Fitness is really giving me a headache! Would lowering my “Z” before calibration help?
1 Like
Yes!
The calibration process depends on the z-axis being all the way down! 
1 Like
That may be the issue!!
1 Like
If calibration isn’t accurate if it’s done with the Z axis up I feel like it might be worth automatically putting that down before calibration or blocking starting calibration until it is lowered. Thoughts?
1 Like
I like that idea. The tricky thing is that because we have stepper motors on the z-axis instead of servos we don’t have any feedback to know if it is actually down or not.
I’m super open to ideas on how to implement something like that if we can find a way.
Interesting. I feel like Z-axis limit switches or a servo would make sense on future hardware. I wonder if that could be done via a probe that could double as an auto z-home finder. Wouldn’t tell you where the bottom of the bit was without some calibration, but could give a known spot for the bottom of the z-axis.
For now, I think not actually enforcing this, but just adding a popup menu reminding the user to lower the Z-axis and giving a chance to cancel starting calibration to go do that would get 90% of the way there. As a brand new user, I wouldn’t have guessed from the software UI that the Z axis needed to be in a specific position for calibration.
2 Likes
That is a great point. A popup seems like a pretty easy way to get that confirmed.
2 Likes
You could run it right up until it is off the screws, to get to a known position, then bring it all the way down. I think a popup to remind you to take the bit out and wind Z down would be a better option. Would also be nice if you could press stop and halt a calibration, if it looks like it’s going to fail or if you want to adjust something mid calibration.
1 Like
Can you not press stop? That seems like an issue for sure!
Do you want to make an issue on github to assign that one to the AI? I feel like it should be able to figure that out
pressing stop stops it, then it restarts in calibrate mode again and continues. only way out is power off and then it’s in state unknown.
1 Like
Ian Abbott wrote:
You could run it right up until it is off the screws, to get to a known
position, then bring it all the way down. I think a popup to remind you to
take the bit out and wind Z down would be a better option. Would also be nice
if you could press stop and halt a calibration, if it looks like it’s going to
fail or if you want to adjust something mid calibration.
off the top is not reliable, the motors can turn a fair bit before engaging the
threads to start moving down and you don’t know how many steps that is.
Currently, the known position is to go all the way down. That is far more
reliable (and evens out the two sides)
you could just drive down for 80mm, but if you are all the way down, this is
going to make a grinding sounds as the steppers miss steps for quite a while.
David Lang
David Lang wrote:
you could just drive down for 80mm, but if you are all the way down, this is
going to make a grinding sounds as the steppers miss steps for quite a while.
This would also cause problems if there is a bit in the machine.
David Lang
I’ve wondered if it would make sense to mount limit switch(es) at the top of travel. Somewhere that’s about 90% up the Z travel, but in a way that’s easy to move for maintenance?
Initial setup - the user could be told to lower to the bottom of z travel and it would auto-set Z-stop pre-calibration, and could sanity check Z-stop by calculation then and before each cut (something that’s tripped me up when messing with hardware).
1 Like
Turns out my z is at the bottom. How do I make changes to my z settings?
Anyone near B’ham Alabama have one of these?
Not sure if I can help or not but I’ll give it a shot. Are you vertical or horizontal? What is going on with it and what have you done so far.
It just keep failing fitness. Vertical application.
Ok. I’m vertical too. So is your frame the default wooden one in the docs or a custom version? Here’s a blast of questions, what’s the firmware and index version you’re running currently, what’s your extend distance, retract force, calibration force, grid size, windows op or something else? WiFi issues?
1 Like
I have never passed calibration, I gave up and entered numbers in by hand. Has worked great for months.
Out of curiosity, I retried recently, got stuck in a loop. If you are careful building your frame, and have someone help you measure the actual mount points afterwards, the math is easy.
Had a (re)calibration issue the other day related to Z-axis settings:
I originally did a 1000x2000 calibration with a 9x9 grid. To do that, I put the sled directly on the floor of my workspace, removed the bit, and lowered the Z axis all the way. Calibration went great and a test square in 1/2” ply had equal length sides as best I could measure. The next day, I went to make a larger cut. I layered a sacrificial 3/4” board plus the 3/4” work piece, then increased the z-home height to fit a longer bit.
On the taller work piece with the higher z-home, the machine refused to tighten slack in the belts, saying that belt lengths were more than 12mm off from calibration. IMO this is expected - given a 50mm+ z change, it would make sense that the belt lengths would all change, though 12mm feels a bit much. I tried removing the work pieces + bit and pulling in slack - worked like a charm with no error. Tried again just moving the z-axis up half way to the top and tightening slack - that failed.
My question is: what should I do to avoid this going forward? I can imagine either:
- Calibrate with the Z-axis roughly at the expected level for cutting - this feels a bit hacky, and from what I understand may compromise the calibration quality. Also means re-calibration might be required for much taller/shorter pieces + bits.
- Update the belt length check - either update that code to account for consistent increases in belt lengths due to z-changes, or turn the belt length error into a warning so it can be bypassed. I’d be up for trying to contribute this, but would love to know what’s been tried here before.
- Make pins for my setup where z-height of the belt ends can be adjusted to always be approximately flat. This could be workable, but feels like it could be unnecessary.
Thoughts?
1 Like
I have an excellent solution 
. This bug is fixed in 1.16 so if you update it should be solved! The z-axis height was not being taken into account by the “apply tension” measuring system but now it is.
Someone else reported the same issue so we fixed it. Thank you for pointing it out so clearly, it’s exactly this kind of thing which is really helpful for finding bugs like that!
1 Like