I just got my Maslow 4.1. Got it assembled and now I’m trying to get it setup and calibrated. Here is the situation.
I’m running in horizontal orientation with anchors in a concrete garage floor. The anchors are laid out to cover about 13 ft x 9.5 ft. The machine is calibrating on a 0.5" spoilboard + 0.5" workboard, and I have both thickness settings set to 12.7 mm. I’m using an extend distance of 2470 mm. For calibration I’ve been doing it with the router bit removed and the Z physically all the way down. The force setting is at 1000. I have the latest 1.20 (firmware, maslow.yaml, index.html.gz, and fluidnc) installed.
I have tried many cycles of: power off, power on, retract all, extend all, find anchors. It doesn’t end up with a fitness, that it keeps, that is beyond 0.45. After waypoint 5, it does get a fitness >0.45 (usually about 0.5), and it says it sent a command to save that value.
But it immediately automatically does find anchors again, starting at waypoint 6, and that pass has failed every time with a max fitness between 0.35 and 0.4 after waypoint 14.
I’m not sure if this is something I should be concerned with, but after I do retract all, top right isn’t pulling tight to an offset close to 0. The rest of them are, but it typically has a value like this Top Right pulled tight with offset 791.657] between 600 and 800. If I immediately retract all again, it will report all of them as pulled tight very close to 0.
What should I be trying next to get this to work? Thanks in advance for the help!
Hello, I am not an expert at the programming side but it might be worth a try to step back through the firmware. Remember to update all three yaml, firmware, and index of you do. I have been cutting with 1.6 with some success. Concrete shouldn’t flex so that shouldn’t be affecting calibration. Are your anchor ends just a tiny bit loose so that they can swing back and forth on your anchor bolts? But not so much that they can slide up and down.
I’m running in horizontal orientation with anchors in a concrete garage floor.
The anchors are laid out to cover about 13 ft x 9.5 ft. The machine is
calibrating on a 0.5" spoilboard + 0.5" workboard, and I have both thickness
settings set to 12.7 mm. I’m using an extend distance of 2470 mm. For
calibration I’ve been doing it with the router bit removed and the Z
physically all the way down. The force setting is at 1000. I have the latest
1.20 (firmware, maslow.yaml, index.html.gz, and fluidnc) installed.
this is a perfect report, giving us all the details we need. Thanks a lot for
this.
I’ll start by saying that this sounds like you did everything right.
I’m not sure if this is something I should be concerned with, but after I do
retract all, top right isn’t pulling tight to an offset close to 0. The rest
of them are, but it typically has a value like this Top Right pulled tight with offset 791.657] between 600 and 800. If I immediately retract all again,
it will report all of them as pulled tight very close to 0.
I think this is the key. your arms are just a little too tight to operate nicely
at a retraction of 1000.
you can try bumping up the retraction force a bit. you want it to relyably
retract all the way first try.
or you can disassemble the arms and make sure that the spools turn freely and
then figure out how low you can set the tension and have it retract all the way
I would try bumping the current limit to 1200 and see if that works, and then
try cutting to see how accurate the machine is. you can always disassemble the
arms later if you want to lower the current limit (which seems to correlate to
better accuracy)
This sounds like you might have a loose magnet on the Top Right. To get a better indication do an extend all, pull out your cable (only need do the top right) retract (Retract will be greyed out as you haven’t pulled out all the cables, click on it anyway), extend all, repeat this a couple of times. If it is a loose magnet, you will get a value greater than 1 each time.
Thanks for the advice! I’ll try upping the force to 1200 and see if I can get it to reliably retract to a value less than 1. It is always retracting fully, but the values it reports suggest it doesn’t know that. If that doesn’t work, I’ll disassemble that arm to make sure everything moves well and the magnet is seated correctly (the kit had the magnets already glued in though).
Thanks for the advice! I’ll try upping the force to 1200 and see if I can get
it to reliably retract to a value less than 1. It is always retracting fully,
but the values it reports suggest it doesn’t know that. If that doesn’t work,
I’ll disassemble that arm to make sure everything moves well and the magnet is
seated correctly (the kit had the magnets already glued in though).
If the magnet is in the wrong place, it can also cause that.
the first time that you retract, the machine doesn’t know what the prior belt
lengths were, so the adjusted distance shown can be anything.
but if you then extend and retract, it should report a very small number (<1mm,
frequently <0.1mm) if you think something is wrong, do this a few times and if
it’s not consistant, there is no chance of the machine working and you need to
disassemble the problem arm(s) to figure out what’s going wrong (which is the
perfect time to get them moving freely so that retraction forces <=700 can be
used)
I’ve made some progress, but I’m still not at a place where I can cut with my Maslow 4.1 yet.
My TR retraction issue has been solved. I ended up taking it apart and I was missing a bearing inside. It probably fell out when I was putting the two halves together. I only noticed the issue when putting it back together and being one bearing short. I still had a bag with the extra bearings. Now TR seems to be working well, and they all retract <1mm.
But it still was failing to calibrate. The BL center point was off around 50-60 most of the time. I tried raising the force (1000 > 1300 > 1500) but that made it mostly worse. Center point error improved more, but measurement repeatability got worse. I saw “measurements are not within 2.5 mm” errors with max deviation up to about 10.4 mm. Best about 0.4224. I’ve attached log from that.
Reducing the force to 800 is what made it complete a find anchor locations calibration, but it took 88 waypoints, and refreshing my browser after 86 waypoints due to it saying Calibration data not acknowledged by computer, resending Parsing and the JS console printing Parsing the GRBL CLBM message failed, the calibration data has not been ‘handled’. This is probably a programmer error.
Unfortunately, I forgot to save the log before refreshing the browser. The second log I attached is the end of that calibration after refresh.
These are the final fitness values it ended on:
Initial center point off was approximately TL 0.000, TR 0.000, BL -6.584, BR 3.454.
Final saved values:
Fitness: 0.4604681
Maslow_tlX: 7.8
Maslow_tlY: 2914.1
Maslow_trX: 3957.5
Maslow_trY: 2912.6
Maslow_blX: 0.0
Maslow_blY: 0.0
Maslow_brX: 3956.3
Maslow_brY: 0.0
So, it now says that it is “Ready to Cut” but when I move the Maslow around the work area it does not move smoothly at all. TL and BL both don’t move smoothly. They kind of pulse and it makes it so it doesn’t move in straight lines. What should I be investigating to correct that? I’m really not looking forward to taking this thing apart again…