RE: Tension pulley broken
yeah, gantry was square before I started tensioning. Actually there's now a small amount of play on the right side. Original it was on the left. might have to unhook pulleys and try the bending thing again.
RE: Tension pulley broken
ok without the tensioning being done, and the belts pretty lose, the gantry is perfectly square. I'll mess with this some more tomorrow. The y calibration actually seemed to behave better but still both x and y ultimately fail calibration. Tomorrow I'll try the tensioning again and see.
RE: Tension pulley broken
ok without the tensioning being done, and the belts pretty lose, the gantry is perfectly square. I'll mess with this some more tomorrow. The y calibration actually seemed to behave better but still both x and y ultimately fail calibration. Tomorrow I'll try the tensioning again and see.
Good luck. Slow and steady - aim for the 6Hz difference as early as possible, once the belts start registering at, say 50Hz, then just gradually increase each one until you get to 98/92. I also make sure that I move the print head around a bit after each tension adjustment, just in case there's any residual friction somewhere that could lead to an incorrect measurement.
RE:
ok without the tensioning being done, and the belts pretty lose, the gantry is perfectly square. I'll mess with this some more tomorrow. The y calibration actually seemed to behave better but still both x and y ultimately fail calibration. Tomorrow I'll try the tensioning again and see.
Good luck. Slow and steady - aim for the 6Hz difference as early as possible, once the belts start registering at, say 50Hz, then just gradually increase each one until you get to 98/92. I also make sure that I move the print head around a bit after each tension adjustment, just in case there's any residual friction somewhere that could lead to an incorrect measurement.
In my case, it's quite odd. Using the older tensioning method (print head centered and to the back) I can easily get both left and right belts at 85hz. If I try the "new" method (print head front and to the right), the left lower belt is always 6hz higher than the left upper belt. No adjustment can get the belts at the right frequency. If I tune the upper belt to 98hz then the lower belt is 104hz and if I set the lower belt to 92hz the upper belt is 86hz. There is always a 6hz offset. I'm sure I was able to get the correct tension a couple of weeks ago.
The only change is recently swapping out the tensioner pulley with a version that has the heatset insert instead of the the nut. Not sure why this would cause this issue. It's printing perfectly fine on the "old" setting although I haven't checked if the VFA's are back or not.
The other issue I have is the gantry is no longer square so I must be doing something wrong (it was perfectly square when the pulleys were slack). It's square again now that I've tuned the belts to be the same using the "old" method.
RE: Tension pulley broken
In my case, it's quite odd. Using the older tensioning method (print head centered and to the back) I can easily get both left and right belts at 85hz. If I try the "new" method (print head front and to the right), the left lower belt is always 6hz higher than the left upper belt. No adjustment can get the belts at the right frequency. If I tune the upper belt to 98hz then the lower belt is 104hz and if I set the lower belt to 92hz the upper belt is 86hz. There is always a 6hz offset. I'm sure I was able to get the correct tension a couple of weeks ago.
The only change is recently swapping out the tensioner pulley with a version that has the heatset insert instead of the the nut. Not sure why this would cause this issue. It's printing perfectly fine on the "old" setting although I haven't checked if the VFA's are back or not.
The other issue I have is the gantry is no longer square so I must be doing something wrong (it was perfectly square when the pulleys were slack). It's square again now that I've tuned the belts to be the same using the "old" method.
If the gantry is now square with the belts both measuring 85Hz in the 'old' position, it should be a short step from there to the 98/92 setting in the 'new' position. With equal tension in the old position, you should already have a difference of ~6Hz between top and bottom in the new position. You should be able to slowly and evenly increase each belt, keeping 6Hz between them, until you get to 98/92.
If you get to the point where adjusting one tensioner changes the tension in either both belts or neither, it probably means that the tensions got too far apart, and a gap has opened up on one side or the other. If so, it's best to slacken them both off and start again. And make sure that the gantry is square before you start, when there's no tension on either belt. This will give you the best chance of getting it right.
RE:
>I'm curious what software you used to tidy up the lettering. I notice that you've removed the bearing and spindle screw spacers - did the extra play in the lateral movement of the bearing cause any issues?
@chris - Yes I removed the spacers from the STL as my spare MK3 is not as dimensionally accurate and the spindles were too tight. I had a bunch of super thin washers I would have used if here was play in the spindle movement. I used Sketchup to modify the pulley STL. I scaled it up 10 times to easily manipulate some of the finer points.
Also for some reason I can't attach the .STL the post.
RE: Tension pulley broken
Also for some reason I can't attach the .STL the post.
If you zip it, you should be able to attach the zip file.
RE: Tension pulley broken
Try raising your bed.. Mine has issues homing when the bed is low...<shrug, dunno why>
well this is disappointing. Got things back together. Gantry is square as far as I can tell. x and y belts have been tuned to 98 and 92 respectively. It still won't pass x and y calibration.
It moves to the far right then does a series of homing motions that that front right corner. It does this for a minute or two before the machine resets and it gets the red screen of death. On the bright side, the lubricated screws in new parts seem ok... I'm really not sure what's wrong.
RE: Tension pulley broken
@chris I posted my mod to https://www.printables.com/model/1378250-core-one-tension-pulley-with-heat-insert-clean
RE: Tension pulley broken
In my case, it's quite odd. Using the older tensioning method (print head centered and to the back) I can easily get both left and right belts at 85hz. If I try the "new" method (print head front and to the right), the left lower belt is always 6hz higher than the left upper belt. No adjustment can get the belts at the right frequency. If I tune the upper belt to 98hz then the lower belt is 104hz and if I set the lower belt to 92hz the upper belt is 86hz. There is always a 6hz offset. I'm sure I was able to get the correct tension a couple of weeks ago.
The only change is recently swapping out the tensioner pulley with a version that has the heatset insert instead of the the nut. Not sure why this would cause this issue. It's printing perfectly fine on the "old" setting although I haven't checked if the VFA's are back or not.
The other issue I have is the gantry is no longer square so I must be doing something wrong (it was perfectly square when the pulleys were slack). It's square again now that I've tuned the belts to be the same using the "old" method.
If the gantry is now square with the belts both measuring 85Hz in the 'old' position, it should be a short step from there to the 98/92 setting in the 'new' position. With equal tension in the old position, you should already have a difference of ~6Hz between top and bottom in the new position. You should be able to slowly and evenly increase each belt, keeping 6Hz between them, until you get to 98/92.
If you get to the point where adjusting one tensioner changes the tension in either both belts or neither, it probably means that the tensions got too far apart, and a gap has opened up on one side or the other. If so, it's best to slacken them both off and start again. And make sure that the gantry is square before you start, when there's no tension on either belt. This will give you the best chance of getting it right.
I do have a difference, however it's inverted. Currently both left and right belts with the head in the center rear are at 88hz. When I move the head to the front right, the top belt is 75hz and the bottom belt is 86hz. No matter how I adjust it, with the gantry at the front and nextruder to the right, the bottom belt has a higher frequency than the top.
RE: Tension pulley broken
Also for some reason I can't attach the .STL the post.
If you zip it, you should be able to attach the zip file.
Voila, one with spacers, one without.
RE: Tension pulley broken
I do have a difference, however it's inverted. Currently both left and right belts with the head in the center rear are at 88hz. When I move the head to the front right, the top belt is 75hz and the bottom belt is 86hz. No matter how I adjust it, with the gantry at the front and nextruder to the right, the bottom belt has a higher frequency than the top.
I did manage to get it "in tune" (98/92hz) however the gantry is too skewed on the right side by almost 1cm. So, in my case, I don't see how this is possible to achieve this. My guess is the upper belt is slightly shorter than the lower? For now I'm going to stick with 85hz.
RE: Tension pulley broken
I did manage to get it "in tune" (98/92hz) however the gantry is too skewed on the right side by almost 1cm. So, in my case, I don't see how this is possible to achieve this. My guess is the upper belt is slightly shorter than the lower? For now I'm going to stick with 85hz.
That's strange. 1cm is a huge gap.
I don't see how belt length can have any bearing on this, since the role of the tensioners is to change the belt length until the required tension is achieved. The belt teeth are 2mm apart, and by virtue of the 180 degree loop around the tension idlers the tensioners only need to move by 1mm to change the belt length by the equivalent of 1 tooth. With the 0.5 thread pitch of the M3 tensioner screw, that means two turns of the screw will take up 1 tooth of the belt. And there's more than 10 times that available in the tensioner range. I believe Prusa's advice for 'extensive adjustment' ( https://help.prusa3d.com/article/adjusting-belt-tension-core-one_845048), where they describe adjusting the belt length, is for cases where the full range of adjustment has been used up (i.e. an adjuster has reached its end stop) before being able to achieve the correct tension.
I can't really see too many reasons why your belt tension is behaving like this. My best guesses:
1) Your gantry wasn't square before you started, i.e. before there was any belt tension. If that was the case you'll always be fighting that distortion with belt tension, so I feel you're bound to end up with either a skewed gantry or mismatched belt tensions. Keep in mind that the 'new' method is just a new and better way to measure the tension - the goal (or at least the result that I've observed) is to keep the belt tension the same between both belts when measured at the 'old' locations. The new 98/92 values do, however, result in a higher tension than before on both belts (~100Hz?) when measured at the old location. Even though you're getting 85Hz on either side in the old locations, it's possible your gantry is already skewed at this point.
2) Perhaps there's an issue with the reliability of the frequency measurements on your phone. Are you getting clean and reproducible frequency measurements? Is your phone perhaps one of the ones that's been identified as using noise filtering? Which app are you using? Have you tried the 'Carbon Drive' app from Gates (the belt/pulley manufacturers)?
3) It's worth saying again that if you get to the point where adjusting one tensioner seems to affect either both belts or neither, it's an indication that a gap may already be opening up on one side. With a properly squared gantry at the start, the main reason for this to happen is that the adjusters weren't kept in sync - one side was perhaps allowed to get too far ahead of the other one.
My money is on your gantry not being square at the start (even if it was, after applying enough tension to open up a 1cm gap it likely isn't any more). I'm sure you've seen the procedure for squaring it - it involves using a spacer on one side so you can pull the other side hard enough (with belt tension removed completely) to bend the brackets until the gantry is squared up.
RE:
@chris - Yes I removed the spacers from the STL as my spare MK3 is not as dimensionally accurate and the spindles were too tight. I had a bunch of super thin washers I would have used if here was play in the spindle movement. I used Sketchup to modify the pulley STL. I scaled it up 10 times to easily manipulate some of the finer points.
Interesting - I haven't used Sketchup in a very long time. Filling in the de-bossed test on a curved surface would be a non-trivial task in OpenSCAD, since it's not really aimed at modifying existing STLs.
Just a thought on your model - in your Printables listing that insert does look a lot like the one I designed my model around. Are those dimensions (5mm diameter, 6mm length) accurate, or nominal/approximate? With callipers mine measure 4.6 x 5.7. Is it possible yours are actually the same? Also, if your pulley didn't fit into the model with the spacers, both of these things suggest that your printer was slightly over-extruding, rather than being inherently inaccurate. Did you use the built-in profile for PC-CF? I don't know what the default extrusion modifier for this is on the MK3, but on the CORE One it's 1.04 for some reason, and this certainly does lead to over-sized prints. I had to reset that to 1.00, and could have got away with even lower. I'm really curious whether your model is simply correcting for this over-extrusion - I know there's another model on Printables that seems to be doing just that - the one with greater clearance for the original square nut.
RE: Tension pulley broken
As a reference point, I printed Chris version 2 with Prusa PC-CF blend, o,15mm structural, 0.6mm standard obxidian nozzle, not high flow, 100% infill and extrusion modifier 1.0. I may have bumped up perimeters as well but need to check that, I am not sure it matters with 100%infill. I used the CNCKitchen M3-5,7 which is 5,7x4,6mm. This is tight but fits perfectly, it takes a bit of effort to push in place but with patience it works. There is sufficient room for idlers to run free without grinding plastics.
/Anders
RE: Tension pulley broken
@chris-hill
How is the ABS holding up mate ? Any signs of creep or too early to tell ?
RE: Tension pulley broken
@chris-hill
How is the ABS holding up mate ? Any signs of creep or too early to tell ?
Still looking ok. I did some ASA prints yesterday and overnight, which have a slightly higher bed temperature than ABS. There may be a 1Hz drop in both belts today, but that's within the measurement error, so I'm not sure. I'm not going to re-adjust the tension - I'll leave them for a while longer and see what happens. But first impressions are that plain old ABS is behaving exactly as my PC-CF parts did. I do still want to do another VFA tower though, to see if they've changed anything.
RE: Tension pulley broken
@chris-hill
How is the ABS holding up mate ? Any signs of creep or too early to tell ?Still looking ok. I did some ASA prints yesterday and overnight, which have a slightly higher bed temperature than ABS. There may be a 1Hz drop in both belts today, but that's within the measurement error, so I'm not sure. I'm not going to re-adjust the tension - I'll leave them for a while longer and see what happens. But first impressions are that plain old ABS is behaving exactly as my PC-CF parts did. I do still want to do another VFA tower though, to see if they've changed anything.
Yeah I think long term I'm going to have to reluctantly invest in some PCCF, because I sure as hell aint paying Prusa shipping prices when I'm eventually out of warranty. It's just annoying that PCCF is so expensive across the board, it's not just Prusa PCCF.
If anyone is aware of any PCCF that isn't priced on par with cocaine/gold per Kg (and been pleased with the results) then please let me know.
RE: Tension pulley broken
42$/1Kg Spool. I printed a few Core One mods, and spares w/ it. I can't say how satisfied I am yet, as the mods have only been in the printer a week, and haven't used the spares, but they seem to be adequate.
RE: Tension pulley broken
>Interesting - I haven't used Sketchup in a very long time. Filling in the de-bossed test on a curved surface would be a non-trivial task in OpenSCAD, since it's not really aimed at modifying existing STLs.
Sketchup can import STLs and clean them up nicely.
>Just a thought on your model - in your Printables listing that insert does look a lot like the one I designed my model around. Are those dimensions (5mm diameter, 6mm length) accurate, or nominal/approximate? With callipers mine measure 4.6 x 5.7. Is it possible yours are actually the same?
The measurements are taken from the packaging; I then measured them with a micrometer to be 5.018 x 6.012 with the smooth diameter top of the insert being 4.227
Yes, correct. The extrusion modifier was 1.05 I think. Also I dried the PA-CF for 24 hours and it still had some anomalous layering where there were artefacts. (I scraped them off with a Xacto blade).
Once I got the Core ONE working with the new tension pulley I printed the model again and it came out *perfectly*, however, no way am I going to take it apart. If it ain't broke....
>Did you use the built-in profile for PC-CF?
I had eSun PA-CF and I got the profile from internet sources with some slight mods. I figure PA-CF is just fine to use.