Think I tried this yesterday with sub 1mm results, but I’ll give it a shot again and report back.
Think I’ll also try a little more angle on the frame and better supports on the bottom — it’s currently on a hinge with some 2x4s wedged in the bottom to kick it out so maybe they’re allowing some flex. Also heard that 15° might not be enough.
Yeah, sub 1mm on the extend and retract.
Running another calibration now with sturdier bottom supports and about 17° angle — I think I had incorrectly measured the surface angle previously, so it was probably around 10-12°. I could see those two contributing to the 8mm deviation in BL, so fingers crossed!
Using slightly different values for the initial guess for the width and height could help too…just a way to shake things up a little
Kyle wrote:
Think I tried this yesterday with sub 1mm results, but I’ll give it a shot again and report back.
Think I’ll also try a little more angle on the frame and better supports on the bottom — it’s currently on a hinge with some 2x4s wedged in the bottom to kick it out so maybe they’re allowing some flex. Also heard that 15° might not be enough.
a frame that flexes is a problem
take a look at http://lang.hm/maslow/maslow4_frame.html it’s not difinitive (you
don’t lose all accuracy by going outside the green, but the further you go
outside the green, the more problems you are likely to have, and it’s
non-linear, going 2x as far outside the green may be 2x the problem or it could
be 10x the problem, it’s a nasty curve)
David Lang
Unfortunately same results, about a -8mm and change on BL. It says that sub 15mm is within spec, is this reasonable for the BL belt? It’s so floppy it seems kinda useless.
Next two tests I’ll likely run are:
Kinda stumped otherwise 
@dlang No flex in the frame whatsoever now.
Kyle wrote:
Unfortunately same results, about a -8mm and change on BL. It says that sub 15mm is within spec, is this reasonable for the BL belt? It’s so floppy it seems kinda useless.
if a extend/retract is giving results around 8mm, there is something wrong, it
really should be <1mm
the 15mm error is a limit that when the machine sees the belt lengths different
from the calculated lengths it is to the point where it can’t figure out what’s
going on as the encoders are off by > 1/2 turn (approximately). That is a very
different error.
Next two tests I’ll likely run are:
- Increase tension to 1500 then calibrate, as @bar mentioned before.
note that you need to increase both the retration and calibration tension
limits.
David Lang
A full extend and retract is sub 1mm. If I jog the machine then hit apply tension it shows -8mm+. Or if I just hit apply tension once or twice it also shows -8mm plus.
Guys I got it! I deleted and uploaded the maslow.yaml file from the most recent firmware, recalibrated, and the problem went away. I think I might not have realized that the yaml file needs to be replaced when upgrading the firmware, so I think it was trying to run a v0.74 yaml file on v0.83 firmware. That or just completely resetting the file did the trick. Either way, all of the belts stay tight now and looks great when it jogs. Excited to make some real cuts now.
THANKS for all the help!
Thats great news!
Interesting development for @bar and @anna, had the machine crash a few times and the loose axis on jogging issue came back, but on the BR this time. Replaced the maslow.yaml file modified with my calibration specs and that fixed the issue. I scanned the file with the stock firmware file and nothing was different. So must be something deeper in the software thats getting corrupt but resets when uploading the maslow.yaml file.
Just to keep the community informed, i’m running into same kind of problems.
I’ll start overt with fresh calibration ans Maslow file to check if anyway get better.
It’s a BR slack problems with 3000 cal treshold , and think that my cut is getting lower quality ( small dog bone cut in the middle of the frame
Jonathan Dion wrote:
It’s a BR slack problems with 3000 cal treshold , and think that my cut is
getting lower quality ( small dog bone cut in the middle of the frame
If you are needing to go up to 3000 to retract the belts, that’s a strong
indication that something is wrong with that arm.
David Lang
For calibration only, i get better fitness score at 3000.
Retract is going very well at 1500.
What results do you get if you hit apply tension two or more times in a row?
Hi everyone,
this is my first post here. After years of watching the Maslow-Project I finally got the time to get the Maslow4. Everything went fine, but I have one problem I hope someone could help me.
After the calibration all belts are tight, but after jogging around the BR-belt gets loose.
I use firmware 0.84 and replaces the maslow.yaml with the latest version.
I am a bit confused, because if I tighten the belt by hand, it stays tight while jogging the machine. It seems as if the sensor works, but the length of the belt is not correct.
I tried several firmwares, several scenarios of updating, uploading, restarting and calibrating, but every time I got the same problem. Only once the BL-belt was loose.
The extend / retract seem to work fine, I do not thing it is a problem of the frame (wooden one, like in the setup-guide). I disassembled the arm, but I have not found any suspicious there.
Where should I look to solve this problem?
Greeting from Bavaria
Franz
Here the latest log-file:
Maslow-serial (4).log (21.6 KB)
Possibly something wrong with your calibration setup/process. Are you calibrating with no bit in the machine and the Z axis all the way down? What type of corner anchor setup do you have? Any movement in the corners can mess with the calibration.
Another thing to check is to tension the machine, as posted above get the machine homed, then hit the tension button a couple of times and post the offset numbers here. That might help along with some pictures of your frame setup.
Z-axis is down, no bit. Here are some pictures of my setup:
Nice lookin frame! Well… except for the MDF board, looks like it’s pretty beat up in some areas.
Few things to check:
My guess is the frame is too vertical and the warped MDF might be skewing the calibration results. Also watch when it’s tensioning the belts to see if any of the corner anchors move around, that can mess up the results too. I don’t have the ability to map your calibration report, maybe someone else can and provide some extra insights. Only chiming in with fixes from what I’ve encountered.
Thank you, I will need some time for testing and will tell the results.
Hi,
But after all, I´ve got the same problem. I tried to retract and loose serveral times, but that won´t fix it. When I run a gcode BL get´s loose, too.
I´m a bit confused about this line:
[MSG:INFO: Center point deviation: TL: 0.125 TR: 0.214 BL: -1.581 BR: -6.961]
Every time BR has a difference -+5mm. This is about the length the belt is to loose.
Or do I clutch at straws?
