Separate calibrations for Y and A axis??

[Beardie7]

One extremely frustrated person here.

I've spent thousands of dollars and made the most accurate machine that I have ever built and I can't square the machine. I think that I've isolated it down to the need for separate calibration numbers for my sides of the Y axis. But it may not even be a calibration issue since it won't go back to the same postition after jogging the length of the table. It will come back 1/2 to 5/8" from the start point. But it comes back long by the amount, not short! if I send it back up the table it will be short by an 1/8" (Y48). Have it come back (Y0) and it is long an additional 1/8", so 3/4" long. I have loosened the lower bearings on both sides of the Y gantry plates so that I am just running on the top bearings and the R&P, and no change.

all the nema 34 motors are the same and use the same rack and pinion. The calibration for the X axis is 1389 while the Y axis is 1378. The calibration that I get A axis is 1339.

Please tell me this is something simple!!!!!!!!!!

[ger21]

If it's off by 1/2", then you are losing steps, plain and simple.

Are you direct driving the pinion? ±1400 steps/inch is a very low number.

How long is your machine? A 39 step/inch difference means your A and Y axis are different by .028" per inch, and would be 1.35" different in 48". This is not possible, unless your Y axis rack is over an inch longer than the A, with the same number of teeth.

Calculate the steps/inch using math, and set the Y and A the same, then work on why it's losing steps. Start be setting the acceleration very low, and test repeatability. Keep lowering acceleration until it's not losing position.

[Robertspark]

What kernel speed setting have you set? Try 100khz for a start.

Have you wired the drives up correctly (+ & - for step signals {check again, don't assume, a couple of mins checking things over at the basic level rules it out hopefully}). An inverted signal will be a very short pulse....

I cannot see how it would be the case but have you tried swapping the drives over? Does the problem move?
Again worth a few minutes to rule out.

Check your wiring to the drives from the power supply..... Don't Daisey chain the drives, feed each one from the PSU directly.

[spumco]

This is the sequence I'd take before trying to use a sledgehammer.

First thing - if you are using closed-loop steppers and you can program the drives via computer - ensure the settings are identical for both motors. This bit me once.

Moving on:
1. Check the driver micro-stepp settings and ensure they're the same.
2. Disconnect the A motor mechanically from the drive system and command a move of measurable distance on Y - move it an inch or something you can measure pretty close with an indicator. Move it slow to reduce possibility of lost steps.
3. Repeat for A, and see if there's a difference.
4. Swap the drives between A and Y. Test again and see if the problem changes axis.

If there's a difference and the drive microstep setting and steps/unit are the same, then you have a mechanical problem or you're losing steps

5. Remove the gantry and repeat the above if you're still pulling your hair out. Just set it on the frame or something so the Y/A trollies can move completely free of any resistance - this should eliminate lost steps as an issue.

Other mechanical stuff to check:
Pinions or pulleys loose on shafts - that bit me a couple times.
Rack loose on frame - rare, but trying to eliminate everything.

[Beardie7]

Well I may have found the problem....After posting the problem I saw that one test would be to switch two of the rack and stepper motors. That did move the problem from the Y axis to the Y axis.

I said that it is the most precision machine that I have made....well I lied. I had originally designed my "Joe's Evo" to run on Nema 23's, but after getting back my water jetted plates was persuaded to go up to Nema 34's. I had already bought the Rack an pinion sized for 23's. Well i debated sending the rack system back and get the 34"s geared rack, but the gears wouldn't fit the openings that I had designed ( I now understand Joe's design better, but wanted a more compact design). So I ended up drilling out the stepper gear from 3/8" to 1/2" on my mill. The first one wasn't the best but the second two I was happy with. It was that one gear that I wasn't happy with binding on the off center hole. It could get through a rotation but it is totally out of specs.

I have ordered new gears and am going to try again at enlarging the bore to 1/2". I'm using 637oz/in steppers and don't think that the belt will be the problem if I can get a true hole. To get "proper" Nema 34 gears I would have to tear the whole gantry apart and butcher my aluminum plates, plus have to buy new gears and belts.

By the way I'm using an UC400eth, to a Leadshine MX4660 BOB, to run Leadshine Nema 34's steppers (4A, 3.6Mh, 637oz/in)

I will post again after I have installed the new gears. At least I know what I'm fighting now.

Thanks,
John

[Beardie7]

Back to the drawing board.... I have greatly improved my run out on the redrilling the new gear from 3/8 to 1/2. When installed in the rack I can freely turn the the gears with the power off and the belt on. I am still losing steps somewhere. After jogging the gantry around the table, when I bring everything back to where I started I was off 1/2 on the X and 1/16" on the Y. The lost steps are from going in the positive direction, so I ended up closer to front left position.

The calibration Numbers are still way off from it's brothers. I have switched the motor from the A to the X and the problem Stepper brought it's issues with it. When running the calibration on all of the motors I get 1378, 1389, and 1320. It is the 1320 that is giving me the problems.

I'm using imperial measurements and am using the rack and pinion from finelineautomatation.com.

Any ideas are greatly welcome no matter how odd they may be.

John

[Beardie7]

Well I finally have a properly functioning CNC now. I bought a mini lathe to bore out the 3/8" bore to a half inch to handle the NEMA 34's properly. And that still wasn't the problem. I was getting slippage in certain parts of the table. I narrowed it down to the left R&P and found that one of the gears was loose and was wobbling. Tightened the bolt and everything is working as well as I had hoped.
I added a "IMP model#3" 5" vise in the far left front of my table and have found it very useful. It only sticks up about 2" to the bottom of the clamping area so I still have close to 4" usable clearance.

Now to try to master feeds and speeds.....

John