[elmeys]

Hello everybody, we're new in cnc machining and recently we're having trouble with a cnc mill that have Fanuc 6mb control, the issue comes when make a circular interpolation and the result is not a perfect circle, example 6" inside diameter if we measure in Y axis is 6.005" and X axis 5.998", we obtain and oval. We are loooking in internet for information about this issue and for maintenance manual of fanuc 6mb without any luck. We think that servos don

[tc429]

program a 45 degree feed move with a m99 so it moves back and forth... look at diagnostics 800-801, they NEED to be identical (sometimes exactly half depending on tach 3.5/7 v per 1000 jumper or parameters)

if you run yellow (6047 drive) or red cap(6050 drive) servos (blackcaps are too old for me to remember) there are three pots on the cards... either vertical (top-to-bottom) or horizontal (left to right) the order will be gain-offset-tach

no guarantees, use at own risk, but heres how I always adjust them:

*turn off auto drift comp parameter- reset drift comps (if any) to zero before starting- sorry I dont recall parameter numbers...

1) set gain at halfway point. (top or left pot)
2) jog axis + at low speed, stop, note error at dgn 800-802 (800=x axis...)
3) jog - at same speed, stop, note error again.
4) adjust 'offset' pot if needed, recheck, repeat until both stop at 0-1, or at least equal/opposite signs
5) run the 45 degree loop
6) adjust 'tach' pots to get same values +/- in both axes. (turning pot cw decreases error in 800-802)
I never really worried about THE number, just that they match... try to get them exact.
7) cut feedrate in half, if they still match, great... run it.

if they dont match at different feeds, you will need to dick around with gain, load, counterbalance type things... I'd just readjust running the 45 at the finish feedrate you intend on precision cutting at, otherwise, more careful tuning will be needed... I type slow

once all done, turn auto drift comp back on if desired...it will self adjust pretty well...usually we'll get a couple years between adjustments unless a drive or control gets swapped (do NOT put the old drift comps back in- control will if it needs to)


if you run yellow cap dc motors a common problem that can make them 'buzzy' or hard to tune is pulsecoder coupling wear... powerdown, pop the yellow cap, MARK the encoder orientation, mark the motor shaft position, THEN remove the three screws on the pulsecode and gently pull it STRAIGHT out... slowly turn it over (idea is NOT to roll it so fast as to turn the shaft...if you put it back on 180 out, grid offset/home, everything will need messed with) and mark the shaft/housing so you know how it came off... pull out the little brass oldham coupling stuck either to the motor or encoder shaft, youll probably find one or both sides are a little loose in their slots... take a dab of thick grease, put on the key, see if that takes all the shake out, if not take a small punch and put a few dimples in the copper- not much, just a few dots to snug it up- still grease it. reassemble paying attention to shaft/encoder/housing marks and all the buzzing will 99.9% of the time be gone. no need to do this on redcaps as they use a taper fitted rigid connection, or preloaded plastic oldham coupler... much better motors


Biggest failure cause on yellowcaps is magnets coming unglued, then sticking to the rotor like a brakeshoe- until the keeper comes off and crunches into the rotor destroying it... brushes can cause issues too, but rarely- they last a looooooong time, but wouldnt hurt to pop them out for a look-see.
if you ever get a rebuilt DC that wont 'offset' properly, pull it off, run off a 12v battery and tach the shaft- reverse leads, rpm should be same, if not, under the yellow cap are 4 studs...loosen them and rotate the brush holder/endbell to match the rpm in both directions. if you get one thats overheating/noisy, might have a magnet unglued/ dragging... we were able to save a few by pulling apart and epoxying sections of hydraulic tubing between the magnets to form a 'arch' and hold the magnets against the shell... seems silly putting a big motor together with glue, but Fanuc only glued them to the shells on the old yellow caps

[zavateandu]

backlash in the ballscrews when it changes direction

[generaldisarray]

I am using this guide to tune some 6047s.
I cannot set the offset any lower on my X axis. It uses a H002 drive with a 5M motor. DGN 801 is -6 after the slow jog then stop, either direction.
I was able to get Y and Z easily to around zero.
the 45 move shows Y and Z right around 140 where X is 139~142 in one direction and 146~149 the other.
any thoughts?

[generaldisarray]

I swapped out the drive with a spare I have. With the offset cranked fully CCW on the new drive I am at -2 to -3 using the slow jog test. On the 45 move X is within one or two of Y and Z.
This is a Pulsecoder machine BTW.
They way I understand it. My problem with this RV2 is that the control is sending some kind of "hot" VCMD to the X drive.

Is it a hardware problem or a settings issue?

I test cut a 3.375 dia circle. It measures round within about a half using a 50 millionths test indicator, most of the error is at the X quadrants. The X screw makes some noise on rapid, I am not sure if it is the bearings or the screw itself. So at this point I dont know if I can get it any better than this without investigating the Ballscrew noise.
It is miles better than it was.

[tc429]

hardly ever check email or in here anymore, but saw a post to this old thread Anyways, you can fudge the drift comp parameter even if auto drift comp is turned off - doing that can recenter your offset pot. kinda guessing a old value may still be in there causing the offset bottoming...
if you are round within a half in 3 3/8 on a older machine, youre probably not bad... if boxway, have you checked lateral clearance on gibs? screw backlash, servo coupling backlash (yellow cap encoder oldham coupling backlash).
if the offset is varying or the servo gets to acting too odd, the little 8pin 455 op amps near the axis LSIs have caused me grief before- age seems to affect them even more than capacitors for some reason...Radio shack used to sell them, mail order these days... get sockets too