Loading

Open Pick-N-Place Tape Feeder

by Erik, published

Open Pick-N-Place Tape Feeder by Erik Mar 24, 2011

Featured Thing!

Description

The files have been moved to YouMagine, a really nice place to share your 3D designs! They can be found here:
----------
youmagine.com/designs/open-pick-n-place-tape-feeder
----------
This is part of my efforts to automate assembly of PCBs, using an existing 3D printer with a pick-and-place toolhead instead of acquiring a 40K euro machine. This is a parametric design that can be easily adjusted for various sizes of SMD tape. The gear will also be adjusted depending on the pitch of the tape and components.

A tape feeder like this can cost about 1 euro instead of 200 euro's.

Like most of my design, it was designed in OpenSCAD, openSCAD.org a wonderful tool, especially if you're more of a TUI person than a GUI person.

Personally, I really like the recursion that you can print and cut these parts on open source machines and that they will allow you to automatically assemble the PCBs for 3D printers and laser cutters. Also, the laser cutter can be open source, which is what our friends in Utrecht are working on: see laoslaser.org/

Video: youtube.com/watch?v=fsWpC6L91qo

Derivatives, additions, improvements and suggestions may also be prized as part of the Ultimaker challenge:
Ultimaker.com/Challenge !

Recent Comments

view all
very cool model. One suggestion though, if you want to keep cost to a minimum, then surely using a spring+solenoid to drive this would be cheaper (1 transistor to control). Also if you had a takeup wheel for the excess covering film you could ratchet that off of the same mechanism, so cost would only be an extra plastic part, and a maybe cf spring if you need some compliance.
THANKS!!!!!!!!
- buffering the membrane pump somehow. Any ideas?

Use a membrane pump which blows air OUT, put it in a sealed plastic box, and you'll get a "large amount of under-pressure" which should draw air in more smoothly.

Haven't tried it, tho...

Makes

Liked By

view all

License

GNU - GPL
Open Pick-N-Place Tape Feeder by Erik is licensed under the GNU - GPL license.

Give a Shout Out

If you print this Thing and display it in public proudly give attribution by printing and displaying this tag. Print Thing Tag

Instructions

0. See the video for more instructions.
1. Adjust the parts in openSCAD to fit your tape. Export them to STLs.
2. Print the STL files.
3. Assemble them and ensure that it slides well. The gear, lever and the tape holder have a tight M3 bore.
4. It's better to print it a little too narrow and use a little file to get the exact tolerance you need.
5. Laser cut the gear from POM Delrin (tm).
6. Write control and computer vision software :)
very cool model. One suggestion though, if you want to keep cost to a minimum, then surely using a spring+solenoid to drive this would be cheaper (1 transistor to control). Also if you had a takeup wheel for the excess covering film you could ratchet that off of the same mechanism, so cost would only be an extra plastic part, and a maybe cf spring if you need some compliance.
Could you measure the minimum and maximum angles required for dispensing a new part and upload the animation as a scad file I'm writing some code. I would like to participate in this project more. :)
In principle I would like to support any degree, it wouldn't be too difficult, we just need a tip controlled by a servo or stepper motor. Since most RepRap boards have a 4 or 5 axis stepper board anyway, this shouldn't be a problem. Only the relative angle is important, we don't need to home the suction tip as it's round.

What kind of animation do you want? This is just one part of the puzzle, there are some picking tools designed already, though not many of there are printable designs. I haven't made a picking tool design, but I have a non-rotating vacuum tip at the moment that already works for simple parts.
Thank you for the detailed information! I was actually curious about using micro servos to individually control the dispensing of each part. The degrees that I am curious about are the rotational degrees on the tape dispenser required for dispensing a new part. The animation that was shown in the video would be great. I think that this open pick and place project could turn into maybe (dare I say) its own wiki! I have written some arduino code that could control 3 servos via the serial port. (I can easily adapt it for up to 16 servos.) I also have a schematic (done in fritzing). My concept involves that the tape dispenser bank uses its own controller board (atmel's 328 chipset with arduino bootloader) to control the servos via serial commands from the motherboard. I would like the rotational degrees for programing the micro servos. I have Cupcake CNC but alas no laser cutter. So I have no way to test this or build it unless I send off for some laser cut parts from shapeways or ponko. :)
How about this concept for the head??? The dxf files could be extruded in OpenSCAD so that the head is partially printable. The extra stepper drivers would come in handy here.

thingiverse.com/thing:4070
Yep, that's one of those designs that got me inspired. For non-360 degree rotation an off the shelf non-continuous servo seems it should work great. I think we can get away with the limitations this imposes (which are few).

You're right it deserves its own wiki. I started collecting information on the RepRap wiki already, which is a great resource to put this in:
http://www.reprap.org/wiki/SMT_Pick-n-Place_System

The current feeder was designed so that the lever would be pulled by the toolhead (not the vacuum grippe
r, but some other strong part designed for it and hooked up to the moving head.
That page looks great! Should I upload some code??? And maybe design a pcb for the controller?

I thought about the idea of using the toolhead to move the lever, but I would think that using servos would make the machine a bit faster at placing components. These are the servos I am thinking of using. solarbotics.com/products/25500/ It shouldn't take more than 1.6 kg to operate t
he dispenser. Would you like to continue this conversation over email?
I like the idea of using servo's, but the lever pulling doesn't need to take any time. It can be part of the normal action where the toolhead moves out of the pickup site. The toolhead can produce quite some force (at least on my setup). On the other hand I do like the use of servo's, given that you decouple two problems and can make them reliable independently of each other.

Please do add your code! I'd love to start testing.

I think it's best to continue the conversation online. This page is good for discussion, the RepRap wiki is a good place for collecting ideas that are developed (slightly) further.
Ok I added my code to the page I put it under the tape feeder section. I also created a google code project page for firmware and code hosting. I would like to add you add as owner of the page. If you wouldn't to PMing me your gmail I can add you as an owner. Here is the google code page. :) code.google.com/p/smt-pick-n-place-system/
I'm also still looking for an affordable way to detect a change in pressure, to find out whether a part was grabbed or not. Perhaps something capacitive or some other effect might also allow you to detect this, so that's an option too. Any comments are welcome.
A light sensor that gets obscured with the grabbing of a part would be another way to do it. This could also tell you when the part has been released or dropped prematurely.
Erik - in reply to Guest
I'm considering computer vision as a good way to detect if a part was grabbed. The parts might be smaller than 1x1mm... so there's not a whole lot to obscure. But if you have a small enough light source, this might still work. Perhaps we could even shine light through the vacuum nozzle... :)
This is very good...but the lower ratchet mechanism will be putting a lot of stress on the fragile wheel nubs...rubbing across the tops, then pushing on the side. I think maybe a second set of stronger nubs should be considered, or maybe design a wheel where the nubs can just be some pins or small brads pushed into the edge of the wheel.
Erik - in reply to
I was worried about this, but POM is very durable in general. I will see how it goes. I want them to last at least as long as a full reel, or 4000 components placed. I can test it with out the tape and run it 10000 cycles. The three holes were meant for aligning a second piece of POM that has a smoother profile. I will definitely pursue this if wear becomes a problem.

Other issues I'd love some ideas on:
I've actually placed a few components with an Ultimaker running hand-coded GCode and a using a vacuum membrane pump for the head (a syringe needle). The pump's flow, however, oscillates too much. It tends to push away parts sometimes, when they're not fully pi
cked up yet. I'm looking into the following options:
- venturi (aspirator): using an air pressure source and a T-junction, this seems to work.
- two mechanically coupled pistons, one that is driven by a flow, the other is pulled outward to create a vacuum (this is not continous, problematic)
- buffe
ring the membrane pump somehow. Any ideas?
- some other off-the-shelf pump that doesn't oscillate? Or one with multiple pumping blades.

I'm also deciding on how to achieve accurate positioning. I guess I could use the SMT tape itself as a centering pit. Similar to this:
http://www.ciciora.com/pickn
place.html
jcw - in reply to Erik
- buffering the membrane pump somehow. Any ideas?

Use a membrane pump which blows air OUT, put it in a sealed plastic box, and you'll get a "large amount of under-pressure" which should draw air in more smoothly.

Haven't tried it, tho...
could you use the piston that is pulled out to create a vacuum to also calculate the air pressure inside the chamber and detect whether or not it is still holding the part??
Erik - in reply to Guest
That's a great idea. I'm not using a piston at the moment, but using one, with a spring would probably work to detect a part. But it's mostly useful if you use the piston for pumping as well, as you're suggesting.

If I'm not using a piston for pumping, I guess another method to measure pressure might be better.
fenn - in reply to Erik
put a jar inline with the vacuum pump. it will act like a capacitor and smooth out the air pressure. alternatively, you could use a really long length of hose.
Erik - in reply to fenn
I might try this. A really long hose is also a good option, I'll give that a try first. Slightly longer didn't change it much, though.
There was a design for a spiral pump on here a few weeks back, basically two spirals that have their axes offset, but still able to mesh and turn. Printable or millable, the hard part is probably balancing and sealing the hollow turning axis.
Cool idea!

This could also be useful as a hand tool when hand-soldering smt parts... The hardest part for me is usually getting the parts off the tape without losing them =-O
Top