Author Topic: ABB Revolution Counter  (Read 218 times)

0 Members and 1 Guest are viewing this topic.

Offline yannick

  • Newbie
  • *
  • Thank You
  • -Given: 4
  • -Receive: 0
  • Posts: 3
ABB Revolution Counter
« on: June 14, 2018, 01:47:06 PM »
Hi,

ABB IRB 120: I understand the revolution counter, as it's name says, only reset the counter for the actual number of turns the motor makes. So, as long as it's done within one motor revolution (visual alignment mark on robot arm), teached positions are not supposed to be affected.

Here's our problem: If we do an update on the revolution counter (Proper procedure with the alignment mark),  the already teached positions are not the same afterward. I've tried really small changes within the alignment mark and result is a change in the teached positions.

So: I'm a missing something? Do we have to do a fine calibration with the values on the robot base after an update to the revolution counters? Or is there something else I should know?

Thank you 

Offline Lemster68

  • Full Member
  • ***
  • Thank You
  • -Given: 0
  • -Receive: 26
  • Posts: 147
  • Ayotollah of the controlla
Re: ABB Revolution Counter
« Reply #1 on: June 14, 2018, 02:24:01 PM »
Did you verify that the calibration offsets are correct?  NEVER fine calibrate the robot axes unless you have changed a motor.

Offline yannick

  • Newbie
  • *
  • Thank You
  • -Given: 4
  • -Receive: 0
  • Posts: 3
Re: ABB Revolution Counter
« Reply #2 on: June 14, 2018, 03:25:32 PM »
I've checked the values of the calibration Offset after the revolution counter update and they're not the same as the ones written on the base plate. Should they be?

If so, it means I should always:
  1 - Do a revolution counter update base on the calibration marks AND
  2-  Manually correct the calibration offsets values so they're the sames as the values written on the base plate

Is that it?

Thank you

Offline Lemster68

  • Full Member
  • ***
  • Thank You
  • -Given: 0
  • -Receive: 26
  • Posts: 147
  • Ayotollah of the controlla
Re: ABB Revolution Counter
« Reply #3 on: June 14, 2018, 04:51:53 PM »
Yes, they should match.  Problem is that if you change them to match, all your programs will be off.  Hopefully, you don't have many.  First step is to correct the numbers, then update rev counters.  Also, if your TCP was taught with bad Cal offsets, it is probably bad too.  Make a backup before changing anything, just in case.

Offline yannick

  • Newbie
  • *
  • Thank You
  • -Given: 4
  • -Receive: 0
  • Posts: 3
Re: ABB Revolution Counter
« Reply #4 on: June 14, 2018, 06:58:35 PM »
Thanks again,

Just to make sure: I correct these numbers on the teach pendants to match those on the base and then do a revolution counter update? After that, if I have to do another update, positions won't change?

I get the idea of the revolution counter but I have a hard time understanding what is the offset exactly. A resolver should be very precise on one turn  and the counter keeps the count of every turn but the function of the offset?




Offline Lemster68

  • Full Member
  • ***
  • Thank You
  • -Given: 0
  • -Receive: 26
  • Posts: 147
  • Ayotollah of the controlla
Re: ABB Revolution Counter
« Reply #5 on: June 14, 2018, 08:26:02 PM »
 After that, if I have to do another update, positions won't change?  Yes, if you mean in the future after you have corrected your programs and then somehow loose rev counters.

I get the idea of the revolution counter but I have a hard time understanding what is the offset exactly. A resolver should be very precise on one turn  and the counter keeps the count of every turn but the function of the offset?  I am not sure, I think that it is the offset from commutation.

Offline SAABoholic

  • Full Member
  • ***
  • Thank You
  • -Given: 0
  • -Receive: 17
  • Posts: 126
Re: ABB Revolution Counter
« Reply #6 on: June 14, 2018, 11:51:12 PM »
Most ABB robot use resolvers for motor positioning (the 120 is one exception as it uses encoders) and the calibration offset is simply put a reference to the signal values that is read from the resolver when the robot is (fine) calibrated.

While it is true that you can update the counters within one motor revolution and it should pick up the correct 0-position, you should not confuse the visual marks on the robot with actual motor revolutions as the range depends on the gear ratio.

One possible issue is that on axis' with no mechanical limit (Ax.4 on some models and even more so, Ax.6) you can be a full 360 degrees off without knowing it.
Hopefully you have some kind of dresspack so you can visually figure out which direction you should go, but on models where both ax4 and 6 are free rotating(1600 for example) it's easy to get things mixed up.
It's been a while since I tinkered with IRC5 so I don't remember 100% but I do recall that IRC5 will (in most cases) retain the last know position and flash that value n the pendant.
But in cases where it doesn't or on S4, then for example, lets say your axis six was at 345 degrees when you loose your counters... well... was it at -15 degrees, 345 (or worse) 705 ?  ???

What you can do in this case is to update the counter, jog axis 4 (OR) 6 360 degrees in one direction (remember which) and update them again, if that doesn't help  you can go two steps the other way (one to get back to where you started, one to go one step the other way from where you started). If you're unlucky it might be both axis that are off so if you're lost it might be worth a bit of trial and error before having to pay to have someone come out and calibrate the robot.


Assuming your robot was properly calibrated to begin with you should be able to run your programs w/o any problems once it's re-calibrated.

Offline hermann

  • Sr. Member
  • ****
  • Thank You
  • -Given: 0
  • -Receive: 57
  • Posts: 273
Re: ABB Revolution Counter
« Reply #7 on: June 15, 2018, 12:39:06 PM »
One more hint:
After updating revolution counters (and doing some other stuff like changing the offset values...) try to measure a TCP (has not to be the real one) with the 4 point method. If you are not able to do this with an error lower than, let's say 0.4 mm or so, there is a problem with the calibration.

Offline Skooter

  • Hero Member
  • *****
  • Thank You
  • -Given: 36
  • -Receive: 91
  • Posts: 513
Re: ABB Revolution Counter
« Reply #8 on: June 16, 2018, 10:52:17 AM »
Also be careful when updating the rev. counter on axes that have fine calibration positions close to the zero crossing point.  An example would be fine calibration numbers above 6.0 and below 0.3.  Being slightly off can put you on the wrong side of the zero.  Move the robot to its calibration position and verify each axis after updating the rev counters.  Using a known reference point to verify is a good practice