Test Alignment

[geo0rpo]

Hello friends.

I am testing my bottom camera and after picking up a part I do the "test alignment".

The part is recognized 100% but the crosshair does not show the center of the part and the crosshair orientation does not match the part orientation.

Please look at the video.

https://www.youtube.com/watch?v=Cvxr-2U7y_0

[tony...@att.net]

I'm not positive what you're asking but have you told it to draw the center and show the orientation?  You probably have a DrawRotatedRects towards the very end of your pipeline, set both drawRectCenter and showOrientation to true.

Tony

[ma...@makr.zone]

The last computer vision image is frozen for a few seconds, while the cross-hairs keep responding to machine motion. After alignment, the part is centered and rotated, but the last computer vision image is still from the unadjusted rotation/offset of the part, i.e. when it was still slightly off.

If you look closely the two more or less align a moment later.

_Mark

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/f596e69a-a382-4aba-bcd6-efe5be95e3b9n%40googlegroups.com.

[geo0rpo]

Please look at the pictures. Both parts are recognized perfectly but are not placed right.

[ma...@makr.zone]

Have you checked Units per Pixels? It seems the machines compensates, but not nearly enough.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/61a88c7d-d7d9-49a2-b7e7-df406dbba4d5n%40googlegroups.com.

[geo0rpo]

Actually I set the units per pixel for the bottom camera with a graph paper but after doing so, the ruler cross hair did not match the graph paper steps (1mm).

So I manually set units per pixel to match the graph paper.

Is there an "optimal resolution" for the camera? I have set the bottom to 1280x720.

Also please tell me what should I expect from the test alignment?

Check the part and then align the part center and part rotation with the cross hair?

[Markus Oberle]

Problem with offset between camera and nozzle could also be an case.

I would also thing about nozzle calibration, but IMO the gap is too high

Camera/Nozzle offset evaluated?

[geo0rpo]

I will adjust the offset again but my main question remains that why does not the cross hair rotate to the rotation of the recognized part.

Also what do you mean the gap is too high?

[ma...@makr.zone]

> why does not the cross hair rotate to the rotation of the recognized part.

I explained why the rotation is not happening at the same time. Please re-read and tell me what you don't understand.

Also you have to realize that the cross-hairs show the nozzle rotation, not the part rotation. The Bottom vision aligns the part, by rotating the nozzle. The part will not "slip-rotate" on the nozzle.

What I cannot explain yet is the large linear offset you still get.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/dbed039a-2217-4af3-8d9a-0614d973bf09n%40googlegroups.com.

[geo0rpo]

Yes of course I understand that the part will not slip rotate on the nozzle.

What I thought should happen , was that the nozzle with the part on it, would center exactly the bottom camera cross hair and the bottom camera cross hair would match exactly the rotation of the part on the nozzle.

Look at an example I just did on Corel.

[geo0rpo]

This is what I thought I should see on the bottom camera after vision check.

The BOTTOM CAMERA cross hair "should" follow the part.

As about the offset I will calibrate everything again.

[geo0rpo]

I just loaded my example.png which I rotated 35 degrees in corel.

The nozzle indeed turns 35 degrees CC as seen in the DRO but the bottom cross hair does not follow but I guess this does not matter since the part rotation in recognized correctly.

[ma...@makr.zone]

> Yes of course I understand that the part will not slip rotate on the nozzle.

Then what don't you understand in the sentence: "you have to realize that the cross-hairs show the nozzle rotation, not the part rotation"?

As Tony said, you can enable some orientation indicators in the stage.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/0598ec28-36f1-4993-be2e-658e8218a15dn%40googlegroups.com.

[geo0rpo]

OK, I got it :)

I will adjust offsets again and retry.

Also I was wondering about the parts. Shouldn't the part bottom vision pipelines be at the "packages" tab and then just assign packages to the parts?

As it is now, there must be for example  a part for 0805 1K and another part for 0805 10K and another part for 0805 100K even if they have the same pipeline.

I think that if the bottom vision part pipelines were kept in the packages tab, then you could assign package type to the parts and the pipeline would follow. That is how it is in the Philips  assembleon machines.

I mean 0805 ceramic capacitors always "look" the same with (minor adjustments) and a 0805 resistors always "look" the same also.

What am I missing?

Sorry for the long post.

[ma...@makr.zone]

> shouldn't the part bottom vision pipelines be at the "packages" tab

Yes, absolutely, I started a new discussion here:

https://groups.google.com/g/openpnp/c/7DeSdX4cFUE

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/a938bde8-10c7-40e0-93e2-999267c9a5cdn%40googlegroups.com.