BlindsFeeder - Blinds open/close - Nozzle pushes catching on edge of blinds as it decends

[uavfligh...@gmail.com]

When pushing a blind to open/close there is no horizontal gap between the nozzle and the blind as the nozzle descends towards.  This often results in the nozzle pinging the end of the blind and popping parts out of the pockets.  That or it lands on top of the blind and slides along it, resulting in a bad open.

The code here shows that position is exactly the nozzle diameter plus push distance.  No tolerance included at all.

https://github.com/openpnp/openpnp/blob/370b98b3043fa4b735039899ea1ebee76621ddc3/src/main/java/org/openpnp/machine/reference/feeder/BlindsFeeder.java#L1619

Had this problem a long time and only just figured out what is happening.

I could add a fudge factor to the nozzle diameter but that feels like the wrong place to do it.  Modifying the existing behavior will probably mess up lots of setups.

Is there another workaround?

/Regards Matt

[uavfligh...@gmail.com]

Quick update.  Just changed to:

.add(nozzleTipDiameter.multiply(0.75)))

and it makes all the difference

[mark maker]

Hi Matt,

This should not happen, it specifically goes 0.5 x sprocket pitch = 2mm further than the edge, on top of half the nozzle tip diameter.

I suspect your Edge Distances here (see screen shot) are wrong, perhaps because your edge calibration eloped at some time. It can happen, usually due to vision problems (see below):

Vision problems tend to be related to these points (available in the testing version):

1. Camera White Balance
2. 3D Units per Pixel, then setting Feeder Z precisely.

Note: the default 2mm Edge Distance is correct under the assumption that you printed the BlindsFeeder with a length of a multiple of 4mm, as documented. For other lengths, you need to adjust the Edge Distances accordingly.

_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/84c1ffee-8863-4e78-9abf-f7129ea39078n%40googlegroups.com.

[uavfligh...@gmail.com]

Mark,

I don't get it.  Both X0 and X1 have edge distance added.  Edge distance trims the final position of the blind for open or close.  The difference between open and closed is sprocketPitch.multiply(0.5).  2mm for 4mm pitch.  That leaves no extra space for the nozzle.  What am I getting wrong?

Tape lengths are 84mm.

/Matt

[mark maker]

Hi Matt,

I can't believe it, I think you're right, the half part pocket pitch distance is missing. But I don't understand how this could ever have worked. I must have tested with only 2mm and 4mm pitch tapes, where this just barely works. But I thought multiple users are using these feeders for larger parts... apparently not.

I will fix this fast.

Can I assume you can help test it physically right away? My machine is not currently ready.

Thanks for pointing this out and for insisting.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/8d46d542-629d-48f9-8516-90c1e633aef7n%40googlegroups.com.

[mark maker]

This should now be fixed in the testing version.

https://openpnp.org/test-downloads/

See the PR:

https://github.com/openpnp/openpnp/pull/1418

Wait a few minutes for deployment.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/466ac05f-60bb-4afa-a910-34e8e33dd5f2%40makr.zone.

[uavfligh...@gmail.com]

Thanks Mark.  Good to know that I was not having a non existent problem.

I am using a much bigger nozzle to do the pushes.  It has deeper straight sides.  You might have been getting away with that finer nozzle since it kind of tapers.

I am in mid production.  Will test your fix once I get that done.

/Matt