Posts by mekeale

Quote from kwakisaki

mekeale

Remember you have come to Robot Forum for assistance.

Many members are allocating their experience and free time to assist you.

How much time are they spending writing replies in comparison to contacting the OEM and paying for a service/application engineer to assist.......just have a think about that for a minute.

Some of your responses do suggest you are just casually brushing suggestions aside, not necessarily actioning them.

Remember, we are not seeing what you are seeing, and if you cannot convey that to us in text, provide video, pictures and evidence of results from the suggestions offered.

If it continues, there is a possibility these members may seek life elsewhere rather than assisting you.

So I suggest with your responses, to try and provide some sort of evidence/proof that you are indeed trying these things as opposed to comments like 'I assure you the code is correct', 'everything is fine' and 'no I do everything you say'

I could just offer a suggestion by saying 'Your robot is broke, buy a new one', do you see my point?

I am not suggesting to remaster or teach anything, read what I have suggested and post back videos, pictures and results and we can advise further.

You and us are sitting in an area with MANY unknowns at the moment:

1. Fanuc Robot, uncertainty of mastering being correct.

2. Fanuc Robot, TCP no evidence of XYZ motions to prove TCP XY vectors are orthogonal to milling bit.

3. Fanuc Robot, TCP appears inaccurate for rotational, suggests TCP XY vectors may not be orthogonal.

4. No evidence of TCP teaching, showing how you are doing it, wrist posture, fixed pointer position etc.

5. Fanuc Robot, recent image show wrist posture in singularity.

6. Recent image, suggests setup conditions is not ideal for TCP teaching.

7. RoboDK, is the code correct, Fanuc robot will execute what you tell it......shit in, shit out......

So my suggestion is to eliminate the majority of these unknowns and see if there is a fundamental issue with the robot, then at least that can be eliminated and shelved.

Display More

Thanks again everyone for your answers and help, I appreciate everyone's time and thanks for the help. I will do everything that was written here and provide a photo, video and code if something does not work out again.

Quote from hermann

You are asking for help, and ignore nearly every tipp for enhancements.

My best wishes for solving your problem.

no, I do everything you say. Thank you.

Quote from kwakisaki

hermann is 100% correct here:

- Your wrist posture is in a ridiculous position for TCP teaching.

- Your spindle would also be better mounted at 45-90 from it's current position - forcing wrist posture.

- Your pointer/work is very central to the physical robot.

- Mr Robot prefers when ALL joints have freedom to move, your image suggests otherwise.

- Mr Robot will hate you for forcing him to teach/work TCP in these conditions, poor Mr Robot..

1. Can you post a videos of:

- Current setup actually milling showing the full tool and wrist posture during the cutting.

- RoboDK simulation showing the full tool and wrist posture during the cutting as control comparison.

** RoboDK is not a Fanuc application, so creating projects in RoboDK is no guarantee it's correct **

2. Have you tried removing the spindle.

- Move all joints to mechanical zero.

- Fitting a small pointer to central to the robot flange.

- Reset user frame to null.

- Reset tool frame and just have Z value representing the length of the pointer.

- Make sure your pointer is perpendicular to robot and bias more left or right and lower than current.

- Move to your pointer showing us the wrist posture.

- Using Tool and test Rx, Ry and Rz and show video of robot posture during Rx, Ry, Rz

- Post back video of you doing this, showing wrist posture.

- This may highlight any possible mastering issues as there is minimal external influence.

Display More

I did not say that I taught TCP in this position. In this position, the axes J5, J4, J3, J2 are being calibrated. I will do everything you wrote, I will inform you about the results. Thank you.

Quote from hermann

Is this seriously your measuring setup?

You want a precise result, so use at least half way precise measurement setup.

Attach your magnetic holder to something stiff, stable and magnetic.

Move it to a place at least 50 cm lower than now or more away from robot, so that the robot is not near singularity when measuring.

Try all again.

Display More

I agree. doesn't look very good. but everything is fine and does not move. I followed the instructions and set up the robot according to the instructions.

a photo...

tool frame i have only 3 points and 6 points

Quote from pdl

I believe you guys have your Endians crossed.

Erik is referring to teaching tool frames.

HawkME is referring to teaching user frames.

HawkME, I too did not know about this new method until I learned about it from Erik. It's in the HandlingTool manual, section 6.6.2.3.

what is the name of section 6.6.2.3.?

Thank you

Quote from Erik Olsen

Having very accurate payload and inertia values should help your robot's accuracy and repeatability. Would you try teaching a 4-point TCP and tell us what mean and max errors you are getting?

Is 4 points training more accurate than 6? Tomorrow I will re-master the payload and calibrate according to the above instructions, followed by 4-point training.

may affect incorrect PAYLOAD?

Quote from hermann

Seems that the movement goes through singularity. Try a different movement, is the deviation the same?

I'll try, but immediately after the calibration, I checked and filmed what happened. trained the tool in another place and rotated the 4th axis around Z, the offset is the same.

Quote from pdl

Вы переучивали свой TCP после этого?

Как вы этому научили?

Yes, I taught, 6 points.

Good day everyone. Did as described in Robot-Accuracy-Enhancement . calibrated J5, J4 calibrated several times, the result is on the video. rotation around the axis by 180 degrees. In my opinion, there are no changes after calibration.

Quote from pdl

Any of those mechanical unit manuals will have a complete section dedicated to each type of mastering.

RTFM

I'm still busy. When I have time, I'll use your instructions first. I'll post the results.

Quote from pdl

I'll make this a very specific question:

What Fanuc manuals do you have?

Here's what I have.

there is an instruction. I just didn't quite understand the calibration order. First, it is desirable to train the instrument and then follow theinstructions?

can you explain where to calibrate? where is the tcp training for 6 points or the master cal?

Quote from pdl

Can you post a picture of your current master counts from the tp screen? They are under dmr_grp in the system variables.

did not quite understand what you need, photo ...

and this

There is a learning guide. is it worth using it?

project in robodk