I recently re-built my 4.1 with turtle clamps, sanded the spools which were binding and a few other things. I upgraded to 1.21 and re-calibrated and I’m seeing a max fitness of 0.71 whereas previously I remember fitness numbers as high as 1.50.
What is considered a “good enough” fitness value these days? I saw there’s a threshold of 0.45 for cancelling calibration, but not sure what the high-end of accuracy looks like.
Zach wrote:
I recently re-built my 4.1 with turtle clamps, sanded the spools which were
binding and a few other things. I upgraded to 1.21 and re-calibrated and I’m
seeing a max fitness of 0.71 whereas previously I remember fitness numbers as
high as 1.50.What is considered a “good enough” fitness value these days? I saw there’s a
threshold of 0.45 for cancelling calibration, but not sure what the high-end
of accuracy looks like.
“good enough” is “does it do the work you are trying to do” for some people
being 10mm off is not a problem (theater scenary) for others 0.5mm is
unacceptable.
the fitness is an estimate of how accurate the machint thinks it has the anchor
locations, it’s “1/estimated error across all points”
However, it’s not a direct measure of accuracy, if you do a run over a small
error, the accuracy of the estimate is not going to be as good as if you run it
over a larger area, even if the fitness number over the small area is higher.
We’ve had a lot of people do a run over a small area in the center and get a
high fitness score, but if they do the same thing over a larger area, the score
goes down. That doesn’t mean that the machine is less accurate, just that the
initial run over the small area didn’t properly show the limitations.
0.71 is a pretty decent result, try it and see if it works for your type of
work.
David Lang
Appreciate you taking the time, this is super helpful!
I’ll go ahead and give it a shot and see how it goes. 
Update:
Turns out the spools were still binding despite sanding. With the fitness of 0.71, circles came out super crooked and straight lines wobbled. I got several warnings that motor force exceeded 4000 on multiple belts and cancelled the cut.
Rebuilt with brand new spools & fresh belts from the shop. I used 14mm screws so they would engage the nyloc and backed them off so they weren’t too tight. I made sure the spools were spinning freely and verified that they retracted well with retraction force of 600. Started a fresh calibration. Everything looked fine, then the machine snapped a belt mid calibration…
I’m not sure what to do from here. I’m lost as to why this would happen during calibration. Quality control issue with the belts? I didn’t see any motor force warnings, so that would be my only guess. Anyone have ideas?
I’m sorry! I’m not sure what could have caused that to happen. Any chance that you have the log file from it? That should help us to diagnose better.
I will DM you a code for a free belt replacement.
I’m also not sure. Unfortunately I yanked the plug since the machine was still moving on 3 belts and rage quit my browser. I think that means I lost the log file - let me know if there’s a way to recover that after reboot.
One thing I did notice was that the fitness dropped significantly the round before the belt snapped. After the first round it was ~1.5, second round it dropped to ~0.8, third round dropped to 0.47, then it snapped part way through the 4th round. From what I understand some amount of fitness drop as points are added is normal behavior though, so I’m not sure if that is a red flag or not.
Really appreciate the free belt replacement - I’ll get that spool rebuilt and see if that resolves it.
Where did it break, if it’s close to the end you can re-terminate the belt end. You can afford to lose a reasonable amount of belt. If it broke too far in, then you will need a replacement.
Good advice! It looks like about 50" of belt snapped off. My guess is that’s too much to keep running a 4x8 work area (8x12 anchors), but I’ll have to do some math to be sure on that.
Interestingly it looks like the break happened right at the spool exit. Doesn’t look like a sharp/clean cut though, more of a rip.
Set the retract force to a low value 200, extend and retract (it shouldn’t retract) until all the belt is off the spool and then check to see if it can reach the anchor point with the Maslow as far away as possible. No maths needed.