Tree J425 tuning

Hello, I have a Tree J425 that I have been banging my head against the wall for some time trying to tune the PID loop. I went through the steps provided by tommylight to accomplish the tuning. Once I figured I had it right I put my dial indicator on an axis and found that the commanded position vs the actual movement was off .003in on a .1" movement. Through continued tweaking of the PID and the FFn values I was able to get it down to .0015 but nothing further. Part of me thinks the hardware is to blame but part of me thinks I am pooching the tune. Any help is appreciated.

Hi
Welcome
some more infos needed as you may have retrofited the system to a SERVO one
what are the Encoders
there are 2 ways as the Machine itself are deliverd in more then one version
you are facing the problem of the Leadscrew mount not in exact pos and fot fixed on one side only tight to the motor oposit side
Fixed and free run Bearing called

Please Tell what system you got running and maybe the controller type you are dealing

The best way to get this system to a real accurate CNC is to use the Liniar Scale direct Axis mount Glass scales for setprice at 250USD
with 10uM Scale steps that can b devided by 4 signals into the Controller

actual movement was off .003in on a .1" movement

How much error are you getting on longer moves? As an example, if you came up .030 short at 1" move, it would suggest you have your encoder scale wrong. If the error is consistent when moving in the same direction, as an example, indicator overshoots by .0015" then creeps back to zero, that is more likely PID tuning. If you're testing in different directions, backlash comes into play. I'm guessing here, your machine has the encoders fitted on the end of the ball screw? Any play in the gib, ballscrew mount, and screw will be masked. I would also check how much lost motion you have during axis reversal. As an example, you're encoder may be counting .003" of movement before the table moves.

Ok, so I am able to get back to this. I checked the encoder count and adjusted it and it looks like it fixed the x and y. I am now tuning the z and it is acting strange. It seems to lose the encoder on the up stroke. Comparing the up stroke and the down stroke it seems that the down stroke is moving slower than the up. Do you guys know what would cause this?

The velocity for the down stroke is much flatter than the up stroke. The upstroke peaks much faster and it coasts to the commanded position. I will see if I can get a pic of the halscope.

Just for curiosity I probed the Aout2 on the board to read the voltage it was giving to the drive as a position command. I was surprised to find it was about half the voltage for the down stroke vs the upstroke. Do you know what would cause this?