Tray feeders - big IC center

[Krzysztof Kozłowski]

For now I'm designating IC center using 4 points

Is there any quicker method to designate IC center in tray feeder?

Now I’m checking X/Y coordinate values, (Xmax-Xmin)/2 and (Ymax-Ymin)/2 gives me center values.

I’m using many tray feeders and in case of setting 10-15 feeders like this, it takes some time.

Maybe there is any more efficient way to do this?

It’s easy in case of small IC like SO-8, but large IC does not fit camera view.

[Michael Anton]

The commercial machine I have has the option to pick 2 or 4 points, and then it calculates the center.  I've generally found that picking 2 diagonal corners on the large parts is sufficient.  I'm not sure if OpenPnP has a similar method.

Mike

[Krzysztof Kozłowski]

That sounds like much quicker option.

--
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/01fe5f74-e6d9-4d46-8608-b81ab90af151n%40googlegroups.com.

[mark maker]

OpenPnP does not have a function to capture a mid-point via two corners. But it has a function to measure relative distances, which might help a bit. Click on the DRO, it turns blueish, and them it shows the relative distance from there when you jog away. Click again to go back to the absolute display.

But...

Can't you derive the pick location from the tray geometry? It would be a strange tray if it wasn't centering parts inside its grid cell, and if it did not use a well-known/documented grid unit (a round millimeter number), and if the tray was not carefully aligned along X/Y. 

Capture one corner, the rest should be a very simple "plus half grid unit" per axis.

Note, pick locations for large parts do not need to be super-duper precise. Parts usually have quite some wiggle room inside their pocket anyways, that's why bottom vision is needed, and that's why you can take it easy when capturing the pick location (or in this case the tray corner).

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CADPs_Pk_c-D_988v%2B8681d8wrkBwdSWGjWW1mmrfqqDixKL8%2Bw%40mail.gmail.com.

[Krzysztof Kozłowski]

DRO relative measurement might help a little - I didn't know about that feature.

In case of detrive center from tray geometry - I think I haven't seen factory tray which would have maked center point. It's good idea for me to mark it on my custom trays, but it is not really universal way to resolve this problem.

For example I attach photo of factory tray for QFP208 IC (this is one of examples which does not fit camera view).

Even hole under IC in this tray is too big to fit camera view.

You're right, those does not have to be 100% precise, but when IC is 32x32mm like in this case, it's easy to run out of pick tolerance - even if it's set quite large value.

[mark maker]

Photo seems not to work.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/db3a4ff4-59cd-4c34-a3f8-ad1b9dc055e5n%40googlegroups.com.

[Krzysztof Kozłowski]

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/b123900d-06b6-4e74-8313-b091b0c2af0d%40makr.zone.

[mark maker]

It seems to have a ~37mm grid, and surely you have a data sheet that gives you exact measure. Or you can measure relative across many slots using the camera and divide by number of cells.

Then capture one of these screw holes with the camera, that should be easy, keeping in mind that super-duper precision is not needed.

Then add half the grid size in each dimension. 

Subtract number of tray slots times grid to the left/top for the first slot.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CADPs_Pn8Ju79-58Mrw7QsYYO1vW_1uLXL0zdr%3DpJNi%2Bfr_dN2w%40mail.gmail.com.