[samu]

I have a 3 axis vertical machining center with fanuc o-md controller. When the tool path is a small radius or a sharp corner at high feed (about 100 ipm) difference between programmed tool path and actual tool path is quite big. (about 1/32"). I know that a part of this error is caused by acceleration/deceleration parameter. But can i improve accurancy by changing servo tuning parameter. Actually, parameters are the same for all 3 axis. Loop gain 2500, proportionnal gain -816, integer gain 91. I have some notion in process regulation but i am a new with servo tunning.

[HuFlungDung]

How old is your machine? What is it reasonably capable of doing for feedrate, according to the spec sheets?

If you machine an outside square corner, do you get a bellied out corner? Does it occur equally in both X and Y? You'd need to machine different test cuts to determine this because the axis making the straight cut is the one that will experience the overshoot, in other words, the overshoot is confined to moving axis, you most likely won't see both axis overshoot simultaneously. You need to be sure that both axis are behaving with the same degree of error so that you know if one needs to be tuned differently than the other, because altering the tuning of one axis can affect how circular your circular interpolations are.

Having said that, I don't see a figure for your derivative gain, I'm not sure if Loop gain = Derivative, it may be different.

In the tuning I have fooled with, I am led to believe that the proportional gain is the one you would alter to improve the stiffness of response of the servo. Integral gain is fairly touchy, and it kind of counters the effect of the proportional. Derivative gain is kind of a damper on the ringing effect created by sudden stopping of the motor at position. It is not so sensitive to tweaks.

You'd have to double check that your Loop gain might actually be the amplifier gain which may increase the max current flowing in the circuit. It may (or may not) be safe to alter this, but giving the motors a bit more juice may also detune the system a little bit, because it will affect the rapids, and accel/decel without giving you better contouring response.

It may also be the case that your PID tuning is fine, but the accel/decel ramps could be set more aggressively. The only downside of boosting these ramps would be servo amplifier overload faults. I think for testing purposes, it is probably okay to bump the accel/decel rate until you get to the point of amplifiers cutting out on overload, then back off a little bit. You need to test accel/decel with moves that are long enough to allow the motor to obtain full speed (in rapid). This might be 5 inches (or more), depending on how much power your servo system has.

Maybe Al or somebody will happen along with some more insightful comments.

[Al_The_Man]

It is possible tuning is an issue, it is a while since I got involved in a similar issue, have you looked at whether G09/G61 can correct it?
Al.

[Bubba]

Just guessing, but is CV turned on? It might be possible if using CV that it cannot follow the path exactly. May need to use an exact path or exact stop type of generation.

Like I say strictly a guess on my part as I don't really know what I am talking about.

[keebler303]

you could also put in a short dwell (G4?) at the corner to remove the rounded effect.

Matt

[austinnathan21]

CHECK THIS OUT. I HAVE AN ISSUE WITH MY MACHINE WHILE IM ENGRAVING AND I THINK THIS WILL ANSWER YOUR ?

[austinnathan21]

CHECK THIS OUT. I HAVE AN ISSUE WITH MY MACHINE WHILE IM ENGRAVING AND I THINK THIS WILL ANSWER YOUR ?

http://www.geindustrial.com/services...ts/GFT-320.pdf