Anyone had Mach3 gain .001 per .100 move?

[superchargerjim]

I have an IH CNC mill and Mach3 controller. For some reason, I can tune the motors to say, .200" and make a few test moves using my dial indicator check it. When I do the test moves, it always gains .001" per on my indicator for every .100" of move on Mach3. In fact when tune the motors, the steps required to make the same specified move always change as much as 800 steps or more. Has anybody had this problem or know how to fix it. Tommy at IH mill doesn't seem to know why it's doing it.

James

PS - the backlash is >.0001 and everything is trammed up well, so that's not it

[DMF_TomB]

in emc i set the steps per inch rate so when i ask for 0.100 i get 0.100
.
steppers motor slipping with no feedback, will travel short or
<0.100 when asked for 0.100
.
a controller with feedback resolver or encoder if encoder says it did not travel 0.100 it will travel extra til it reads it moved 0.100
.
i had a mill where feedback encoder if stuck the mill would jump far, sometimes inches. basically if encoder is not reading movement it gave it full power til it read movement of required amount.
.
i had encoder replaced and problem went away. although sometimes though encoder is dirty or has loose screws.

[superchargerjim]

Thanks for your reply. I have inspected the encoders for debris and found none, however I have not tried replacing them. The other strange thing is that the Y axis will start to gain .001" per .100" of instructed move and when returning back to Y0, sometimes it keeps the gained amount and often times it does return back to Y0 without the gain. I'm at a total loss as to how to fix this and it has brought my small business to a stand still. I will replace the encoders, but if anyone else has suggestions, please let me know!

[MrWild]

FIRST! Is this on all Axis, or just one?

If you move ten inches, the amount out is .100? .001 per .100 x 100 = a gain of .100". If you are gaining that much, it sounds like a noise issue. Are you sharing a building with a welder, or Plasma cutter? Are your encoder cables shielded cables? Is it grounded on both ends? (HUGE no no).

If you were losing in all directions, that is a stepper issue. As you have encoders, I'm assuming you have servos. With servos, and Mach, you are locked in with not knowing where the move has gone. The drive reads the encoder and keeps moving until it gets to where it is told to go. If it gets bad/misses signals from the encoder, it will move the servo until it is in reads the right number of pulses.

What happens if you attempt to move the lead screw with a wrench? (I'm assuming no hand wheels). Bad signals from the encoder, my allow you to move a few thou from zero.

If you get no displacement, the encoder cables are properly shielded, wire connections are tight, and no intermittent breaks in the encoder cables (Old cables become brittle), it's time to look at the cable from computer to your drives.

Is it shielded? Is it grounded both ends? (again, a big no no). Did your dog chew on it? is it old? What are you doing when Mach is running? Is it a dedicated computer?

Good Luck

[superchargerjim]

Thanks for all the info and I will try all of your suggestions. The CNC mill came as a turnkey servo setup from International Hobbies and is brand new. It's done it from the get to. The ball nuts where not tensioned well on the X and Y, so I tensioned them down. Y is the only axis that gains. I will check over a wider move to see if it continues to gain. I cannot think of anything that would mess with the the signal from the encoders, however I will carefully check to make sure there is no interference from somewhere that I'm not thinking off. We don't run any other high energy systems in the shop that would radiate. Anyways, I will do all that you recommended and post back.

[kmacht]

Are you sure you have your dial indicator perfectly perpendicular to the axis you are measuring? If it is tilted even just a little to one side or the other of perpendicular you will get readings greater than your move. Think of it as a triangle with your indicator as the hypotenuse. The greater the angle between the axis being measured and the indicator the greater the hypotenuse of the triangle has to be.

Keith

[jmullett]

Maybe too simplistic, but is there any coincidence that a Mach control set-up to use a 5 mm pitch lead screw on a machine that actually has a 0.200" pitch lead screw will produce similar over travel results. Turn your screw by hand 20 or 30 revolutions and check distance traveled with a calipers or even a tape measure and see if it jibes with the Mach set-up.

[rowbare]

This might not be your problem since your error seems inconsistent but you might need to calibrate your steps per : YouTube - Improved Mach 3 Axis Calibration
bob

[superchargerjim]

As an update, IH did replace the mill after many many delays and constant calls. The new mill works better, but the new ball screws have .002 in backlash in X and Y with .005 in Z...when I was promised close to zero backlash. Based upon my experience, I will never buy another IH mill. Tommy returns maybe 1 out of 15 phone calls, doesn't do anything he promises when he promises. I don't know if he is totally unorganized or he just lies all the time. He sends out wrong or incomplete parts and takes months to do that. Such a disappointment.