Except that the measurements in the yaml are manual measurements of my frame, not auto-calculated.
Taking this thing apart simply on a guess that something like a magnet might be loose is going to be the end of me. Seriously. If the thing retracts all and extends all perfectly to measurements I give it, every single time, then how can the magnets and sensors be off? If I change extend distance to 6000mm and extend all, I can reliably get the belts to extend all the way to bare spools. I can then retract all and they all pull tight with low offsets. If I then change the extend distance to 2100mm (the distance to my arms), they all extend perfectly to that distance. Every time. These basic things work every time as they should. Itās only when I try to calibrate or actually use the machine that I have problems.
Well maybe once Iām producing dust and actually moving then the hot glue might matter but this thing has been basically stationary (aside from calibration attempts) and the connections seem fine - as stated, all four belts extract and retract all exactly as they should - correct distances, no interruptions, proper tightness. Iāve changed the extraction distance many times over the course of testing and never had a problem with that aspect.
When using manual calibration measurements, my problems always start at center point deviation:
[MSG:INFO: Retracting all belts]
[MSG:INFO: Requesting state change from 0 to 1]
[MSG:INFO: Bottom Left pulled tight with offset -0.011]
[MSG:INFO: Top Right pulled tight with offset -0.032]
[MSG:INFO: Top Left pulled tight with offset -0.086]
[MSG:INFO: Bottom Right pulled tight with offset -0.054]
[MSG:INFO: Requesting state change from 1 to 2]
Extend All
[MSG:INFO: Extending all belts]
[MSG:INFO: Requesting state change from 2 to 3]
[MSG:INFO: All belts extended to 2100.000mm]
[MSG:INFO: Requesting state change from 3 to 4]
Apply Tension
[MSG:INFO: Requesting state change from 4 to 5]
[MSG:INFO: Measured waypoint 2] [MSG:INFO: Center point deviation: TL: 0.000 TR: 0.000 BL: 96.271 BR: -15.049]
[MSG:ERR: Center point deviation over 12.000mm, your coordinate system is not accurate, maybe try running calibration again?]
[MSG:INFO: Requesting state change from 5 to 4]
When I auto-calibrate, my problems are the same:
Calibrate
[MSG:INFO: Requesting state change from 4 to 6]
[MSG:INFO: Setting z-stop position]
[MSG:INFO: Machine Position found as X: 4.328 Y: -110.802]
[MSG:INFO: Measured waypoint 0] [MSG:INFO: Center point off by: TL: 0.000 TR: -0.000 BL: 86.285 BR: 0.411]
[MSG:INFO: Machine Position computed as X: 2.535 Y: -94.105]
[MSG:INFO: Measuring Frame Flex]
[MSG:INFO: Flex measurement: TLBR: 0.309 TRBL: 97.470]
[MSG:INFO: Measured waypoint 1]
[MSG:INFO: Measured waypoint 2]
[MSG:INFO: Measured waypoint 3]
[MSG:INFO: Measured waypoint 4]
[MSG:INFO: Measured waypoint 5]
CLBM:[{bl:2232.54, br:2139.24, tr:2252.92, tl:2251.36},{bl:2270.25, br:2016.93, tr:2136.93, tl:2374.97},{bl:2347.70, br:2100.53, tr:2046.67, tl:2297.60},{bl:2232.62, br:2222.09, tr:2161.13, tl:2171.37},{bl:2111.23, br:2346.01, tr:2289.20, tl:2046.37},{bl:2022.82, br:2270.25, tr:2372.03, tl:2131.68},]
Computingā¦ This may take several minutes
[MSG:INFO: Requesting state change from 6 to 7]Fitness: 0.1247685 in 100
Fitness: 0.1247685 in 200
Fitness: 0.1247685 in 300
Fitness: 0.1247685 in 400
Fitness: 0.1247685 in 500
Fitness: 0.1247685 in 600
Fitness: 0.1247685 in 700
Fitness: 0.1247685 in 800
Fitness: 0.1247685 in 900
Fitness: 0.1247685 in 1000
WARNING FITNESS TOO LOW. DO NOT USE THESE CALIBRATION VALUES!
So it doesnāt seem like anyone can figure out what is causing my problems.
I assume the next step is to disassemble and check other things (which I am really dreading).
Is there a disassembly guide of what I need to take apart (and what I can leave as-is)?
Once I disassemble, what are ALL of the things I should check/fix on a stock standard v4?, eg glue the rj-45 connectors, something I saw about soldering encoder boards, etc etc. Would be ideal if all of these things were in one place but they seem to be scattered all across the forums without an easy way to find them. For example, if I had known during assembly about the solder issue on the boards I would have dealt with it then. I only found that accidentally while trying to get this thing to work.
I assume the next step is to disassemble and check other things (which I am really dreading).
Is there a disassembly guide of what I need to take apart (and what I can leave as-is)?
Once I disassemble, what are ALL of the things I should check/fix on a stock standard v4?, eg glue the rj-45 connectors, something I saw about soldering encoder boards, etc etc. Would be ideal if all of these things were in one place but they seem to be scattered all across the forums without an easy way to find them. For example, if I had known during assembly about the solder issue on the boards I would have dealt with it then. I only found that accidentally while trying to get this thing to work.
solder on the encoders to ground a pin on the chip (a blob between the pin and
the surface mount cap next to it works, there are pictures showing what)
No way Iām buying an upgrade if I canāt get 4.0 to work first.
I tried printing and installing upgraded belt guards when I built my machine originally. Found the installation process worse than the standard 4.0 belt guards. Not sure I want to go through that again. Belts have retracted and extended dozens and dozens of times now without incident (all the way to bare reel and back again) so I think the belt guards are safe for now.
A few more observations before I break this thing (erā¦ disassemble):
I re-measured everything, plugged numbers into @dlangās calibration site, retract all, extend all, apply tension.
at idle, the rj45 connectors on my arms have steady green and orange lights - except for one arm, which only has green light. Sometimes itās the lower right, sometimes upper right.
When I pressed Apply Tension, as with previous tries, 3 arms tightened all the way. Upper Right did not - stayed very slack. I then get a center point deviation error. Nothing new for me. That arm has no problem retracting tight and extracting to my frame before calibration. Not sure why it stays slack when I try to calibrate or jog the machine. Is this normal behavior?
If I press Jog while the upper right belt is slack, I get fans spinning etc but no movement. Pressing Apply Tension again has the upper right loosening even more before trying to re-tighten but it always goes back to its slack position. Never tightens like the other arms.
Also - if I have to take the reels apart to check for loose magnets, whatās the minimum I have to disassemble to do that? I want to leave as much of this thing together as possible. Is there anything else that could cause my problems other than loose magnets? If there are no loose magnets then what else could be causing this thing to fail?
I re-measured everything, plugged numbers into @dlangās calibration site, retract all, extend all, apply tension.
at idle, the rj45 connectors on my arms have steady green and orange lights - except for one arm, which only has green light. Sometimes itās the lower right, sometimes upper right.
the Green LED is saying that there is a connection, as they rotate the other
light goes on and off, just to indicate that itās moving.
When I pressed Apply Tension, as with previous tries, 3 arms tightened all
the way. Upper Right did not - stayed very slack. I then get a center point
deviation error. Nothing new for me. That arm has no problem retracting tight
and extracting to my frame before calibration. Not sure why it stays slack
when I try to calibrate or jog the machine. Is this normal behavior?
with that upper right arm, when you disassemble it, move the motor slightly
further from the idler and it should move much easier.
I think the fact that this belt is not getting tight is causing a lot of yoru
problems.
If I press Jog while the upper right belt is slack, I get fans spinning etc
but no movement. Pressing Apply Tension again has the upper right loosening
even more before trying to re-tighten but it always goes back to its slack
position. Never tightens like the other arms.
Also - if I have to take the reels apart to check for loose magnets, whatās
the minimum I have to disassemble to do that? I want to leave as much of this
thing together as possible. Is there anything else that could cause my
problems other than loose magnets? If there are no loose magnets then what
else could be causing this thing to fail?
you have to either remove the plastic towers from the sled or take the tops off
the tower (IMHO taking the towers from the sled is easiest, you can leave the
three bolts on one side loose, but not fully out)
detach wires from the arms
then you can remove the arm/router/clamp assembly from the sled (run it up until
you hear it clicking, then you should be able to lift it off
you need to detach the 4 towers from one of the clamps (doesnāt matter top or
bottom)
you have to losen the bottom clamp so it can come off the router
this will let you remove the bottom clamp so the arms can come off
then you need to take out the arms and work on them individually (just so you
know whatās what and if you keep having problems with the same arm after you
work on them, Iād label them in some way)
The 3rd arm gets very tight when I retract all and extends perfectly during extend all. It just doesnāt get tight when I apply tension and try to jog. For some reason, it just stays extended.
Yes - my point is that the orange light also stays steady at idle on 3 of my 4 arms. On one arm, the orange light is dark until the unit retracts or extends. Not sure why that is.
Yes, itās my problem but Iām still not convinced that itās caused by the arm itself. It reliably retracts tightly against the machine when I retract all, even after full extension. It only doesnāt retract all the way when I extend all and then Apply Tension or try to calibrate. Weāll see if itās a magnet issue but if itās not then it seems like thereās a bug in software somewhere.
the Green LED is saying that there is a connection, as they rotate the other
light goes on and off, just to indicate that itĀ¢s moving.
[/quote]
Yes - my point is that the orange light also stays steady at idle on 3 of my 4 arms. On one arm, the orange light is dark until the unit retracts or extends. Not sure why that is.
itās coincidence. the brightness of the light reflects the angle of the magnet
to the sensor. since you didnāt attempt to align the magnets (and Iām not sure
how you would) the lights are not going to be in sync. the important thing is
that as you extend/retract, they will fade in and out.
Yes, itās my problem but Iām still not convinced that itās caused by the arm
itself. It reliably retracts tightly against the machine when I retract all,
even after full extension. It only doesnāt retract all the way when I extend
all and then Apply Tension or try to calibrate. Weāll see if itās a magnet
issue but if itās not then it seems like thereās a bug in software somewhere.
The part I hated most about disassembling was removing the hot glue. The 4.1 update freed me from this, but I wasnāt even thinking about the update until, during the disassembly-assembly process, I came to understand Maslowās dynamics much better. But not before hating it a lot, haha.
Because I used too little glue, sometimes the magnet would come out by itself when I took the arm apart, and other times I discovered that it wouldnāt come out, but it could be rotated.
This is the expected behavior, just the lower belts should pull tight for apply tension. The tension in the lower belts should be enough to pull the upper ones tight. To help them out I like to pull the machine down a bit before applying tension. I will make an updated calibration walk-through video next week.
I donāt have an exact sense of what is going on from what you are describing, but I very much wouldnāt recommend taking anything apart just yet. Iām not convinced that there is any sort of hardware issue.
I think that the next thing to do is that I will make a video walkthrough for the calibration process and then if you follow along with the same steps we can see where things behave differently from what is in the video.
Iām not sure exactly what is happening, but until weāre comparing apples to apples with knowing that weāre doing all the same things itās tough to make any sort of guess.
Thanks @bar Looking forward to the video. Per above, I tried manually entering all of my digits with @dlang help and his website but had the same behavior. Hopefully youāll see something in my video when I follow along with yours.
Iām highly skeptical that this is a factor in my problems. Maybe after I start using the machine and generating dust and static, but right now the machine is essentially stationary and the problem seems isolated to one specific issue - the machine canāt figure out the size of my frame even when I tell it what the size is.
After I run through Barās tutorial vid, Iāll post my own vid and if the problem isnāt something obvious I will disassemble and start checking magnets and soldering.
Try increasing the calibration force.
I had heaps of trouble getting through the calibration process consistently and would get stuck in fitness loops too. I never got a single calibration to work with less than 1300 calibration force. Once i put it to 1300 i started getting through some calibrations but they would then have belt slack as i moved the maslow around the frame.
Today i upped it to 1800 and had my smoothest calibration yet, fitness calculations took no longer than a couple minutes,1800 x 900 calibration area, 9 x 9 grid and no belt slack post calibration. No idea if its the ācorrectā thing to do but so far so good for me
I finally got a successful calibration. Hereās what I did:
Every time I calibrated previously, I noticed that 1 arm did not tighten.
I disassembled and switched the position of that arm, adding a piece of tape so I was sure which one to focus on.
I re-assembled and attempted calibration again. The same arm stayed slack even though it was in a different position.
I disassembled again and took that specific arm apart. The magnet was glued tightly and in the right place. The sensor seemed perfectly aligned but I tried loosening the motor anyway to follow @dlangās suggestion to move it. 2 of the screws holding the motor would not loosen.
I re-assembled the arm without moving the motor but I assume Iāll have to return to this problem later if I hope to solder the board as suggested.
I changed nothing else in the arm or in the machine.
I booted up and retracted/extended again a few times until I got good values.
I Applied Tension but I got the dreaded center point deviation error again.
I pressed Calibrate just to torture myself but THIS TIME it actually successfully calibrated. I have no idea why.
My calibration settings (which I had tried many times previously) were 300x300, 3x3 grid, 1500 for both retraction force and 1000 for calibration force.
I calibrated on my garage floor, no spoil or work board underneath the router.
I assume I will have to calibrate again when I add a spoil board and work board. Hopefully whatever caused the problem previously will now be solved forever.
Hoping one of you experts can figure out why everything worked this time but never before without any changes.
My hunch is that there was something funky going on with that arm such that it wasnāt tensioning fully.
Not necessarily, I donāt bother to re-calibrate for different wood thickness, but it is probably the best practice. If you do decide to re-calibrate Iād recommend downloading and backing up a copy of your maslow.yaml file so that you can always get back to this calibration if you need to.