Close

Does this project spark your interest?

Become a member to follow this project and don't miss any updates

Retro Populator

an electronics pick-n-place retrofit for 3D printers

16 97 73
Enjoy this project?
Share on twitter   Share on Facebook

This project was created on 06/26/2014 and last updated a month ago.

Description
A jig and software for allowing a 3D printer to do electronics pick-n-place assembly. It populates boards, by retrofitting a 3D printer, hence Retro Populator.

Motivation: surface mount soldering is great, but the process of placing parts is tedious and requires great manual dexterity as well good vision. Doing a few boards by hand is practical, but doing more than about 20 is hell on earth. Yet it's not practical (read: affordable) to pay industry to do it until about 500 boards - the setup charges just kill you. So there is this huge chasm between what's doable for hobbyists by hand, and where industry can take over, and this volume is commonly desired in the maker community - a small run of 50 to 100 boards is in fact typical. Thus, the desire to build machines that hobbyists can own to do electronics pick-n-place. Our cost-saving idea is to retrofit similar machines that makers already own: 3D printers.

Team:
Rob Gilson
Adam Evenden
Charles Hartlaub
Eric Boyd
Details

Video explanation of RetroPopulator:

Version 1.0 : Essentially complete now.  CNC milled acrylic base plate, with milled tape holders.  Vacuum syringe for picking up parts, vacuum supplied by SMT rework station.  Milled nozzle mount which attaches to the side of the existing extruder.  Z-axis rezero jig to easily change re-zero height.  3D printed cam-lock parts to holds boards in place on 10mm peg-grid base.  Code parses a yml file for board & tape placement info, and then an eagle .brd file directly for part types and locations.  It generate a gcode file which goes from pick to place location for all of the parts which can be populated.  Support for multiple boards.  Demo of RetroPopulator in operation, with multiple components and multiple boards:

Connected Device

The RetroPopulator connects mechanically to a 3D printer, which in turns connects to a computer. For our software we use Tegh, a wifi 3D printer client/server application D1plo1d's been building. For now we're using it locally on a single laptop, but ultimately it will allow us to access the RetroPopulator from our laptops anywhere in the lab and extra ultimately, the world, which is awesome! The ability to quickly load the RetroPopulator jig into a 3D printer, combined with wifi accessibility via Tegh, makes the RetroPopulator perfect for a modern connected workspace like hacklab.to or Ultimate Workshop.

Future Plans

Version 2.0: currently in planning/design. Addition of tape advancement, including plastic cover peeling. Addition of a second nozzle which will do solder paste dispensing. Software will have 'position confirm' feature where it will move to the corner of each board, and the first component on each tape, so that you can verify key locations directly before running the entire job.

Version 3.0: part rotation. Numerous other things we'll learn about as we do V2.0, no doubt...

Licensing

We're fully open source! Our PopCAM and Tegh Daemon software are licensed under GPL3. The Tegh GUI is licensed under MIT. Our hardware designs are licensed under MIT as well. All text, images and videos CC-BY V4.0.

System Design Doc

Charles has done up a nice brain-map of the RetroPopulator, download full-size PDF, or see a small version here:

Components
  • 1 × 3D printer (e.g. ultimaker)
  • 1 × surface mount rework station with intake port (e.g. for fume extraction, we use it for vacuum)
  • 1 × jig base, milled, 10mm grid, with vertical board walls
  • 7 × tape holder jigs, grooved acrylic, mounted on the base
  • 4 × 3D printed board holders w/ pegs
  • 1 × vacuum nozzle, with tubing (medical supply)
  • 1 × nozzle mount bracket (screws unto side of 3D printer extruder)
  • 1 × z-axis homing adjustment jig
  • 1 × c-clamps, to hold the jig to the 3D printer bed

Project logs
  • All About Retro Populator

    2 months ago • 0 comments

    Here's a <2 minute video which explains our motivations and aspirations for the Retro Populator:

  • ​RetroPopulator is kicking it up to 11!

    2 months ago • 2 comments

    Today we're showcasing 12,000 mm / minute of sweet FOSS pick and place automation.

    We've also upped the complexity in this run by pick and placing components for a circular board. And we're placing 4 different components in 2 different rotations.

    And we're doing 2 boards at once!

    Why all of that at once? Because why not that's why!

    All I'm saying is this isn't last week's pick and place!

    Final Score: 9 / 12 Pick and Places successful.

  • How to use the RetroPopulator v1

    2 months ago • 0 comments

    Check it out! It's Eric in all his splendour showing off how we set up the RetroPopulator!

    Part 1/2Part 2/2

View all 5 project logs

Discussions

bobcousins42 wrote a month ago null point

"We're fully open source! "

It is great that you are committed to Open Source. However, the CC-BY-NC-SA V4.0 license is not an Open Source license.
Please see http://www.oshwa.org/2014/05/21/cc-oshw/ for a longer explanation.

Are you sure? [yes] / [no]

Eric Boyd wrote a month ago null point

Hey thanks for your reply. We talked it over, and we've decided on a new license for the hardware: MIT. Enjoy your freedom!

Are you sure? [yes] / [no]

Pixel Pirate wrote 2 months ago null point

How much would this cost, and how accurate would this be? I am in dire need of a PnP, because shaky hands do not a good engineer make.

Are you sure? [yes] / [no]

D1plo1d wrote 2 months ago null point

If we were to do a kickstarter today it would be looking around $200 plus the vacuum pump.

If you're down to mill the parts yourself and don't count your time I'm not as sure. Depends on how cheap you can get machine time mostly.

We're going to get documentation for the mechanical design online soon so you can see for yourself.

PS. If you do mill one yourself post pics and let us know how much it cost! I'm excited to see the first RetroPopulator made in the wild!

Are you sure? [yes] / [no]

Adam Fabio wrote 2 months ago 1 point

Thanks for entering The Hackaday Prize Eric! Don't worry about having competition in the Pick and Place realm - the more the merrier! There are aspects of both machines I really like. I'm loving your videos so far - Keep the updates coming!
Good luck, and keep the past (re)flowing!

Are you sure? [yes] / [no]

D1plo1d wrote 2 months ago null point

Thanks Adam! I just posted an instruction video on how to use the RetroPopulator. We're on tack to keep those videos flowing so check back soon!

Not so sure about reflowing though, I think that might be version 4 ;)

It's super-cool to see competition taking off in open source pick and placing. I see it as a kind of multiple points of failure / heterogenous engineering. One way or another the future of small production run electronics is about to get automated.

Are you sure? [yes] / [no]

D1plo1d wrote 2 months ago null point

/s/tack/track

Are you sure? [yes] / [no]

cBake wrote 2 months ago 1 point

Hey, a video of it actually working! Congrats. This is an awesome project. Can't wait to see more iterations :)

Are you sure? [yes] / [no]

D1plo1d wrote 2 months ago null point

Thanks cBake, I just posted 2 more videos showing how we set up the machine just for you*

* you and the rest of the internet. Enjoy ;)

Are you sure? [yes] / [no]

Stanos wrote 3 months ago null point

Looking great, but what about the ability to rotate?

Are you sure? [yes] / [no]

D1plo1d wrote 3 months ago 1 point

Yeah, that's definitely on the roadmap. Eric makes boards with all sorts of weird angle components for aesthetics so we're very aware of the rotation limitations of the current prototype.

The plan as it stands is to build up our rotation feature through a couple iterations.

So in the near future version 1 will be able to do rotation in 90 degree increment by having tapes already at the right rotation on the x/y axes (it's really just a software limitation preventing this right now).

Then version 2 is other stuff (reel advancement, etc.) but once we get to version 3 we get into precise rotation control (so any angle, 360 degrees).

What is for sure is that we don't want to add anything that would need to interface with the 3d printers' electronics because we want this to be really easy to set up. So version 3's going to have to be done through clever stuff where we use the the jig and the cartesian movement of the printer to rotate the pick and place head.

Are you sure? [yes] / [no]

PointyOintment wrote 2 months ago 1 point

Don't most 3D printers have at least a channel for a fan that you could use for rotation?

Are you sure? [yes] / [no]

Rogan Dawes wrote 2 months ago null point

Many 3D printer controllers have 2 extruder channels, you could use one for controlling rotation of the component.

Are you sure? [yes] / [no]

D1plo1d wrote 2 months ago null point

notexactly/Rogan: we thought about wiring up to the fan and extruder pins. The fan pins are especially nice because it's already pwm so a rc servo would be able to hook into it directly.

The issue we saw with going that route was that it was going to complicate the set up process and it would be very specific to each 3d printer's electronics. We want the RetroPopulator to be quick to setup and quick to tare down so people can easily switch between 3D printing and pick and placing.

Also as the Makerbots of this world get progressively more blackbox we wanted something awesome and open source that would just work anywhere regardless of how closed sourced companies try to make the 3D printing space.

Are you sure? [yes] / [no]

PointyOintment wrote 3 months ago 1 point

Good works so far, but you've got serious competition here: http://hackaday.io/project/963-%24300-Pick-and-Place-%2F-3D-printer

Are you sure? [yes] / [no]

D1plo1d wrote 3 months ago null point

Yeah, I'm super excited to see their first pick and place operation. There is some pretty grandiose CAD work going on over there.

I see us as being ahead on the implementation front in that we've already written a gcode generator ("slicer") program, built our pick and place machine and had successful pick and place runs (if you haven't, check out the 1st video!).

Anyways, competition is definitely a good thing and hopefully it encourages both teams. All the power to 'em! :)

Are you sure? [yes] / [no]