Posts by Ceta11

None of the links work.

Response to selection FYI:

'404 Attachment not found'

I would reinstall KUKA KSS... rule that out. Make sure you choose the robot type per sticker.

I wanted to have a permanent Start Type, ie: Cold Start, but doesn't retain the settings (saved) and for the next shut down it defaults back to Hibernate...

When is on Hibernate it doesn't want to start right, including many times brings a WINDOS BSoD.

Batteries are good and charged. I replaced the Motherboard battery as well.

I tried different memory and no difference.

I think Leon meant 0.05 m/s (3000mm/min) over soft plastic like materials should be easy. But there is always a learning curve.

The material we are talking here is quite soft, no external forces have any significance, and error on position or tolerances of + - 1mm are quite achievable, provided good practices are exercised.

It is and has been quite a learning curve for me to learn how important all the details are... just like in any machining.

I always remember (and keep in mind) many decades ago, that the one machinist in my home town that had the oldest and more basic lathe in town, whenever you wanted an accurate part within tight tolerances you will take it to him. You can recognize a message there.

Excellent. Look somewhere else.

Note: and again jogging the robot produces the same X-Y pulsing than during program running.

Here is a small piece of a code for rough cut on foam I just did.

Thanks for your comments and ideas.

Robot mastering can create that kind of errors in my case that is what happened. So I would take a manual Micrometer and do a remastering. That done should give you a reasonable accurate Robot... and would point out tours your EMT tool.

This Load parameter is for Dynamic positioning, but it should reflect your Tool configuration, ex: B: -90 deg. ....if your spindle is at 90 deg. you know that.

The pic shows that you have not positioned the CG correctly, it is coincident with flange face center point and the spindle is pointing straight out normal to robot flange...

For added dynamic precision is a good idea to add the inertial values. Is that the case ?

IF your error is static... then check your Config file in the software you are using... and in the $config.dat in the System folder I believe.

Other than that.. maybe your RDC card ? Mine died would NOT master anymore ... so I replaced it robot was unusable.

Unfortunately, my KR210 doesn't have AA option ...as far as I know... don't see any sticker. New River Kinematics' product is for position accuracy, not what we talk here, a dynamic behavior of the robot+controller's control software.

I am aware of the Control may have been calibrated for 50-75% load, I see that as a logical way of setting up for basic use of the rated Load. I lieu of this reasoning, I tried loading a parameter of load of 75kg mass at 3... you guess it right the pulsing was worse. I also tried negative values and although message says 'illegal' on screen I see behavior changing... but as you know it doesn't seem to have an even behavior across the distances. I think is interesting.

What I believe, because I have seen it done, is that the Control algorithms have gains parameters that can be changed. This is probably the most influential one that they tune at factory.... AND I suspect they may have a combo that can provide me...albeit not perfect but far away better than the pulsing I see at certain speeds.

I also can see videos posted by member Mr. V. Velikov in Youtube that show a smooth running milling path X and Y ... i.e. no pulsing at similar speeds where I get it.

I just have to find the right people in Engineering at KUKA or elsewhere to get me the set up. If as you mention there are parameters for elastic constants, there are also many structural and system damping, etc in the Control that are adjustable. We know that is what a Control Engineer does when designing a system.

I will keep you posted. Regards.

Well the response from the manufacturer support was less than satisfactory, to put it in a nice way.

* "your robot will never run smooth because is designed for higher load and you just have 9 kg of load".
* positionally accurate option software and installation, calibration and comissioning will cost $10K +
* "but your robot still needs to work with added weight to be smooth, even with that option".
* you are running too slow (20 mm/s) you should run at 60 mm/s (this in spite of having provided information to them and clear explanation that amplitude is larger at 40 mm/s)
* "your load is only active in PTP not in LINEAR, this only is activated with the software option"

I said to him that this type of robots are used in Europe and the US without adding ANY additional load, I have seen it and there are videos everywhere.
I tried to ask several times what the Control variables are available to adjust for light load... He could not answer and would deviate to other subject.
Load IS active in LINEAR instructions, I can see that the robot acts differently if I use different load parameters.

Maybe you can have a good lough at this... but the fact is that the support guy was not very knowledgeable of Control's Engineering, to me he was just an installer/integrator with "a lot of milling installation experience"...

Maybe his Engineering background was not there. Period.

It would be nice to get some comment from a knowledgeable european KUKA Engineer. I asked for that: Could you please consult with KUKA Germany...?

This is were things are so far.

Yes indeed. I also triple checked for the torque tightness of bolts.

We will see what KUKA has to offer.

There is no sticker "absolut vermessen"...

I did replace the R1 contents (machine + robot dats) with the ones stored on D: and the pulsing is gone apparently up to 75% TCP... which is a noticeable improvement. They are the same files that were supposed to be, but somehow were not in good shape.

I also contacted KUKA and awaiting what they suggest.

At least things run better so far. Thanks !

Yes shutting thru Win keeps your mastering fine.

What I do safely:
1) Safety Stop on KCP (very important when I got Communication errors)
2) Shut down thru windows, select Cold, Warm or Hibernate
3) Switch OFF

This has been the best for me.
Win XPe 2.2 KSS 5.6.10

Velichko, Thanks.

The pulsing happens not because of software but under higher speeds, i.e. running with pendant or with soft, on the X-Y axis only, not Z. I already corroborated this at lengths.

Individually all axis are silky smooth as well and Loads have been checked correctly.

I can email you a program but I am sure is not the problem. Just let me know if you have other ideas.

Thanks. I did not rushed into programming at all, tried to learn and follow Kuka recommendations on everything.

OK, yes I had specified Loads, Mass and Inertia are well calculated. I had done that well for many years in aerospace engineering. Besides It is 0.00kg mass with a fresh load of KSS. I had recently replaced the RDC card that died and configured loads (10kg spindle and plate on 3, 0.0 on 1 and 2kg on 2 at proper CG).

I am not aware of any milling software inserting any Loads, it uses what is on the configuration... but I checked to make sure. So Loads are what is configured after running a program.

It seems to be more than that... I suspect. It is like a 2-4Hz pulse that changes amplitude with speed.

are you sure your loads are correct?
A- forgive my ignorance, what are 'loads' ? electrical ? ...my 'loads' inputs in soft and real are very benign for the foam I am cutting, low engagement, low linear speed, etc. Besides, this X-Y pulsing is happening even without running spindle.

are you sure your programs are using them?
A- I can see the speeds are correlated to soft input, ex: 15, 20 or 30 mm/s with 4mm depth and cut over 12mm dia tool.

If other please explain. Thanks a lot.

OK, my question is 2 fold, is Hardware or Software ?

Software
This happens only in manual moving, in cartesian ?
A- on both manual and Auto. So it happens in Manual at 30% and up and under a program if I run at the same speed. So I have been forced to run at very slow speed for milling.

Have you made changes in machine.dat parameters ?
A- None. It is as it was. The MADA and machine.dat files has been KUKA checked. Robot is mastered and tools calibrate better than 0.8mm

Hardware
How many working hours has your robot ?
A- 15k plus. was given a written clean bill by Kuka 200 hrs ago.

What if move the axis rotating each one, still same shaking?
A- Each axis individually move sillky smooth at any position thru its whole range.

I could suspect some parameters in machine.dat that could make one or more axis to pulse if they are not well setted.
A- Pulsing when moving in Y or X direction is the best description on what seem to happen, amplitude is directly related to speed.

I don't know how to check the values on machine.dat other than replacing it with original file along with the MADA.

Any advice? Thanks a lot.

My KR210 + KRC2 with KR 5.10.6 robot, only when axes are streched is not moving smoothly manually at 100% but in small shaking steps increasing in amplitude as it streches.

Reducing to 30% or better 10% is the only way to minimize it. This is an annoyance when milling with some degree of precision and very slow time consuming.

Since this only happens away from a 1.5m radius approx. I believe that is not gears related.

Any advice as to what is or could be the cause ?

Thanks.

I want to upgrade my KRC2 controller memory to full 3x 512mb chips, and would appreciate your input on memory quality and reiability... other than KUKA's.

My KRC2 is a 2004 with SOYO PCI MBd and handles PC133 memory.

Another question of interest: Since KUKA VX is so sensitive as real time OS, why we are not using ECC memory instead ? Any experiences ?

Thanks a lot.

Play with 1) rapid move heights and 2) links.

See what you get...

Did reset the MFC card and DSE, nothing different. All the LED are green.

A lot of reading here and one opinion tell me that the RDC board could be the culprit.
Upon start the lights flash red and solid green, then stops and show solid red and green. This is the correct behavior according to Manual.

Could this card still be faulty ? wish I had one to swap... but nothing near.

What has been your experiences ? Thanks a lot.