Introducing the HeapFeeder - the slowest feeder ever

285 views
Skip to first unread message

johanne...@formann.de

unread,
Mar 22, 2021, 3:48:44 PM3/22/21
to OpenPnP

Hi,

after some inspiration of the Team14 HeapFeeder (https://www.youtube.com/watch?v=rYFK_RLNU-g) that was never really finished, I put a bit of time in an implementation.
It is a high density feeder (about 3cm² for each feeder, around 500 1206 parts or > 3.000 0603 parts on that area) but really slow, since it requires many movements and lot's of vision.
My goal was for my liteplacer based pnp, that I can have most passives always on the table, since I mostly do very small numbers of pcbs, so the set up times are the biggest factor. (Probably with the time needed to develop this feeder, tweezers would have been faster for the rest of my life, but I hate placing parts with tweezers)

A pull request is planed in the near future, but till then for the guys with IPv6 (or the knowledge how to use an IPv6-proxy) I have written a short(?) introduction with a short video on my homepage: https://www.formann.de/2021/03/openpnp-heapfeeder/

greetings
Johannes

ma...@makr.zone

unread,
Mar 22, 2021, 4:00:30 PM3/22/21
to ope...@googlegroups.com
--
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/a9754432-2ba8-4632-a40a-9b2b51091673n%40googlegroups.com.

johanne...@formann.de

unread,
Mar 22, 2021, 4:05:21 PM3/22/21
to OpenPnP
Strange.
As an ugly work around the pictures attached.
HeapFeeder_FeederConfig.png
HeapFeeder_Overview.png
HeapFeeder-DropBoxSettings.png

Clemens Koller

unread,
Mar 22, 2021, 6:05:52 PM3/22/21
to ope...@googlegroups.com
Bulk case feeding is somehow an interesting thing in the long term.
I believe the prices of these components will become really really low.

There is an ASM marketing video from 2015: https://www.youtube.com/watch?v=fuOnD91CKdQ
At 2:20 she says "customers are not yet present for this technology, but..." 8-)

I am wondering of what becomes possible, if the components are feed directly into the PnP head to reduce movements (= energy, costs).
That's one of my ideas. Any patents pending? 8-)

Clemens
> --
> 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 <mailto:openpnp+u...@googlegroups.com>.
> To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/a9754432-2ba8-4632-a40a-9b2b51091673n%40googlegroups.com <https://groups.google.com/d/msgid/openpnp/a9754432-2ba8-4632-a40a-9b2b51091673n%40googlegroups.com?utm_medium=email&utm_source=footer>.

Arthur Wolf

unread,
Mar 22, 2021, 6:31:36 PM3/22/21
to ope...@googlegroups.com
The way this would allow for massive parts catalogs within the work area is extremely cool. Definitely worth the time cost for some types of uses/businesses where setup is a large part of the work.
With some optimizations, something like 20x20cm would be 400 part types, try to beat that with spools... and you can pretty easily use a few motors to stack things up in the Z axis too

Have you thought about having a separate controller/xy axis actually just dedicated to a catalog built like this, that would just prepare parts in advance in a shared area? (running in parralel to the actual job, or running before the job as an automated setup process). How hard would that be to support in OpenPNP?

I really see a ton of potential in this, at least for how we are planning to use OpenPNP in the future, this is definitely something we are looking forward to setting up.


--
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/a9754432-2ba8-4632-a40a-9b2b51091673n%40googlegroups.com.


--
勇気とユーモア

johanne...@formann.de

unread,
Mar 22, 2021, 6:51:40 PM3/22/21
to OpenPnP
The only way I currently can think something like a independent xy-system with an shared area could work would be, if you define in the shared area some feeders, and then select for each feeder one of the "heaps" as source.
But that would limit the number of parts, that could be used in a single job, also probably a huge pain how to get that running asynchronously  to the main process without the risk of collisions of both heads...
Would not say impossible, but I guess not (much) easier to get openpnp to support multiple independent heads at the same time.

Arthur Wolf

unread,
Mar 22, 2021, 7:12:14 PM3/22/21
to ope...@googlegroups.com
On Mon, Mar 22, 2021 at 11:51 PM johanne...@formann.de <johanne...@formann.de> wrote:
The only way I currently can think something like a independent xy-system with an shared area could work would be, if you define in the shared area some feeders, and then select for each feeder one of the "heaps" as source.
But that would limit the number of parts, that could be used in a single job, also probably a huge pain how to get that running asynchronously  to the main process without the risk of collisions of both heads...

Just have a small belt between the two work areas, you don't need to have any risk of the heads colliding. You could have two separate machines really, one that has 10000 heaps, gets informed about what will be needed, places that on a belt, and then a belt that feeds that to the actual pick and place machine.

I feel like it might be possible to make the "catalog" machine cheap (and potentially compact) enough that you might have *several* of them next to your pick and place machine, essentially *partially* taking the place/role of some of the feeders.
Multiplying those machines (if they are inexpensive enough) would solve/reduce the speed issue, as everything could be done in parralel.

On the software/async side, it sounds like a pretty manageable issue to solve/work on, just have to have some sort of way for the machines to communicate and manage a "queue" of parts, so they know when parts are available or when they need to wait for some more to be added to a given belt.

Would not say impossible, but I guess not (much) easier to get openpnp to support multiple independent heads at the same time.

The way I see it this would be a separate dedicated "catalog" system, probably based on openpnp source-wise, but really meant just to do the "heap" procedure and that's it.


Arthur Wolf schrieb am Montag, 22. März 2021 um 23:31:36 UTC+1:
The way this would allow for massive parts catalogs within the work area is extremely cool. Definitely worth the time cost for some types of uses/businesses where setup is a large part of the work.
With some optimizations, something like 20x20cm would be 400 part types, try to beat that with spools... and you can pretty easily use a few motors to stack things up in the Z axis too

Have you thought about having a separate controller/xy axis actually just dedicated to a catalog built like this, that would just prepare parts in advance in a shared area? (running in parralel to the actual job, or running before the job as an automated setup process). How hard would that be to support in OpenPNP?

I really see a ton of potential in this, at least for how we are planning to use OpenPNP in the future, this is definitely something we are looking forward to setting up.


On Mon, Mar 22, 2021 at 8:48 PM johanne...@formann.de <johanne...@formann.de> wrote:

Hi,

after some inspiration of the Team14 HeapFeeder (https://www.youtube.com/watch?v=rYFK_RLNU-g) that was never really finished, I put a bit of time in an implementation.
It is a high density feeder (about 3cm² for each feeder, around 500 1206 parts or > 3.000 0603 parts on that area) but really slow, since it requires many movements and lot's of vision.
My goal was for my liteplacer based pnp, that I can have most passives always on the table, since I mostly do very small numbers of pcbs, so the set up times are the biggest factor. (Probably with the time needed to develop this feeder, tweezers would have been faster for the rest of my life, but I hate placing parts with tweezers)

A pull request is planed in the near future, but till then for the guys with IPv6 (or the knowledge how to use an IPv6-proxy) I have written a short(?) introduction with a short video on my homepage: https://www.formann.de/2021/03/openpnp-heapfeeder/

greetings
Johannes

--
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/a9754432-2ba8-4632-a40a-9b2b51091673n%40googlegroups.com.


--
勇気とユーモア

--
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.


--
勇気とユーモア

johanne...@formann.de

unread,
Mar 28, 2021, 6:05:17 PM3/28/21
to OpenPnP
After it has been merged, I added the short documentation also in the wiki (the same content, except the videos):

Martin Gyurkó

unread,
Mar 29, 2021, 2:53:29 AM3/29/21
to OpenPnP
Hey Johannes,
I love it.
Thanks!
Exactly what we need here.
Can anyone tell me where to buy the blister packed SMD parts that are shown in the bulk feeder video in the ASM marketing videos?
No need to ever hassle around with the roll feeds.
Bye,
Martin

Randy Park

unread,
Sep 14, 2022, 11:41:06 AMSep 14
to OpenPnP
I built our OpenPnP machine pre-pandemic, then the "large" order it was intended for vanished, so haven't used it much. Most of our builds are 5 to 10 boards and setup time is always an issue. PnP was great after it was set up, but...
So I LOVE the idea of the bins and the heap feeder; placing time is not an issue for us, but setup time is.
200 heaps would cover almost all the common parts we use, especially for resistors and capacitors; OR we could set up sets of heaps for each product. Also could be used for inventory by storing sets of heaps in drawers.

QUESTIONS:
Johanne, do you have any updates on this?
Anyone else, have you implemented, any tips?
I'd be happy to contribute my experiences if I start working on this.

Ravi Ganesh

unread,
Sep 15, 2022, 5:01:15 PMSep 15
to OpenPnP
Feeder Groups enables you to setup and finish jobs faster.
More of it in this thread

HeapFeeder is a great idea. But unfortunately my bed area did not permit this family.

Cheers




Johannes Formann

unread,
Sep 29, 2022, 6:43:00 PMSep 29
to openpnp
Hello Randy,

sorry for the late response, I'm a bit behind on the mailing list.
Have it running for my passives (resistors, capacitors) and I'm quite
happy after impleneting the "poke" option for capacitors. (made it even
slower, but avoided damaging them when they lock together).

If you have questions, feel free to ask.

Greetings
Johannes
> <https://www.youtube.com/watch?v=rYFK_RLNU-g>) that was
> never really finished, I put a bit of time in an implementation.
> It is a high density feeder (about 3cm² for each feeder,
> around 500 1206 parts or > 3.000 0603 parts on that area)
> but really slow, since it requires many movements and lot's
> of vision.
> My goal was for my liteplacer based pnp, that I can have
> most passives always on the table, since I mostly do very
> small numbers of pcbs, so the set up times are the biggest
> factor. (Probably with the time needed to develop this
> feeder, tweezers would have been faster for the rest of my
> life, but I hate placing parts with tweezers)
>
> A pull request is planed in the near future, but till then
> for the guys with IPv6 (or the knowledge how to use an
> IPv6-proxy) I have written a short(?) introduction with a
> short video on my homepage:
> https://www.formann.de/2021/03/openpnp-heapfeeder/
> <https://www.formann.de/2021/03/openpnp-heapfeeder/>
>
> greetings
> Johannes
>
> --
> You received this message because you are subscribed to a topic in the
> Google Groups "OpenPnP" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/openpnp/3okTMYjBWio/unsubscribe
> <https://groups.google.com/d/topic/openpnp/3okTMYjBWio/unsubscribe>.
> To unsubscribe from this group and all its topics, send an email to
> openpnp+u...@googlegroups.com
> <mailto:openpnp+u...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/openpnp/0d550d8f-3bc5-49fe-949b-1d6932107ff5n%40googlegroups.com
> <https://groups.google.com/d/msgid/openpnp/0d550d8f-3bc5-49fe-949b-1d6932107ff5n%40googlegroups.com?utm_medium=email&utm_source=footer>.

Reply all
Reply to author
Forward
0 new messages