If the belts won’t retract with the lower force, then calibration won’t work with the lower value either.
Makes sense. When unspooled fully they will retract at 900 but after that I have to switch to 1200 so they all retract at the same time
Was able to complete maxed out grid calibration @1200 force. Will run a cut and report on accuracy.
1 Like
Well, the sled is still lifting up in corners and messes up the cut. I think its due to bottom belts being too tight. I’m out of ideas…
The only thing Ive noticed is that left lower belt is kind of getting slightly stuck in the spool after I stopped the cut, and I have to increase force to like 1700 for it to retract it. So maybe all my issues are still related to those spool/gears/arms not having right tolerances. 900 will only retract belts from 100% extension.
Ian Abbott wrote:
If the belts won’t retract with the lower force, then calibration won’t work with the lower value either.
almost, full retraction with a full spool has the belt rubbing against the
idler, you don’t pull in as much belt when doing the calibration
but as a rule of thumb, you are correct.
David Lang
I was able to calibrate with 850 and retraction of 1200 (5x5 500x500 grid), although I dont think its calibrated at all.
maslow-cal-june18.yaml (5.4 KB)
Are the anchor points need to be on the same plane as the work piece or a spoil board?
Mine are currently at the spoil board height.
I can try messing with the frame angle but I feel like 25º is pretty adequate. I will do some more measuring to check for twists or any other deviations.
1 Like
How much distance is there between the anchor points and where you are cutting? The sled will tip more as it gets close to the anchor points just because of the angle of the belts
About 90mm. Right side perform fine, its only the bottom left that is struggling.
90mm is way less than it should be able to handle. We recommend 12-18 inches which is more like 300 - 450mm of space there
Although you just DM’d me a picture that makes me think this is a miscommunication because in the picture you are way more than 90mm away from the anchor point
Sorry 90cm
1 Like
Ah yeah that is a totally reasonable distance 
UPD:
Spend many hours going through each arm, sanding, and making sure there absolutely no binding anywhere, replaced the screws around the gear with ones that have smaller cap (as suggested by @RandomDave here What a difference 1/10th of a millimetre makes (how I got retraction to work at force 3/4/500)). Got it to retract at 300. Also found a broken magnet gear:
Calibration went ok:
I tried jogging it around, seems better, less stressed. There is still a small lift up of the sled when going from left corner upwards (right side is fine).
Is there a different vacuum shroud design that won’t cause the belt twist?
I don’t think it’s the cause of the lift, but still not ideal.
I also measured the exact offsets for the belts, I think thats how it’s calculated based on what people have suggested.
I’m going to do a test cut to see where things are.
1 Like
I’m not sure on dust port design but I know you can change order of arms. The corresponding z height offsets would need to be adjusted to match.
Dano
Ok I might try that to see if changes anything. I’m about out of ideas how to fix this lift/tight belt issue on the left side. I tried mix and match the arms, same thing.
So I think my issue is within calibration process itself.
I’ve tested everything, modified what was modifiable (sanded arms etc), frame is solid, retraction/extension is nominal.
Right after first calculation things go south, you can hear the motors start to struggle more (increased tension) and once it moves to the bottom and move upwards — things start to look bad (thats all during calibration I’m talking about). Left side being much worse than right. I wiped the yaml file and started fresh. At this point I’m ruling out any hardware problems. Something happens during that first run of numbers and gets worse after that.
maslow (2).yaml (5.4 KB)
Any input will be greatly appreciated.
This makes some sense. After that first round of calibration the machine is switching from moving using only the top two belts to moving using all four belts. If the first round is finding bad numbers it would do something like that.
The question is why is it getting bad numbers in the first go around.
Have you tried measuring your frame and manually entering the values?