About to toss this thing

[krazyking01]

I have spent hours upon hours trying to calibrate this thing, and for some reason, it only seems to get worse. I have tried the hardware calibration via the screws, no luck. I've tried the firmware calibration, no luck. The horizontal radius has been from 110 to 150, recalibrating every .5 step. Nothing has worked. Right now, here is where I'm at. Between X - Y, and X - Z it scrapes the glass. Between Z - X, it jumps up about 5mm. Any ideas what is going wrong?

[BenTheRighteous]

Well for one if you have the stock arms and your horizontal radius is outside of the 125-135 range, you're way off. I'd even say outside of the 129-131 range would be surprising.

Put it back to a sane value, check for loose hardware (cheapskates especially), belt tension, and redo the calibration. It helps if you have a real feeler gauge rather than a piece of paper. (Not sure what you kids are talking about these days with your fancy shmancy newfangled firmware calibrations... back in MY day we did things by hand... yadda yadda yadda).

Assuming you have a Max and not an Orion, right?

[KAS]

If your mechanical calibration is off, you'll drive you're self nuts trying to chase it down with EEPROM settings. I want to say 130 is the perfect number, I'm not sure where I saw that.

I do mine a tad differently, I'll adjust the screws till X,Y,Z are all exactly same height off the glass. I use feeler gauges, but you can use a socket or anything that's consistent. Then measure the center, and compare it to Y. Make the adjustment in the EEPROM for horizontal radius until the Y matches the height of the center. Once Y matches Center, check X and Z. If everything is good and I mean perfect ( just okay or good enough doesnt count), set the overall height from the LCD to the sheet of paper or .08mm from the glass.


Just remember that horizontal radius will not move the location of the nozzle at center. Only X,Y,Z will move when you change that value.

[stonewater]

1- reload your firmware after writing down your pid calibration settings.
2- check for teeth missing on your belts
3- 130 is just about perfect.

once that is done. reset your Z zero, then set your x y z tower height, but do not set the z zero every time, just do the 3 towers, homing after every turn of the screws. then do the horizontal radius. I found myself chasing the Z zero in circles if I reset z height after every screw turn, that was from me misreading the calibration instructions.

Tom C

[krazyking01]

When I tried the EEPROM method of calibration, I removed the screws so that the skates themselves were the ones triggering the endstops.
Here's how I do my calibration:
1. Set initial Z-height.
2. Set to home.
3. Move to the X tower, 5mm off the glass.
4. Slowly lower until I have the same grab as center.
5. Adjust either screw or EEPROM(with screw removed).
6. Check Z height at center and adjust horizontal radius as needed.
7. Repeat 2-7 until all 4 points are even.

I've checked for anything being loose, and everything is just snugged up. Even so, with over 12 hours spent this past couple days trying to calibrate it, I still get the results from the first post.

[Nylocke]

Sounds like you got some tower lean, you could check your towers perpendicularlarity to the bed with a digital angle gauge (Wixy or something like that). They are $30 or so and they are very useful for getting an accurate angle between 2 planes.

[DLS3141]

I had a problem with mine when building it and found that one of the stepper drive pulleys was broken. That was with the older, 2 piece plastic pulleys, but if you don't have those, not having the stepper pulleys secure to the shaft would cause the same thing.

[KAS]

When I tried the EEPROM method of calibration, I removed the screws so that the skates themselves were the ones triggering the endstops.
Here's how I do my calibration:
1. Set initial Z-height.
2. Set to home.
3. Move to the X tower, 5mm off the glass.
4. Slowly lower until I have the same grab as center.
5. Adjust either screw or EEPROM(with screw removed).
6. Check Z height at center and adjust horizontal radius as needed.
7. Repeat 2-7 until all 4 points are even.

I've checked for anything being loose, and everything is just snugged up. Even so, with over 12 hours spent this past couple days trying to calibrate it, I still get the results from the first post.

Question: Following your steps, I thought I found an error until step 7 you said "until all 4 points are even"
If all 4 locations are the "exact" same, you're done it's calibrated. Close or good enough doesn't count with this, has to be exact.

Although,


If you are just doing steps 3-6 before moving on to X & Y, then that might be the issue.

[krazyking01]

Let me amend that some to clarify. I adjust the screw or the numbers in the EEPROM until the tower is at zero, then move to the next tower. Once those three are all at zero, I check the center and adjust the horizontal radius as needed. Then I keep going in that fashion until I don't need to adjust anything and all points are at a perfect zero. Then I run the test pattern, and this is the point that is irritating me. After spending all that time getting everything exact (close doesn't count for me, I'm a perfectionist), X-Y and X-Z press into the glass, while Z-X arcs up almost 5mm. When that happens, I start back at square one.

[KAS]

Okay, If you're getting good calibration at X,Y,Z and Center but not between the towers then it's mechanical calibration. The effector is lifting in some areas and dipping in others.

This is pretty much what another member is going through.

http://forum.seemecnc.com/viewtopic.php ... 68&p=67058

[teoman]

Would you mind pressing home on the lcd so that the motors lock in place and then try to rock the carriages and then the nozzle. If something is loose you should notice it.

[Earthbound]

I have spent hours upon hours trying to calibrate this thing, and for some reason, it only seems to get worse. I have tried the hardware calibration via the screws, no luck. I've tried the firmware calibration, no luck. The horizontal radius has been from 110 to 150, recalibrating every .5 step. Nothing has worked. Right now, here is where I'm at. Between X - Y, and X - Z it scrapes the glass. Between Z - X, it jumps up about 5mm. Any ideas what is going wrong?

It stands to reason that if changing the horizontal radius value is not improving or worsening the calibration, the problem is elsewhere.

Your next to last sentence and the one before it seem to contradict, and there is no mention of state between Y - Z towers. If X-Z was a typo and you meant Y-Z scrapes the glass, then the issue is almost certainly on the Y tower. If Z-X was the typo and you meant Z-Y has a 5mm gap, then the problem would be on the X tower.

I'd look for something that throws off the accuracy of that one tower's travel. A loose pulley, a loose/bad belt, a sticky limit switch. The 5mm deviation you are seeing is too big to be caused by a slightly out of square tower, IMO.

That's my $0.02