Thanks for all the responses.  I now know what the source of the vibration I 
hear during a G33 can be attributed to.

First a correction. I said something yesterday that was incorrect and 
misleading.  The oscillation I was seeing in the velocity waveform (also in the 
DDT of the position information, of course) DID NOT go away on coast (I said 
yesterday that it did).  What I was thinking of was that as the spindle coasted 
down it became low enough to be unnoticeable.  It was happening though after I 
shut off the spindle for several seconds before trailing off, zooming in on the 
graph revealed that.  I was thinking (to continue to blame the vfd :-) that it 
wasn’t truley coasting that it must still be interfering some….

After spending today doing more testing and peering at Halscope traces, 
oscilloscope traces I ended up convincing myself that the VFD was causing the 
speed variation.  I talked to Automation Direct tech support and theyagreed 
that the variation could be due to the VFD having a slight variation of the 
frequency output.  I put a high voltage probe on one of the motor output leads 
from the VFD and could see the PWM fluctuating (in frequency) and talked myself 
into that being the culprit.  Later I realized that the PWM would fluctuate in 
frequency to make up the output waveform but that by itself is NOT the output 
waveform frequency (and hence was a red herring).

Anyway, one of my shop mates had the idea to just wire the motor directly to 
the 3-phase AC line voltage to completely remove the VFD form the picture.  The 
motor spins at 1755 rpm at AC line voltage. With everything wired normally I 
first watched motion.spindle-speed-in in Halscope after commanding 1468rpm in 
Axis (with gear reduction that means motor at 1755rpm) just to be able to 
compare it when connected directly.  I then re-wired directly to AC and fired 
it back up.  I fully expected the fluctuation to be gone.  But it wasn’t!  So 
much for assumptions.  Doing some calculations on the resulting plot revealed 
two things.  The period was almost dead on once per rotation of the spindle, 
the variation is about 9rpm (so the spindle is fluctuating between +4.5 and 
-4.5 rpm of commanded speed), and that the motor itself was not causing it 
because otherwise the period of the variation would be at the motor gear 
reduction (0.8313).   The frequency of the variation is about 25-30Hz and this 
jibes which what you can hear when running a G33 (a low rumbly vibration).

If you look at the link I posted which shows a picture of the spindle you see a 
long mechanism that the spindle motor is turning.  
https://www.flickr.com/photos/37438950@N00/16941293785/in/album-72157651167328249/
This includes the Rohm collet closer, the casting mount (bearings inside) for 
the spindle and the head.  Somewhere in that 18+” long contraption is a slight 
imbalance that is causing the spindle to vary as it spins.  There is no 
noticeable noise with the spindle alone and nothing you can feel - I think the 
variation is just too small given the mass.  BUT, if you run G33 it is trying 
to synchronize with the spindle.  So, my Z-axis stepper is amplifying the 25Hz 
rumble.   I will post a video (in another thread) showing the G33 knurling 
routine but you can’t really hear the vibration in the video, maybe a little, 
it definitely sounds different than a non-coordinated move though... 

Anyway, the G33 runs fine, I know the source of the vibration, and I am 
declaring victory and moving on.

Thanks again for all your help,
-Tom


------------------------------------------------------------------------------
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to