Hello Guest it is May 27, 2019, 05:45:41 AM

Author Topic: Spindle Lock Error G76 Thread Cycle  (Read 7732 times)

0 Members and 1 Guest are viewing this topic.

Re: Spindle Lock Error G76 Thread Cycle
« Reply #20 on: February 11, 2014, 07:12:09 AM »
Yes, having some synchronisation problems on this PC so the error seems logical.  It seems to me that Mach3 is detecting some timing errors and therefore kicking out these lines in the error log file. The thread is indexing ok to start but something is interfering with the progress through the thread length and it sometimes changes pitch toward the end of the thread.  Which is why this error is so intriguing, It may be the PC has its own timing issue.  Is there anyone that can determine what part of the program actually produces these error condition messages?
LP

Offline Hood

*
  •  25,828 25,828
  • Carnoustie, Scotland
    • View Profile
Re: Spindle Lock Error G76 Thread Cycle
« Reply #21 on: February 11, 2014, 02:28:04 PM »
I will see if Art will have a look, he is the only one that will know.
Hood

Offline ART

*
  • *
  •  1,696 1,696
  • Tough as soggy paper.
    • View Profile
Re: Spindle Lock Error G76 Thread Cycle
« Reply #22 on: February 11, 2014, 03:13:58 PM »
Hi Guys:

 Well, lets see what it says..

>>Mon - 22:31:31 ---Spindle Lock: 4721 , Total Ints: 0, Total Spindle Ints: 0, Error: 728, Adder: 0 LastRot: 5830
Mon - 22:31:31 ---Spindle Lock: 4721 , Total Ints: 5221, Total Spindle Ints: 4721, Error: 500, Adder: 0 LastRot: 5964
Mon - 22:31:31 ---Spindle Lock: 4721 , Total Ints: 10606, Total Spindle Ints: 9442, Error: 1164, Adder: 0 LastRot: 5876
Mon - 22:31:31 ---Spindle Lock: 4721 , Total Ints: 15426, Total Spindle Ints: 14163, Error: 1263, Adder: 0 LastRot: 6009
Mon - 22:31:31 ---Spindle Lock: 4721 , Total Ints: 20116, Total Spindle Ints: 18884, Error: 1232, Adder: 0 LastRot: 5937
.....

  Ok, so the spindle was seen to be running at 4721 interrupts per rotation with no load just prior to threading..( thats what the spindle lock means). Not sure of your interrupt freq., assume its 25K so you have
40us per int so .188 seconds per rot for 5.29 per second or approx 320RPM. ( again, assuming 25K interrupt time, always use the lowest interrupt freq you can get away with for stabilities sake..).

  Then you started the trhead. Amazingly it ran slow even before it hit the material.. it took 5221 on the first pass, so Mach likely slowed the axis to compensate..it kept slowing every rotation as the
error kept building. It never really catches up to the expected time. Really, the Total Ints number should try to stay near the Total Spindle Ints. Mach tries to slow things down to make that so. The number
under Error: is the actual number Mach uses to slow down. So its slowing every rotation and by a variable amount.

  Try running a thread in the air, the numbers should be much closer.( Total Ints: *********, Total Spindle Ints: yyy, ) If they are not , then you probably have either noise ( because of a 3..5 volt port maybe being ore sensitive?), or the interrupt
on that pc isnt very stable. Your driver test should display spikes or other patterns instead of a straight line if thats so. The printout seems to simply show that Mach doesnt seem to be able to track very
closely though it IS trying to. The problem seems to be there is too much variation. Id suspect noise or PC interrupt stability as the root issue if the threads wander..

Art
Re: Spindle Lock Error G76 Thread Cycle
« Reply #23 on: February 11, 2014, 03:53:22 PM »
Got it, thanks Art.  Think you are right with your prognosis. This was run in the air!
On the case now!  I will get back with the results as soon as complete or abandoned!
What does the LastRot: parameter signify?
 
I will focus on the Spindle PWM stability first, as it ran slow before hitting the material and get my scope on the pins and trace where the fluctuation is coming from.  This really helps with the starting point for the analysis.  

Thanks Hood for your input too.
LP

Offline ART

*
  • *
  •  1,696 1,696
  • Tough as soggy paper.
    • View Profile
Re: Spindle Lock Error G76 Thread Cycle
« Reply #24 on: February 11, 2014, 06:08:38 PM »
LP:

  Lastrot is the int's of the last rotation counted.

NOTE:  I recommend if threading that you do not allow Mach3 to pwm control the spindle. The reason for this is that most spindles these
days have their own internal regulator so Mach tries to control the speed at the same time and it can create instabiility as the twp controls
fight each other. Most will find stability in threading increases drastically if you set a static speed in the invertor alone.,

  I recommend allowing the operator to dial in a speed for the spindle, and not to let MAch3 control anything other than on/off.

Art
Re: Spindle Lock Error G76 Thread Cycle
« Reply #25 on: February 11, 2014, 06:55:49 PM »
Hi Art,
I am using the Little Machine Shop Spindle interface for the Sieg brushless motor spindle drive.  Current setup uses a PWM output from Mach3 through a low band pass filter and into the 0-5 volt AVI on their board. There is no RPM state or one time frequency to set a spindle speed, so I am pretty sure there is no battle going on between two regulators here. But, thanks for your point and may move over to a Mobus control instead a bit later.  After all, I do get a good thread of the first PC, it is only the second PC that is producing the Spindle Lock error information and causing threads to come out wrong.

FYI
The interrupt handler on the PC is working fine according to the driver test.
Ok there is a significant difference here, the LastRot shows about 30rpm difference, under no load, recon you are right with the interference thought.
Investigation so far, have found a level of background noise on the Index pin, coincidently a shade lower frequency than the spindle pulses at 250rpm, in fact, hit the spindle button, turn the spindle off while the thread cycle is running and the thread still engages after seeing a trigger, the spindle is stooped so the trigger should never arrive. The Index LED stops flashing while the spindle stops. The interference is a spike that lasts a tiny amount of time, much higher voltage so I am guessing the spike is enough to trigger a thread pass.

Still think this is a combination of issues. Found a GND inconsistency on the controller, Opto-isolator with active low not linked to the PC GND.  Needs to be active high and linked to PC GND then filter some of this noise out with strategically placed caps and ferrite rings.    Now I understand the Spindle Lock errors I can test and check the results shown in the log.

No need to reply.
LP

Offline ART

*
  • *
  •  1,696 1,696
  • Tough as soggy paper.
    • View Profile
Re: Spindle Lock Error G76 Thread Cycle
« Reply #26 on: February 11, 2014, 07:52:15 PM »
good luck. :)

Art

Offline RICH

*
  • *
  •  7,334 7,334
    • View Profile
Re: Spindle Lock Error G76 Thread Cycle
« Reply #27 on: February 12, 2014, 06:04:05 AM »
LP,

- RPM stability along with adequate motor torque is most important for threading.
  1 or 2 rpm difference and the thread should be ok. I get an occasional flicker from whatever it is set at to 1 rpm under.

- When or if using the threading Wizard, click the "calc number of passes" as that will show you if your exceeding your velocity.
   Even if it is ok, leave some head room from the max.

- As Art has replied, set the speed manualy. Play as you wish, but, your on your owne.
  Would venture to quess that anything that is manipulating the rpm for output to Mach is not in real time and thus must be totaly repeatable.
  Just a matter of degree as how much influence it will have on the actual threading.

- Scribing is a very effective way of seeing results and can tell you just what your machine is capable of doing since it's a visual of air running.

FWIW,

RICH


Re: Spindle Lock Error G76 Thread Cycle
« Reply #28 on: February 12, 2014, 06:52:27 AM »
Thanks RICH, good advice.