure where you tune by plucking the belt at the X gantry, the frequency and hence the tension will be fully independent of the tensioner position and hence the belt length.
RE: HOMING ISSUES
Nah, sorry -- I don't buy that. With the old "tuning" procedure, where the belts were plucked on the sides, there was a theoretical impact because the plucked belt length did change with the tensioner position. But we are talking about a 1% effect here for belt length and frequency, and about 2% belt tension; that's hardly significant.
With the revised procedure where you tune by plucking the belt at the X gantry, the frequency and hence the tension will be fully independent of the tensioner position and hence the belt length.
Prusa's invented a faulty algorithm and we are all busy developing black magic about it.
Prusa is failing its users by not reverting it.
RE: HOMING ISSUES
Prusa's invented a faulty algorithm and we are all busy developing black magic about it.
I'm actually not sure it's a fully software issue. In my case that was a symptom of a hardware problem (wrong screw, not fully supporting a pulley). Sure, the software could be a bit more informative, but at least in some cases that's diagnosing an existing hardware problem.
RE: HOMING ISSUES
Can it be a hardware (Xbuddy?) problem?
I have a coversion from MK4S to C1 and the first few days it worked OK, only a bit long homing-process, but that seems normal. But now I have only red screens and everything is checked dring a long time with a Prusa chat. Pully, tension, 90deg and so on.
So I think it is probebly an electronics issue.
what do You think?
RE: HOMING ISSUES
Prusa's invented a faulty algorithm and we are all busy developing black magic about it.
I'm actually not sure it's a fully software issue. In my case that was a symptom of a hardware problem (wrong screw, not fully supporting a pulley). Sure, the software could be a bit more informative, but at least in some cases that's diagnosing an existing hardware problem.
The extensive test should be in the test/diagnostics, not on each print.
What do you get now when your printer is fix, a single head bang on X and a single head bang on Y like normal sensorless printers or many more?
RE: HOMING ISSUES
Can it be a hardware (Xbuddy?) problem?
I have a coversion from MK4S to C1 and the first few days it worked OK, only a bit long homing-process, but that seems normal. But now I have only red screens and everything is checked dring a long time with a Prusa chat. Pully, tension, 90deg and so on.
So I think it is probebly an electronics issue.
what do You think?
I believe it's a firmware issue. Prusa added to Marlin a defective algorithm that expect too much from a cheaply built CORE X/Y mechanism with sheet metal, skinny pelts, plastic parts and non end sensors. That algorithm is the cause for the massive print head banging and failures and all that is Prusa need to do is to disable it and use a standard sensroless homing (one bang on X and one on Y, with optional verification of free movement in the expected X and Y ranges).
RE:
Prusa's invented a faulty algorithm and we are all busy developing black magic about it.
Prusa is failing its users by not reverting it.
That is too black & white for my taste.
Determining the position that precisely is not a trivial problem to solve -- especially when you must measure distances by counting steps at the far end of a slightly springy belt with a 2 mm tooth pitch. I am not convinced at all that using a limit switch or an optical sensor instead of the Stallguard sensing would make it any easier.
Determining the position that precisely is also not a pointless exercise. While you and me have no problem with power outages, there are many areas (also in developed countries) where blackouts or brownouts are common enough to be an issue.
What I hold against Prusa is that they are stubbornly refusing to allow the user to select whether they require precise homing, or prefer to forfeit the power loss recovery in fgavor of fast & quiet homing. Granted, they should not add more and more option switches in response to every little request or issue. But ignoring four months of very vocal and common user complaints, while neither being able to improve the measurement nor willing to make it switchable, is disappointing.