Triggering peel from script without feed.

[vespaman]

Hi all,

Recently, I have been testing to pull/expose four pockets with 0402 instead of the usual 2.

My machine has drag feeders, so it takes a significant time to pull the tape for every two pockets (machine also has two nozzles), so I have been testing pulling 4, to see if this would cause more issues (like jumping components etc), but it doesn't look to be a problem. I have not run too many boards, so verdict is still out.

However, while doing so, I realized that it would perhaps be a little bit safer to only pull 4 pockets (8mm) without peeling, arriving to two exposed pockets, leaving the remaining two still covered. 

Looking like this after the actual feed cycle;

The remaining two needs a peel the next time the head is on its way.

I suppose this could be scripted, but thinking about it - it is harder to accomplish than I first envisioned.

I don't want a peel on the same cycle as the feed, so if I used e.g. "feeder after feed" scripting event, it would somehow need to keep track of that. 

So this is the question -  if there's a way to accomplish this, that I haven't thought of?

If not, I'll keep it as is, to see if it causes any side issues, but I figure it is similar to strip feeders that are also kept open, without causing any issues afaik. And I will only do this on passives, perhaps even limiting to the most used components (100n etc), at least to start with.

If the machine would allow dragging 12 or even 16mm, I'd try that as well, but there's a limit to drag max 8mm on 8mm feeders unfortunately.

 - Micael

[Jan]

Hi Micael!
So you're trying to mimic the window of Yamaha feeders? Very
interesting idea. Didn't you previously tried to add some bed below the
picking area to prevent kicking out other already exposed components?
Concerning your actual questions: at present all feeders receive a feed
request for every pick. It's just the question what the feeder does with
it. Most (all?) feeders will at least increment their Feed Count
property. I have not studied the code, but I'm petty sure the
ReferencePushPullFeeder uses some modulo math between Feed Count, Part
Pitch, Feed Pitch and Multiplier to decide whether a push-pull cycle is
actually needed. You could do the same in a Feeder.BeforeFeed or
Feeder.AfterFeed script and decide if/when to peel.

Jan

On 19.10.2025 15:30, vespaman wrote:
> Hi all,
>
> Recently, I have been testing to pull/expose four pockets with 0402
> instead of the usual 2.
>
> My machine has drag feeders, so it takes a significant time to pull the
> tape for every two pockets (machine also has two nozzles), so I have
> been testing pulling 4, to see if this would cause more issues (like
> jumping components etc), but it doesn't look to be a problem. I have not
> run too many boards, so verdict is still out.
>
> However, while doing so, I realized that it would perhaps be a little
> bit safer to only pull 4 pockets (8mm) without peeling, arriving to two
> exposed pockets, leaving the remaining two still covered.
>
> Looking like this after the actual feed cycle;

> PXL_20251015_070100128.jpg
>
> The remaining two needs a peel the next time the head is on its way.
> I suppose this could be scripted, but thinking about it - it is harder
> to accomplish than I first envisioned.
>
> I don't want a peel on the same cycle as the feed, so if I used e.g.
> "feeder after feed" scripting event, it would somehow need to keep track
> of that.
>
> So this is the question -  if there's a way to accomplish this, that I
> haven't thought of?
>
> If not, I'll keep it as is, to see if it causes any side issues, but I
> figure it is similar to strip feeders that are also kept open, without
> causing any issues afaik. And I will only do this on passives, perhaps
> even limiting to the most used components (100n etc), at least to start
> with.
>
> If the machine would allow dragging 12 or even 16mm, I'd try that as
> well, but there's a limit to drag max 8mm on 8mm feeders unfortunately.
>
>
>  - Micael
>

> --
> 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 visit https://groups.google.com/d/msgid/openpnp/
> e641e038-ce7d-4c90-a938-97512c4a462bn%40googlegroups.com <https://
> groups.google.com/d/msgid/openpnp/e641e038-ce7d-4c90-
> a938-97512c4a462bn%40googlegroups.com?utm_medium=email&utm_source=footer>.

[vespaman]

Hi Jan

So you're trying to mimic the window of Yamaha feeders? Very

Interesting, I did not know Yamaha did something similar! (Not that I thought I would be the first with this idea, I was pretty sure this had been used/discussed before).

 

Didn't you previously tried to add some bed below the
picking area to prevent kicking out other already exposed components?

Yes, this is what is what you can see in the picture; a small grey 3d printed plate, with a rather strong spring beneath it, which gives a linear pressure over the whole pick surface. The tape is pressed rather hard to the metal keeper, which is why I believe this might work well. 

 

ReferencePushPullFeeder uses some modulo math between Feed Count, Part
Pitch, Feed Pitch and Multiplier to decide whether a push-pull cycle is
actually needed.

 

Yes, this is exactly what it does (not Feed Count, but a similar counter). But I don't think they are all publicly available. 
I guess I could make the needed stuff public. I have to think a bit more about this, how it will relate to other peels/feeds, so the machine will not be bombarded with unnecessary peel requests.

I will need to make the script generic enough, so it will know if a peel has already been done from the use of another feeder feed or peel request etc.

I also am not sure how long time the peel take, and if it has enough time to finish after the Feeder.BeforeFeed/Feeder.AfterFeed, without delaying the pick cycle. So there are some things to consider.. :-)

I soon need to build some panels, and will test leaving the 4 pockets open to start with, collecting a knowledge on its quality performance.

 - Micael

[Jan]

Hi Micael!

On 20.10.2025 13:12, vespaman wrote:
> Hi Jan
>
> So you're trying to mimic the window of Yamaha feeders? Very
>
> Interesting, I did not know Yamaha did something similar! (Not that I
> thought I would be the first with this idea, I was pretty sure this had
> been used/discussed before).
>

You can watch their operation in the videos of thus guys, that design
controllers form them. They move the tape using a sprocket hole wheel
and an aperture in parallel. Then - when asked - they expose the part by
pulling back the aperture into its initial position. IIRC the cover tape
is removed when feeding and the part is held in its pocket by the
aperture. Thus feeders feed in about 100ms even tiniest parts.

> Didn't you previously tried to add some bed below the
> picking area to prevent kicking out other already exposed components?
>
>
> Yes, this is what is what you can see in the picture; a small grey 3d
> printed plate, with a rather strong spring beneath it, which gives a
> linear pressure over the whole pick surface. The tape is pressed rather
> hard to the metal keeper, which is why I believe this might work well.
>

If you haven't done already, you shall open up a store on some platform
line thingiverse and offer them to others like me...

> ReferencePushPullFeeder uses some modulo math between Feed Count, Part
> Pitch, Feed Pitch and Multiplier to decide whether a push-pull cycle is
> actually needed.
>
> Yes, this is exactly what it does (not Feed Count, but a similar
> counter). But I don't think they are all publicly available.
> I guess I could make the needed stuff public. I have to think a bit more
> about this, how it will relate to other peels/feeds, so the machine will
> not be bombarded with unnecessary peel requests.
> I will need to make the script generic enough, so it will know if a peel
> has already been done from the use of another feeder feed or peel
> request etc.
>
> I also am not sure how long time the peel take, and if it has enough
> time to finish after the Feeder.BeforeFeed/Feeder.AfterFeed, without
> delaying the pick cycle. So there are some things to consider.. :-)
>
> I soon need to build some panels, and will test leaving the 4 pockets
> open to start with, collecting a knowledge on its quality performance.
>

If you let your panel have 2^n boards, you shall not be left with some
exposed parts. And with Tobys new retry feature you could set this
feeders to defer and gracefully handle any parts that "vaporized" while
swapping boards.

Jan

[vespaman]

Hi Jan,

If you haven't done already, you shall open up a store on some platform
line thingiverse and offer them to others like me...

I don't mind sharing the prints, but I also don't think there's many people that might be interested.

You don't have a printer, right? If you do, I can add the missing files to my thingiverse.

Maybe I can send you some printouts if you want something in particular?

I don't remember how much I tested the feeder solution on the machine before I changed to closed loop drivers, I would be worried to pull 8mm without, since esp. without peeler, the acceleration of X (at least on my machine), really makes the feed pull the tape very fast/hard.

But it should be able to use a 0.4mm spring to mitigate that, if this is a concern. Or maybe I still have also a back plate will less pressure, that I use for the thickest paper tapes.

Bottom line - there's always a possibility that tweaks are need to perform well on different machine/setups.

If you let your panel have 2^n boards, you shall not be left with some
exposed parts. And with Tobys new retry feature you could set this
feeders to defer and gracefully handle any parts that "vaporized" while
swapping boards.

Unfortunately, since the peeler motor is common for all feeders, the two remaining parts will be exposed also by the next in line component peel.

And if the start is not with even components (0/2/4), the chance is that at a maximum, 3 components will be exposed. Otherwise 2.

But I am not sure this is really an issue, I suppose the staggered peel will be good, just as a extra precaution.

 

-  Micael

[Jan]

Hi Micael!

On 22.10.2025 16:53, vespaman wrote:
> Hi Jan,
>
> If you haven't done already, you shall open up a store on some platform
> line thingiverse and offer them to others like me...
>
>
> I don't mind sharing the prints, but I also don't think there's many
> people that might be interested.
> You don't have a printer, right? If you do, I can add the missing files
> to my thingiverse.
>
> Maybe I can send you some printouts if you want something in particular?
>

That would be very much appreciated. I'm in particular interest in your
guides to feed the empty tape through the base (had a view issues with
that) and the wheals for the cover tape guide (I also see that it tends
to stick on the rod). I do not have trouble with the springs or the tape
guiding block. Therefore I don't like to modify them (I'm thinking about
building adapters to feed 8mm tape thought 12mm feeders, but that's an
other story).

> If you let your panel have 2^n boards, you shall not be left with some
> exposed parts. And with Tobys new retry feature you could set this
> feeders to defer and gracefully handle any parts that "vaporized" while
> swapping boards.
>
>
> Unfortunately, since the peeler motor is common for all feeders, the two
> remaining parts will be exposed also by the next in line component peel.
> And if the start is not with even components (0/2/4), the chance is that
> at a maximum, 3 components will be exposed. Otherwise 2.
> But I am not sure this is really an issue, I suppose the staggered peel
> will be good, just as a extra precaution.
>

You're right. I did not noticed that earlier. However, with your nozzle
tip doubling techniques, there shall be a high probability to take the
second part from the same feeder.
Finally the question is how much a double feed like you're trying will
improve the speed. For 0402 you have one feed sequence per two parts and
one (longer) feed per four. So the feed itself only takes a little time
compared to the entire pick/place cycle and the double feed takes
something like less then two single feeds but for sure more then one.

Jan

[vespaman]

Hi Jan,

That would be very much appreciated. I'm in particular interest in your
guides to feed the empty tape through the base (had a view issues with
that) and the wheals for the cover tape guide (I also see that it tends
to stick on the rod). I do not have trouble with the springs or the tape
guiding block. Therefore I don't like to modify them (I'm thinking about
building adapters to feed 8mm tape thought 12mm feeders, but that's an
other story).

No problem. But just to be clear, the yellow arrows in this picture;

 

Similar colours to the ones I have ok (some light grey, that is)?

(Don't remember exactly which filament I used for my print) 

Finally the question is how much a double feed like you're trying will
improve the speed. For 0402 you have one feed sequence per two parts and
one (longer) feed per four. So the feed itself only takes a little time
compared to the entire pick/place cycle and the double feed takes
something like less then two single feeds but for sure more then one.

In fact, if you look at the graphs in my camera thread, you'll see that the whole pick+feed operation is the most consuming part of the whole sequence. Of course this is a matter on how you group them, but never the less.

The long feed gives about 250-300 placements per hour extra, over the normal feed, on my machine

 

 - Micael