Author Topic: Measuring A Simple Tool  (Read 1535 times)

designfugitives

  • Robots in Architecture Members
  • Forum User
  • *
  • Posts: 17
    • View Profile
Measuring A Simple Tool
« on: December 06, 2021, 08:23:23 PM »
I 3D printed a "unicorn" to use to learn the process of measuring a tool tip.  I had set the tool to 0X 0Y 228.6Z as shown in the first picture.  When I measure the tool using the 4-points XYZ reference I get a range of error anywhere from 100 up to 250 mm.  The current set up isn't very precise but I was expecting errors in the teens not hundreds.

Did I set the tool wrong initially?  Suggestions?

Johannes @ Robots in Architecture

  • Lead Developer
  • Administrator
  • Forum User
  • *****
  • Posts: 1311
    • View Profile
    • Robots in Architecture
Re: Measuring A Simple Tool
« Reply #1 on: December 07, 2021, 08:41:53 AM »
Hello,

For the 4-point calibration it should not matter what values the tool was set to before. You just pick a static reference point in your robot's workspace and then move the tooltip to it in 4 different orientations. With a small robot I can get to 0.1/0.2mm with a good tool, with larger robots usually to 0.5mm. Doing it very roughly is often around 3mm. So hundreds of millimeters definitely points to some problem.
Was your reference tool placed on a turntable and maybe moved during calibration?
You can move the robot in a straight line in T1, right? That would indicate that the mastering is generally OK, because a wrongly mastered robot could also lead to those results.

Best,
Johannes

designfugitives

  • Robots in Architecture Members
  • Forum User
  • *
  • Posts: 17
    • View Profile
Re: Measuring A Simple Tool
« Reply #2 on: December 10, 2021, 09:15:42 PM »
Joahannes

I understand the 4-point calibration is not tied to the tool values set.  I did that to see what the difference might be after doing the 4-pt calibration.

Some more strange behaviors.  I mapped a base on our turn table.  The origin was first set at the inside corner of the square.  First point was at the end of the long edge.  When I move the tip of tool to the origin after mapping the base, the location reading isn't 0,0,0.  Is it time for mastering?

My training was 5 years ago with little or no use on a robot since.  I might be misunderstanding some basics.

Johannes @ Robots in Architecture

  • Lead Developer
  • Administrator
  • Forum User
  • *****
  • Posts: 1311
    • View Profile
    • Robots in Architecture
Re: Measuring A Simple Tool
« Reply #3 on: December 12, 2021, 11:37:42 AM »
Hello,
If you are actively using the turntable, make sure to calibrate an Offset Base, as that is then rotating along with the turntable.
If you just used it as a static table without any movement and you get a different position when approaching the same XYZ coordinate from different angles, then you may really need to calibrate the robot.
Have you got the mastering kit, or do you need some advice for the "hacky" way to master it?
Best,
Johannes

designfugitives

  • Robots in Architecture Members
  • Forum User
  • *
  • Posts: 17
    • View Profile
Re: Measuring A Simple Tool
« Reply #4 on: December 14, 2021, 06:42:31 PM »
Static stable for now.

Hacky please?

Johannes @ Robots in Architecture

  • Lead Developer
  • Administrator
  • Forum User
  • *****
  • Posts: 1311
    • View Profile
    • Robots in Architecture
Re: Measuring A Simple Tool
« Reply #5 on: December 14, 2021, 10:09:09 PM »
So this is not the "nice" way, but it works good enough. But only do this at your own risk, of course.
Move the robot so that the white paint marks between the axes roughly match up. Unmaster all axes, then as the mastering option choose "Dial" (if I remember correctly). Dial assumes that you have a dial gauge with a thread that fits onto the calibration ports.
The calibration ports are usually behind a metal cap that you need to unscrew. There is a pin with a spring inside, and underneath a "V" shaped grove. The deepest part of the V is where the axis is calibrated.

If you have got the KUKA dial gauge or can fabricate one so that it fits, just move the robot very slowly to find that position. Be careful not to leave the dial gauge in when moving the robot faster.
You can also do it with a less accuracy by hand, by pushing gently down on the pin, e.g. with the back of a pencil. We had to do this a few times and it worked reasonably well. Ideally with two people as you need to press the deadman switch, axis direction and calibration pin at the same time.

You can also check the mastering to see if an axis is off.

Best,
Johannes

designfugitives

  • Robots in Architecture Members
  • Forum User
  • *
  • Posts: 17
    • View Profile
Re: Measuring A Simple Tool
« Reply #6 on: December 22, 2021, 05:58:06 PM »
Thank you for the hack.  I figured if you can do it manually with a dial gauge there must be a similar hack.  However, I think we will try to get an EMT. Can I used this - from ebay https://www.ebay.com/itm/324707156563?hash=item4b9a0e0253:g:61UAAOSwOOZfWlBp ?  Our robots are KR150 with KRC2-ed05.

What kind of values should I see in the robot axes when line up the mastering indicators?  See attached image for our current reading.

Can you undo Unmastering, such as reverting to last set of parameters?

Johannes @ Robots in Architecture

  • Lead Developer
  • Administrator
  • Forum User
  • *****
  • Posts: 1311
    • View Profile
    • Robots in Architecture
Re: Measuring A Simple Tool
« Reply #7 on: December 22, 2021, 10:08:27 PM »
Hmm... In your photo the robot seems to be at around 0,-90,90,0,0,0 and your screenshot says something very different - are you sure that they belong together?
Aligning the mastering indicators (the white/black marks) won't lead to perfectly "round" values, though.
As far as I know, there is only one set of mastering information with no way to revert. You can however check the mastering and then decide whether you take the new values.
Regarding the EMT, unfortunately I haven't ever used an EMT with a KRC2, only the hacky way. The tool for KRC4 (EMD, they changed the name from German to English) looks different because it uses EtherCAT as the bus system. 2.5k seem a bit much to me, though. I think they cost around 3k EUR completely new.

Best,
Johannes