Display Encoder position with button to zero.

Is this a standard cylinder or a hydrostatic one? (the hydraulic servo machines I used to work with used hydrostatic cylinders and managed positioning to microns at huge loads)

I expect that you are right that touch-off works to the commanded rather than measured position. You might be able to get round that by incorporating your own sall offsets within hal (look at the "offset" HAL component.) Mux2 can be used to store and latch a value, if you feed its output to the input.

This is a standard hydraulic cylinder. I added the servovalve to control the table via PID loop. I think it would be better at precise positioning if the ways were linear bearings rather than flat/v-ways. I believe It's the stick/slip on the ways that is the issue. The ways have been flaked and I believe they should not be.

Here is the thread from the original build - www.cnczone.com/forums/linuxcnc-formerly...er-emc2-project.html

I see the offset HAL component has 5 pins but I am a little unclear of their function. Is this correct? (I have a feeling I am way off here)

offset.N.offset - The offset value (I would net this to my zpos-cmd) ?
offset.N.in - This is the input for the value that gets added to the offset ?
offset.N.out - Equal to: (offset.N.offset + offset.N.in) ?
offset.N.fb-in - The feedback input (I would net this to zpos-fb) ?
offset.N.fb-out - Equal to: (offset.N.fb-in - offset.N.offset) ?

So I would use Mux2 to store and latch the offset.N.fb-out value?

Is it possible that there is air in the system, causing problems?

That is a possibility, The main cylinder has "choke coils" at each end that bleed the air out. When the machine starts up after sitting a while I have to run it to each end of the stroke to get the air out of the cylinder. It runs smooth after that. When given a position command, it may stop within 0.0002" of the desired location, then the output will slowly start to climb until it overcomes the friction but it may overshoot by 0.0002" once it moves. This leads me to believe it is the "stick/slip" of the ways. The ways were flaked on this machine which should not have happened and I believe this increases the force required to overcome the static friction. Even if the machine did stop on a dime, I would still want to use a hard stop just to prevent unintended crashes.

If the ways are bad there could be all kinds of binding going on. Both from the friction of the bad ways, and possibly another source that could have caused the over loading and failure of the ways in the first place.

Sounds like the first order of business is fixing the bad ways.

Beezerlm wrote: This leads me to believe it is the "stick/slip" of the ways. The ways were flaked on this machine which should not have happened and I believe this increases the force required to overcome the static friction.

This is not something I have heard before, but that doesn't mean it can't be true.

I wonder if changing the oil would help? Can the Gibs be adjusted to run slightly more freely?

In some other systems I have worked on a dither signal added to the command can help overcome mechanical hysteresis. That might be worth a try (maybe turn it on when the error is small and the I-term hits a threshold value)