Full automatic 3D printable smt feeder

[Bálint Szita]

Hello,

I made a full automatic smt feeder with SG-90 Servo motor.

All documentation and STL file downloadable from github.

https://github.com/szitabalint/SMT_FEEDER_SERVO

Balint

[Bálint Szita]

Video link:

https://www.youtube.com/watch?v=bTL0QI4lp5s

[Mike Harrison]

On Fri, 27 Jan 2017 13:36:31 -0800 (PST), you wrote:

>Video link:
>
>https://www.youtube.com/watch?v=bTL0QI4lp5s

Looks too jerky for plastic tape - parts are likely to bounce out

[Jason von Nieda]

Nice work Bálint! The design is very minimal, which I like. With regards to Mike's comment on jerk, I suspect this could be improved just by controlling the speed at which the servo actuates if needed. I have a couple questions:

* Is there a BOM? 

* The bearing a one way bearing?

* Have you tried it with plastic tapes?

In any case, congratulations on your design! It looks nice and from the video it certainly appears to work well!

Jason

--
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 post to this group, send email to ope...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/nifn8cpopudqkcic7q9to0btte797deu1j%404ax.com.
For more options, visit https://groups.google.com/d/optout.

[SMdude]

Perhaps a little bar that goes across the top of the tape where the tape exits the feeder would help a little too. This way when the cover tape is peeled, the tape cannot lift in the pick pocket causing tiny components to jump.

I love it!
I think the day is very near when I acquire a 3d printer. Pity Christmas and my birthday have been! :D
Thank you for sharing!

Cheers

[Anthony Webb]

This is really cool Balint!  I would love to hear more about what challenges you faced, what things you are happy about, and what pieces still need improvement?  I'm still a little fuzzy on how it all works, any more detail you can share would be great!  Congrats!

--

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+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/39680fbf-1baa-496e-bdf3-3160a6111d8a%40googlegroups.com.

[Bálint Szita]

[Bálint Szita]

Hello Jason

Yes it's works with plastic tapes up to 3mm height.
I'm tested it, with SOT-23 transistor and 10U/0805 capacitor.
The feeder pulley has two 3x10x4mm bearing.

BOM:

2x -  3x10x4mm bearing
1x - control PCB with PIC12F510
8x - M2x 12mm screw
7x - M2x 5mm screw
5x - silicone rubber (form child's play)
1x - SG-90 servo

Balint

[Anthony Webb]

Any plans to support embossed tape?

--

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+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/2bc5c995-662d-4373-9954-2ec0c6c2d286%40googlegroups.com.

[Bálint Szita]

Hello

Perhaps this video shows that it works

https://www.youtube.com/watch?v=-htC3pBVXA0

Problems:

The 3D printed reverse rotation stopper isn't precise.
Sometimes loosen the plastic tape.

Balint

[Bálint Szita]

Yes it's works with embossed tape.

[Bálint Szita]

Test Video:

https://www.youtube.com/watch?v=G3Xg9xvcwfc

[Oz-Ron]

Balint,

 

I am very impressed by your clever design.  WOW, the sequencing is perfect.  Excellent work!

 

The next challenge is to see if the design can be made a little skinnier so more lanes can fit in the available pick space.

There are a myriad of small servos available:

https://hobbyking.com/en_us/servos/sub-micro-0-5g.html?dir=asc&order=price

Maybe some of these might work?

 

Anyway, going to knock some up and have some fun trying out your design.  Thanks heaps for sharing the source!

 

(FYI, the smt_feeder_one_servo_box_a.stl file generates a “Non Manifold” error, all the others are good.)

 

Mike,

The part can’t bounce out because both the main tape and cover tape advance together, then the cover tape is pulled back with the part stationary.  That’s the clever bit.  (Have to watch the video frame by frame because it is so quick.)  I agree with Jason though, it could be slowed down a touch from light speed. J

 

Cheers,

Ron

[Mike Harrison]

On Sat, 28 Jan 2017 01:42:42 -0800 (PST), you wrote:

>Test Video:
>
>https://www.youtube.com/watch?v=G3Xg9xvcwfc

Try it with small, light parts like SOT-23 or SOT-323. I think you will see parts bouncing out.
It may be better with a slower feed - the cover tape on some plastic tapes can sometimes jerk as it
comes off instead of peeling smoothly. It can also be a bit sensitive to peel tension.

I didn't see the earlier part of the thread - how are you taking up the cover tape?

[Bálint Szita]

Upload file to github:

smt_feeder_one_servo_box_av2.stl

[Michael Anton]

You should be able to use a spring loaded arm riding on the same gear on the cover tape pickup spool, except on the bottom side.  This should stop it from spinning backwards when you try to advance it.

The one problem I see, is that as you get lots of cover tape on the spool, it will pull move cover tape when advanced than when the spool is empty.  There needs to be some sort of slip mechanism, otherwise, I suspect that the tape will advance due to too much cover tape being pulled.

[Bálint Szita]

Not the servo is determined by the width dimensions. The plastic tape pulley determine overall width.

[Michael Anton]

On Saturday, January 28, 2017 at 3:29:10 AM UTC-7, Michael Anton wrote:

The one problem I see, is that as you get lots of cover tape on the spool, it will pull move cover tape when advanced than when the spool is empty.  There needs to be some sort of slip mechanism, otherwise, I suspect that the tape will advance due to too much cover tape being pulled.

Never mind, it looks like you have that handled with the slotted hole, and only allowing the cover tape to be pulled by the spring.

[Bálint Szita]

The plastic tape pulled only spring. The servo motor can only pull up the spring.

[Michael Anton]

You can share the stop gear and pulling gear, by putting the stop arm at the bottom on the same side as the pulling arm.  This would save you 2mm.

[Bálint Szita]

Pulley arm and spring

[Bálint Szita]

This is a good idea.

Balint

 

[Oz-Ron]

Michael,

Great suggestion.  If that is the pinch point for a 2mm width penalty, plus it simplifies the printing process, bonus!

 

Maybe the sides could be thinned a little just in critical pinch areas too.

 

Balint,

before you know it, sub 13mm wide if you are happy to try some suggestions that might help.

 

Thanks for updating the STL file so quickly.

Please keep up the good work!

 

Cheers,

Ron

[Bálint Szita]

I didn't pull off the cover tape, because I didn't want to loss the parts.

[Bálint Szita]

Test SOT23-5 Video:

https://www.youtube.com/watch?v=SH2uKZ0hRkc

[Deniz Ertan]

Amazing design. Thanks for your fast sharing policy Bálint :]

[Jason von Nieda]

Absolutely brilliant work! Thank you so much for sharing. I intend to print one out today!

Balint, may I place a link to your feeder on the OpenPnP hardware page at http://openpnp.org/hardware?

Jason

On Sat, Jan 28, 2017 at 10:58 AM Deniz Ertan <dene...@gmail.com> wrote:

Amazing design. Thanks for your fast sharing policy Bálint :]

--

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 post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/78cd03a7-af3a-42b3-96e9-38d6a42f6646%40googlegroups.com.

[Jason von Nieda]

Hi Balint,

I printed one of your feeders out yesterday and have some questions:

* The teeth on the gears are quite small, and it's easy for the pawl to slip out of them. Is there any reason not to make them a little bigger? This might just be a quirk of my printer.

* I'm not quite sure what "silicone rubber" is. Are they just silicone rubber bands? Like this? https://www.amazon.com/gp/product/B0041HTBQ2/ref=crt_ewc_img_gw_2?ie=UTF8&psc=1&smid=A3IDQG8WRUU0VM

* What material are you using for the flat spring that holds down the tape? It looks like it might be copper clad PCB? What thickness?

Thanks,

Jason

[Bálint Szita]

Hello,

I'm using rubber from kid's game.
Example:
https://www.amazon.com/Refill-Loom-Bands-Assorted-Neon/dp/B00EAXKSF0/ref=sr_1_10?ie=UTF8&qid=1485761868&sr=8-10&keywords=Rainbow+Loom+Bands
I'm printed gear with slow speed 30mm/sec. 100% infill and 0,2mm layer hight, material: PLA.  (Printer: prusa i3)

Balint

[Jason von Nieda]

Thanks Balint! Those rubber bands are much cheaper than what I was thinking! :)

I will reprint the gears with higher quality settings. 

Would it be okay if I place a link to your feeder on http://openpnp.org/hardware? I think this will help people find it and get more people interested.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/cfde70f2-4a49-4aef-8f58-eb72e6d0832f%40googlegroups.com.

[dzach]

Nice design.
Looks like the strip tape holes are on the opposite side from where OpenPnP expects them.
I believe this can be corrected by mirroring the design (or the STL files).

[Jason von Nieda]

Hi dzach,

Since this is a full auto feeder, OpenPnP doesn't care where the holes are. No vision is performed on a full auto feeder since it is expected that the pick position never changes.

Jason

--

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 post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/485c7ca0-44b0-4631-a33e-13afae5cf3bd%40googlegroups.com.

[Thomas Langås]

On Tue, Jan 31, 2017 at 4:10 AM, Jason von Nieda <ja...@vonnieda.org> wrote:
> Since this is a full auto feeder, OpenPnP doesn't care where the holes are.
> No vision is performed on a full auto feeder since it is expected that the
> pick position never changes.

What happens when the position actually changes? Like for instance with
DigiKey digireels?


--
Thomas

[Jason von Nieda]

Hi Thomas,

I've never worked with DigiReels. Can you explain to me how it works, and what causes a change of position?

Jason

--
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 post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CANN7TQ%3Dj3v1yOppzoeP6VR%2BqGQkwx8nO9Wts72%2Bdo5WNM__oAA%40mail.gmail.com.

[Thomas Langås]

Digireels are basically cut-tape + empty cut-tape put together by tape, on a reel.  More often than not, they really don't match and you need to detect change or re-adjust manually. This is mostly done with vision on live system I've seen...

On Jan 31, 2017 04:17, "Jason von Nieda" <ja...@vonnieda.org> wrote:

Hi Thomas,

I've never worked with DigiReels. Can you explain to me how it works, and what causes a change of position?

Jason

On Mon, Jan 30, 2017 at 9:15 PM Thomas Langås <thomas....@gmail.com> wrote:

On Tue, Jan 31, 2017 at 4:10 AM, Jason von Nieda <ja...@vonnieda.org> wrote:
> Since this is a full auto feeder, OpenPnP doesn't care where the holes are.
> No vision is performed on a full auto feeder since it is expected that the
> pick position never changes.

What happens when the position actually changes?  Like for instance with
DigiKey digireels?


--
Thomas

--
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+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CANN7TQ%3Dj3v1yOppzoeP6VR%2BqGQkwx8nO9Wts72%2Bdo5WNM__oAA%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

--

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+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jwRatNJZG9jwgk2pH-kx0vE%3D89O5u7BT%2BDJ-SWVJ64ZcA%40mail.gmail.com.

[Jason von Nieda]

Thomas,

Are there multiple non continuous pieces of tape holding parts? I was under the impression it was just cut tape +  leader + trailer. This would mean that once you got the first part fed the rest should be the same. If that's not the case can you provide further detail? 

I would think that if it didn't work like this then DigiReels would be useless on most PnPs as they don't have feeder vision on auto feeders, either?

Jason

On Mon, Jan 30, 2017 at 9:21 PM Thomas Langås <thomas....@gmail.com> wrote:

Digireels are basically cut-tape + empty cut-tape put together by tape, on a reel.  More often than not, they really don't match and you need to detect change or re-adjust manually. This is mostly done with vision on live system I've seen...

On Jan 31, 2017 04:17, "Jason von Nieda" <ja...@vonnieda.org> wrote:

Hi Thomas,

I've never worked with DigiReels. Can you explain to me how it works, and what causes a change of position?

Jason

On Mon, Jan 30, 2017 at 9:15 PM Thomas Langås <thomas....@gmail.com> wrote:

On Tue, Jan 31, 2017 at 4:10 AM, Jason von Nieda <ja...@vonnieda.org> wrote:
> Since this is a full auto feeder, OpenPnP doesn't care where the holes are.
> No vision is performed on a full auto feeder since it is expected that the
> pick position never changes.

What happens when the position actually changes?  Like for instance with
DigiKey digireels?


--
Thomas

--
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 post to this group, send email to ope...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CANN7TQ%3Dj3v1yOppzoeP6VR%2BqGQkwx8nO9Wts72%2Bdo5WNM__oAA%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

--
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 post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jwRatNJZG9jwgk2pH-kx0vE%3D89O5u7BT%2BDJ-SWVJ64ZcA%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

--

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 post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CANN7TQ%3D-wTaLHOO0BVmvxxKV6_sYSNi3n-y%2B%3Doi1aXrpAPgUZg%40mail.gmail.com.

[Michael Anton]

Any of the DigiReels that I've used fed into a standard auto feeder just fine.  The holes should line up well enough for that, or at least they did for the sprocket driven feeders that I use.  Perhaps an auto feeder such as this one with a pin would have some drift.

[trentonc...@gmail.com]

Jason,

I had this question as well, but less on account of the differences in tape spacing between vendors, and more to compensate for any drift/errors in feeding due to mechanical issues (especially considering some of us are experimenting with 3d printed feeders).

I actually am working from scratch to learn open cv for java to implement this feature in open pnp and am getting there slowly.  In my mind, the software would visually identify part location every feed advance, or every number of feed advances to correct for mechanical and other sources of error/drift.

On some level, i suppose a stepper or servo linear feed could correct for this by calibrating the feed rate/step, but that is an expensive option for what i'm trying to set up.

regards,
t

[trentonc...@gmail.com]

Balint,

Great build, and thanks for posting the files.  I had some questions/critiques (that i might play around with myself)

-pictured on the github repository here what is the pcb for?  is this for driving the stepper?
-I notice there are a lot of tiny hardware nuts and bolts required to assemble this, as i scrounge the scrap bin waiting for my order of parts to arrive so i can test this out, I was wondering if there is any way to avoid needing them by including features in the print such as hooks or latches?
-you may be able to print in the spring feature, while on the other hand, it may also be nice to have a way to attach the cover spring such that you can experiment with different spring material/length/pressures.

thanks

On Friday, January 27, 2017 at 1:34:11 PM UTC-8, Bálint Szita wrote:

Hello,

I made a full automatic smt feeder with SG-90 Servo motor.

All documentation and STL file downloadable from github.

https://github.com/szitabalint/SMT_FEEDER_SERVO

Balint

[Jason von Nieda]

Hi Trenton,

Glad to hear you are learning OpenCV. We can use more help here!

There is useful code in both ReferenceDragFeeder and ReferenceStripFeeder. I'll describe the CV capabilities of each a bit:

ReferenceDragFeeder: One of the first pieces of CV I wrote. It uses a template match to find the part in the pocket and then determines an offset to use for when it drops a pin in the hole to pull the tape forward. It works, but it's not great. It suffers from a number of problems, the main one being that templates don't work great when parts are rotated in the tape.

ReferenceStripFeeder: This is a much more modern attempt at a similar thing. It uses hole finding to find the holes in the tape and then uses the relationship between the holes and the tape to find the part. It's not perfect, but overall it works very well. It's on the list to be converted to the CvPipeline system which will make it much more flexible.

For what you are describing, I would suggest looking at ReferenceStripFeeder. Finding the holes is much more accurate and reliable than finding the parts. The holes are always the same. The parts can differ wildly. 

If you haven't already, check out the CvPipeline stuff in OpenPnP: https://github.com/openpnp/openpnp/wiki/CvPipeline

This is an attempt to make it easy to quickly customize complex computer vision pipelines. By using a pipeline and some common UI themes it makes it possible for users to easily customize CV operations for their specific setups. All OpenPnP CV operations are using this system going forward, and several already use it. Bottom vision, in particular, was designed around CvPipeline.

Good luck with your learning, and please let me know if I can assist! The IRC channel is a great place to chat about this kind of stuff in real time: http://webchat.freenode.net/?channels=openpnp. I'm on there as "vonnieda" almost 24/7 :)

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/5407e9b1-7554-4e95-aba1-c27bb289478e%40googlegroups.com.

[trentonc...@gmail.com]

Jason,

thanks for the info and support.  I will certainly ask around the community as i get lost along the cv pipeline.  I realize the error correction will be difficult, however the initial goal that set me out on learning it was identifying blobs boxes and edges so that i can verify that there is/not a part in a location.

Wish me luck

regards,
T

[Cri S]

You can aquire color image, and suppose you use ROI do gaussian blur, normalize, threshold
and findcontours without any gray conversion.
This gives you position of components.
Its possible to store that info.
Now go above first component for lightning reasons, set stronger roi and take image converting the blurred image to gray, normalize, gaussian blur and threshold to 220 + findcontours. Optional
Merge contours that are less distant as 20% of component with or 1.5mm, whatever is less.
This gives you the exact position and angle of component.
This is how I do top vision.

[Bálint Szita]

Hello,

Soon I make a video how to assemble.

Balint

[Bdale Garbee]

Hi Balint!  I really like your design, but would like to tweak a bit for use with my machine.  I see that in addition to the stl files that you also published an stp of the assembly, but that's not a great format for starting modifications from.  What software did you use to do the initial design, and are you willing to share more of the original design files?

In addition to making the whole thing a bit thinner as others have already suggested by combining both ratchet mechanisms on the take-up reel on one side, I'd like to adjust the mounting to work like Alex's design that sits down over a 20x40mm extrusion, and I'd like to eliminate the circuit board pocket since I've recently designed a USB to 16-channel servo board that I think would be ideal for controlling a bank of 16 of these... the servos I have in hand seem to have cables long enough to allow them to reach my new servo board mounted centrally under the mounting rail in the middle of a bank of 16 feeders.

I also need other tape widths besides 8mm, but that is of course pretty easy once the basic design stabilizes.

Anyway, I have some time and some enthusiasm and would love to help make this design even better!

Bdale

[Bálint Szita]

Hi Bdale!

Used cad is Pro/Engineer wildfire 3.  I work in a wider range  feeders (12mm/16mm/24mm). I hope I'll be ready in a few weeks.
For the smt feeders control I make 10 chanel RS-485 control board (Atmega48PA...).

[Bdale Garbee]

Thanks for the quick reply, Balint.  I don't have access to Pro/E, and I'm not sure that leaves us with a common file format that's easily modifiable.  Wish the CAD tools world were a bit less proprietary in that way...  The tape widths you mention are the ones I need, which is great.

Here's a quick screen dump of the 16-channel USB-attached servo control board I whipped up.  I should have boards to load in another week or so.  It uses an STM32F042 (ARM Cortex) SOC driving a PCA9685.  I'll post more about it when I have it running.  The combination of USB attach to the host running OpenPNP and simple 3-wire servo cable attach to each feeder seems like a good combination to me, though...

[pham thang]

Deare balint

How you trigger the feeder work.

Thang

Vào 04:36:32 UTC+7 Thứ Bảy, ngày 28 tháng 1 năm 2017, Bálint Szita đã viết:

Video link:

https://www.youtube.com/watch?v=bTL0QI4lp5s

[Michael G.]

Hello Bálint,

thank you for sharing your design, really cool! I think about to use your feeder on my machine I plan to build.
Did you make any progress or do you have improved the feeder in any kind?
I'd like to adapt your currentmost model to my needs and be glad if you would share your wildfire-files.
Further I think about to control 12 feeders with a simple g-code parser on an arduino uno in combination with the servo-library. Think that would work similar as your solution? It could be controlled directly by the gcode-driver in openpnp.

[Michael G.]

Hello again to all,

i'm fiddling around with the design I imported in solidworks (import doesn't work very well, but it's ok for now).
I thought about to modify the design, primarily to
*enable feeding strips with pockets (eg. sot23) and
*have the strip top-aligned for having them on the same z-level throughout the whole bank regardless of pockets, plastic or paper-strips.

Before I redesign it I would like to receive some feedback of people with more experience in feeder design as me.
I made two screenshots of the principle I intend to go for. Modified parts are colored in the transparent model. Maybe a spring from below doesn't support the stripe enough? One could integrate a few more springs near the pick-place and support it in grooves within the feeder, I think.
Further there should be an subtractive extrusion for the pockets as indicated in the second image. To have this, the whole feeder would need become higher. It would follow the principle realized here: https://www.youmagine.com/designs/smd-tape-strip-holder-8mm-12mm-16mm-24mm

Do I miss some drawbacks I introduce by changing the design this way?

michael.

[Anthony Webb]

A couple thoughts, if you are going to go with support from the bottom it does need to be nearer to where the pick actually happens.

One more thing to consider as this seems to be one of the trickier components, the balance of everything tends to be very difficult.  Meaning, you have to secure the tape by some means, this creates friction that needs to be overcome by the force you are advancing the tape with.  Too much force and things are jerky and parts jump around on you.  Peter and I have been playing around with using a motor to advance the tape and pick up the cover which lets us dial in the forces and has been working relatively well.

--

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+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/1bb6c80d-56f4-418d-b874-4464c885f5f9%40googlegroups.com.

[Trampas Stern]

On idea which I had is to make the part "shoe" changeable.  That is print a custom shoe for each part. Then the depth of the part could be set with no spring needed from the underside as it always matches the part height. 

Another idea I have seen is people putting foam under the parts, the foam compresses and provides some upward force. 

To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

[Michael G.]

You mean this one: https://groups.google.com/d/msg/openpnp/Kg57XAz-_IA/jFsMG4N5BwAJ ?
I did not find it in the group before. Do you have some more information, videos or something else how it works? You are using that feeder currently on your machine?

Began to remodel the cad-model. It's not yet fitting everything together for now, but just to keep the idea about it.
The spring to push the tape upwards could be printed off PETG. Its flexible and has smooth surfaces. Maybe some of the rubbers can be replaced by such type of springs, too.
For a higher force, the rotating arm of the servo could be shortened a bit since it's currently rotating only about 90° for advancing the tape. In combination with an arduino I plan to use, one could drive a ramp to smoothen the advancing of the tape?



I'm currently waiting for the servo-motor to test the feeder-design and see what I can do to improve it. Would like a feeder which is fully 3d printable.

To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

[Michael G.]

forgot the pics :)

hmm, maybe if using specific shoes or just grooves in the sidewall, there is no other kind of support from the bottom necessary?

[Trampas Stern]

So the best thing is to start printing designs and testing. 

You can improve the tape take up ratcheting system, see here:

https://groups.google.com/forum/#!topic/openpnp/ywNMvo6WBr8

Note making a working feeder design is simple, making a feeder design that is 3D printable across various quality of printers is hard.  Making an automatic feeder that is cheaper than a $50 Yamaha feeder is really hard. 

That is an automatic feed 14mm in width for an 8mm tape is easy, getting the cost down to under $50 is not too hard, getting it down to under $20 is really hard. So the big question is if you could go out and buy a proven working automatic feeder today what would you be willing to pay for 10 of them? 

I had started doing a design with an SG90 servo: https://groups.google.com/forum/#!searchin/openpnp/trampas$20feeder%7Csort:relevance/openpnp/bJOfE4xI1Pg/-fAtPOKGCAAJ

The issue here is that teeth for the tape should  to be made of metal for long life, the gears need to be better quality than a FDM 3D printer, etc.  So the question is not about can we design an automatic feeder, the question is how much are we willing to spend for feeders. 

For myself it is hard to compete with $50 Yamaha feeders. I would love to have a $20 or $30 automatic feeder, I would rather pay the money than print one myself, however I might not be like most people. 

 

[Michael G.]

Good evening,

just printed a part. Looks good so far. The tape is well supported but easy to move manually.
I will consider your concept for the ratchet system later. Will work the next days on the construction, but I have no motor yet to try advancing automatically.
Next construction site is setup an controller. I made a fast draft for a shield (arduino uno/mega). Will keep posting an update when I've gone further and don't hijack this thread any more ;)

kind regards,
michael

[Anthony Webb]

Michael, I'm really keen on seeing how this works for you.  Do you have an STL somewhere for the motor driven version?

--
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+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/696359b6-d584-456f-8b4d-b407ff367f79%40googlegroups.com.

[Michael G.]

Hi Anthony.
Do you mean the original files from Bálint? You can find them on github.
 
I will give my parts/drawings away when I stabilized the reworked design. If you want them for a review, just write a mail, I could export an stp-file for you - feedback is welcome.
Went to the city today to get a little servo and wrote a first prototype program for the arduino (see vid). Advancing the tape works but not 100% reliable yet (haha, indeed I have the dim feeling I will need a vacuum sensor one day ;). Have to adjust the springs etc. to improve advancing. Since paper tapes are much thicker than plastic tapes I think there has to be a dedicated version for both of them. After stabilizing the advancing I will look into the tape peeling mechanism. Current design you can see in the picture.

[Michael G.]

hi @all.

much time went by since april and many kilos printed plastics later, I think I have found a 3d printed full automatic feeder that hopefully works not only for me well.
I put much effort into building the feeder as well as having setup a controller and documenting everything properly. If you are interested, I set up a small wiki with documents and downloads: https://wiki-public.mgrl.mywire.org/maschine:pickandplace:feeder:automaticfeeders
please click through the site and see yourself.

beware that the site and the feeder are still work in progress: currently I'm building a series of 8mm tape feeder. furthermore I am waiting for the n20 motor-pcbs to be delivered as well as the arduino-shield pcb.

[Anthony Webb]

This is awesome!  Really looking forward to giving it a try!

--
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+unsubscribe@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/f3ff77fe-1e8f-4492-a882-65173cce4c1a%40googlegroups.com.

[Anthony Webb]

@michael. Nice work!  I have my plastic parts printed and the seem to be very well thought out and a great fit!

Took me  exactly 8 hours to print them at normal resolution. Will need to source some parts to test it out but so far so good!

Sent from my iPhone

[Anthony Webb]

Got the feeder assembled and feeding parts. The design is a work of art and has been tweaked and fine tuned by Michael to be a reliable and accurate solution to economically feeding parts.  Still some enhancements on the way but this is as good as I have seen and highly recommended. We are all wired in to openpnp with gcode driver and feeding parts flawlessly.  More details to follow. https://youtu.be/vJzb3llKgjA

Sent from my iPhone

On Nov 11, 2017, at 10:52 PM, Anthony Webb <anthon...@gmail.com> wrote:

@michael. Nice work!  I have my plastic parts printed and the seem to be very well thought out and a great fit!

Took me  exactly 8 hours to print them at normal resolution. Will need to source some parts to test it out but so far so good!

[Bradley Parcels]

Looks amazing Michael - nice work!

[Jason von Nieda]

This slipped past me when you posted it a month ago - but now I'm finally looking at it. This is really amazing work. The documentation is completely off the charts good. Great work on this Michael!

Would you mind if I added a link to the OpenPnP Hardware page at http://openpnp.org/hardware/? If it's okay, let me know how you'd like the link to appear.

Looking forward to building one, or ten, of these! :)

Jason

--
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 post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/b6a608eb-7294-477f-ba93-04fc7be9dc21%40googlegroups.com.

[Michael G.]

Hi jason, thanks for your kind words. I built already 12 of them. they work pretty well for 0.6mm paper tapes already but didn't work for thicker tapes (e.g. 1mm for 0603 caps) and still have problems with plastic tapes.
currently working on an optimized version and finalizing the design. when this is finished (or at least I can't make it any better), I update documentation for latest design and start a new thread to introduce the feeder "officially". afterwards I would be proud to get a link on the site :)

[Mike Harrison]

On Thu, 14 Dec 2017 00:41:04 -0800 (PST), you wrote:

>Hi jason, thanks for your kind words. I built already 12 of them. they work
>pretty well for 0.6mm paper tapes already but didn't work for thicker tapes
>(e.g. 1mm for 0603 caps) and still have problems with plastic tapes.
>currently working on an optimized version and finalizing the design. when
>this is finished (or at least I can't make it any better), I update
>documentation for latest design and start a new thread to introduce the
>feeder "officially". afterwards I would be proud to get a link on the site
>:)

Have you done any endurance testing on them to see what fails first ?

[Michael G.]

no I have not done any endurance-testing yet.

just updated the page with the latest version and created a new topic: https://groups.google.com/forum/#!topic/openpnp/0ZMk60x51tw
I think I have a stable design now and am happy to share with others...