On 7/4/23 18:18, gene heskett wrote: > On 7/4/23 16:49, John Dammeyer wrote: Hi Gene, I think you are overthinking the problem. .
> Lets clarify where the error is, because the turn around at the bottom > of the hole is synchronized, the turn around at the top is not. > So where is it when the error was reported? The error occurred after the I3 completed. Recall I3 tells the spindle to turn 3x as fast CCW as it turned in the other direction CW so 600 RPM. The spindle is started earlier with an S200 M5. If the spindle is up to speed at 200 RPM the knee heads up accelerating up to 10 IPM and the tap threads the hole. At the bottom of the hole the spindle is instructed to slow down and the Z axis tracks that from 10 IPM down to 0 IPM so that the thread isn't ruined or tap broken. Once the spindle is completely stopped (and the Z axis would be too) the spindle is accelerated up to 600 RPM CCW and at the same time the knee is accelerated to 30 IPM which is also 3x the amount when it went down. It has no trouble following that because I suspect the acceleration is likely based smaller of the two accelerations which is the Z axis acceleration. After all it's a tracking operation so it may not even use the max accel values or the min of the two so they can both keep up. Now, at the top of the hole once the tap is out the synchronized motion is complete and the spindle direction now has to be reversed to return to where it was set with the M5 CW rotation. Here's where it screws up. It's no longer _required_ to have the Z axis track so I suspect the Spindle MAX_ACCEL value is now used. However from the Z axis perspective it may still be trying to track the CCW spindle until the spindle is stopped. Therefore we have the spindle with a MAX_ACCEL value of 300 and a Z axis MAX_ACCEL value of 10. The Joint2 axis (Z) just can't keep up and that’s when the following error happens. So by reducing my spindle acceleration to a more reasonable value which doesn't really appear to create problems the Z axis, which I suspect still has tracking enabled can now keep up as the spindle slows. And now I can speed out at I6 or 1200 RPM which it does without any Joint 2 tracking errors. John >> >> That reads like the spindle stop time is overly long to me. Nope. >> >> Actually the spindle stops so quickly that I can hear the splines >> rattle in the pulley driver. > > The stop should be so violent the rattle is preloaded to silence. > > Do like I've done, add 2 timers in your hal file, one triggered by the > motions direction output, to start when the reverse come in, and is > stopped when the spindle has has stopped as indicated by another timer > set to retrigger for 10 milliseonds on every edge the comes out of the > spindle encoder. The idea it to use this 10 ms lag to indicate the > spindle is slowed enough to be reversed w/o tripping breakers all the > way to the substation, and display the time the first timer is stopped > at in a pyvcp box. I have additional hal stuff to block the reverse from > the spindle control until t is everything but dea stopped, then and only > then do I allow the reversed direction signal into the spindle control, > and the same limit3 then ramps it back up to the chosen speed, with a > mux4 steering the input to the limit3. So the stop is a very fast ramp > down, and the accel ramps it back up to the same speed in the other > direction. 3k max spindle speed fwd to 3k speed in reverse, in under 400 > ms. z s/b able to follow that. If not, slow the limit3, giving z a > chance to catch up.� That is assuming the following fault is being > blamed on z.. > > I'm still convinced a lengthy stop is the problem. > Z is following the encoder. > So block the direction change by making a stop out of it until its > stopped, then allow the direction change thru and ramp it back up to > speed.� To do that right needs a 4 quadrant control. > > Your stepper/servo's may be quite precise, but are they true 4 quadrant > control? IDK, but very few stepper drives we can afford are. > > Jon's pwm-servo is, a vfd programmed correctly is. Both can stop the > motor by sucking the spin back out, storing that recovered energy in the > psu's filter caps, then using that over voltage to re-accell the motor > in the other direction and the only problem is the few ms of the > resultant over voltage on the filter caps. The over voltage spike is > there and gone again long before the caps fail short from overheating > due to the leakage at that voltage peak. I think that high voltage pulse > helps to keep an electrolytic formed. Both of my bigger machines are > past due for a 5 year recommended shotgun replacement with no sign of > ageing yet. > > You may need to check the addf order in your half file, the basic rule > is that all addf's should be in an order where a single signal coming in > from the machine, should fall thru all processing and go back out to the > machine in the write at the bottom of same invocation of that thread. > Violate that rule and all sorts of problems can crawl out of the woodwork. > > We need an analyzer script to check that Andy.� Hint hint ;o)> > > > � I suspect that the spindle stops and then starts back in the > clockwise direction� too quickly and the Z axis can't do that fast enough. >> >> My spindle is an AC Servo run with step/dir so I could change some >> values in the ini file. >> >> [SPINDLE_9] >> MAX_VELOCITY = 50.0 >> #MAX_ACCELERATION = 30.0 >> MAX_ACCELERATION = 300.0 >> # The values below should be 25% larger than MAX_VELOCITY and >> MAX_ACCELERATION >> # If using BACKLASH compensation STEPGEN_MAXACCEL should be 100% larger. >> STEPGEN_MAXVEL = 62.50 >> STEPGEN_MAXACCEL = 375.00 >> >> The acceleration values of the spindle are dramatically faster than >> the Z axis. >> [JOINT_2] >> TYPE = LINEAR >> HOME = 0.0 >> FERROR = 0.05 >> MIN_FERROR = 0.01 >> # 3.2:1 Max Speed >> MAX_VELOCITY = 2.50 >> MAX_ACCELERATION = 10.0 >> # The values below should be 25% larger than MAX_VELOCITY and >> MAX_ACCELERATION >> # If using BACKLASH compensation STEPGEN_MAXACCEL should be 100% larger. >> STEPGEN_MAXVEL = 2.75 >> STEPGEN_MAXACCEL = 20.0 >> >> >> >> >> _______________________________________________ >> Emc-users mailing list >> Emc-users@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/emc-users >> . > > Cheers, Gene Heskett. Cheers, Gene Heskett. -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author, 1940) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis Genes Web page <http://geneslinuxbox.net:6309/> _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
