Calibrating my belts for the first time since building on latest software. When I press retract, the fan starts for a second but then turns off.
I also noticed the z seemed to move quite slowly and quietly compared to the calibration video. Not sure if that is a problem or not?
I’m guessing the machine is sensing tension but I can move my belts in an out through the entrance hole of the arm easily.
Could anyone offer any suggestions?
Thanks in advance,
G
Those later offsets are way too high. Have you tried increasing the threshold current?
Sorry, I’m not really sure what that means. I’m following the YouTube guide and this hasn’t seemed to have come up. Is it a software or hardware issue do you think?
click/tap setup button, then click/tap config here:
Then change the retraction force here:
and save, then try retract again.
Ron Lawrence wrote:
click/tap setup button, then click/tap config here:
Then change the retraction force here:
and save, then try retract again.
does that change both retraction and calibration force? or just retraction
force?
David Lang
just retraction. You have to go up in the fluidNC menu to change the calibration force.
Derek Graham wrote:
just retraction. You have to go up in the fluidNC menu to change the calibration force.
we should probably change this to do both from this UI, if you need the higher
current for retraction, you probably need more for calibration (although since
you are not fully filling the spool, you may get by with a little less)
probably should be combined with a frame flex test.
David Lang
So what do we think I should change that value to?
I’d start with 1800 and work up by 100 until all the belts retract fully. If the number is over 2200 I’d try a few cycles of extend / retract and try setting it down again. From my experiences with this, it will loosen up. I think when I put together my most recent arm and put it on I had to go up to 2200, but now I’m set at 1800 with no extension issues any more.
Regarding the retraction force value, I am curious how that translates to milliamps. Anyone know?
Cat6 can only carry so much power and not sure if these cables are shielded and I was starting to wonder if these ethernet cables for powering servos are causing or attracting EMI.
I’ve assumed it is milliamps. Isn’t the max current for the motors 4000 where they will alarm or something if exceeded?
Maslow #3 yesterday couldn’t get one of the arms to fully retract. Finally worked at 2400. After that I backed it down to 2200, then 2000, then 1800 after an extend/retract at each level.
Maslow#2 I upgraded to 0.74 and tried to re-run calibration. Its retract current had been the default 1300 and had calibrated previously on 0.73. I need to look when I get on-site today, but I may have upped the calibration current from that because after ~30 minutes of cutting it was stopping itself from what seemed like loose belts. Hence new firmware and upped the retract current.
Anyway, it failed calibration. And one thing we noticed was a creaking during calibration that hadn’t been present during the first calibration. Last week we heard more of that on Maslow#1 but attributed it to a different frame. Now I’m wondering if it was the fact that Maslow#1 was on newer firmware with me setting calibration current to match what was required for retract all.
Seeing how the belts wind, especially when they are too long, I realize that additional current may not be necessary for calibration beyond the bump from 1300 to 1500 that happened in 0.73 because the belts are not fully retracted during calibration. I’ll have notes later on that.
If the drive gear assembly bearings are not seated all the way down the drive gear can clip the opposing bearing. Not saying its the cause of the creaking but its a potential cause of noise and depending on the wobulation of the plane based on cutting forces the creaking from gear clipping the bearing may be intermittent.
Regarding the variability of the retraction current levels across your maslows and across arms, I think we could use a test capability to ramp up current so we can give a health score to an arm that indicates inherent resistance differences due to manufacturing or assembly inconsistencies. I think @dlang mentioned this before…
I don’t know but 4amps down cat6 cable seems like a lot of current for cat6…
And I noticed our M4 cat6 cables are unsheilded (UTP) <— replacing these with shielded cat6 cables may be something to explore
Changing that value to 2400 seems to have fixed that problem. Onto the next! Thank you guys!
gazinux wrote:
I don’t know but 4amps down cat6 cable seems like a lot of current for cat6…
the motor current doesn’t go through the cat6 cable, that goes through a pair of
dedicated wires.
And I noticed our M4 cat6 cables are unsheilded (UTP) <— replacing these with shielded cat6 cables may be something to explore
yes, that would help
David Lang
I am going to buy this 5-pack of 1ft shielded cat6:
It’s not mA, it’s just reading from the ESP32 ADC. We could convert it if we wanted to though.
The power for the motors doesn’t actually flow through the Ethernet cable (for exactly that reason)
That’s a great find. Let us know if you can tell the difference!
I am a month or so away from being able to finish my assembly and test… Hoping others also try those shielded cables…
