Author Topic: Industrial-Grade Pick'n'Place-Machine DIY  (Read 1569 times)

0 Members and 1 Guest are viewing this topic.

Offline TynogcTopic starter

  • Newbie
  • Posts: 2
  • Country: de
Industrial-Grade Pick'n'Place-Machine DIY
« on: January 29, 2019, 08:15:46 pm »
Hi everyone!

In this topic I want to share my work on building a DIY Pick-and-Place-Machine. But none of those ordinary DIY-ones, the end-goal is building a Machine on industrial level, complete with Software and open-source hardware.

General Specifications:
Target would be around 120-240 Components per Minute on an 8-Spindel, rotating Placement-Head.
Accuracy up to 4um, full Software Vision System for part detection up to 0201.
Carbon Y-Axis, Servo-Driven.
Build-Space 400x600mm, 50 Feeders, automatic Tray-Feeders, and so on.

Open source:
Hardware and Software will be under CC-BY-SA, however with the exclusion of 'selling' the Software (a Programmer got to live from something :P ).

The Software:
Will be posted to Github as soon as it is presentable, as of now it is a lot of Proof-Of-Concept glued together.
The software is written in a Java-Frontend with Open-Computer-Vision and OpenGL backends.
In General it is build to fix some issues I had with OpenPNP.

Just a Disclaimer: I am working around 1.5 years on all of it... and am still not very far. So don't expect it to go anywhere quickly :D .
Of cause I'm happy to answer Your questions and follow Ideas given!

Images: My Software detecting a Part and the Carbon-Fiber Y-Axis (Still in the making ofc.)
 

Offline Doctorandus_P

  • Super Contributor
  • ***
  • Posts: 4004
  • Country: nl
Re: Industrial-Grade Pick'n'Place-Machine DIY
« Reply #1 on: January 30, 2019, 07:06:12 am »
I wish you lots of fun with your project. I really do but:

You seem to be designing and then building your machine one part at a time.
This is a horrible way to design complex machines, especially from expensive material such as carbon fiber plate.

I stopped doing that quite some time ago.
I've been working on a fully 6 axis miniature industial robot, with sort of the idea in my mind that I want resolution and rigididy good enough that I can use to (also) use it as an P&P for electronics part for components of 0603 format and 0.5mm pitch QFP.

I've been working on it on and off for almost 3 years and I still have not built any real parts.
The whole is "just" a 3D drawing in FreeCAD.
In those years I've designed and scrapped quite a few parts, because I got better Ideas during the design. This thing would have been very expensive If I had made them all out of carbon fiber.

Sometimes all the 3D drawing gets a bit boring / confusing. I have also made a few mockups out of cardboard by printing out parts on paper, gluing them on cardboard and cutting them out with a knife, and then gluig the parts into a 3D mockup.

With your fancy 3D router you can also make mockup parts out of plywood or other cheap material, and then iterate and improve your design. You can also use tape or hot snot to keep this together. It only has to be just strong enough to keep the parts together easonably well.

And then, when the design is complete route all parts out of carbon fiber or other high quality material for final assembly.

This is a tremendous speed up of the work flow, and this allows you to test many more ideas and improve upon them. No need of carefully deburring the Carbon fibre. No time spent in carefull alignment of parts etc.
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15794
  • Country: fr
Re: Industrial-Grade Pick'n'Place-Machine DIY
« Reply #2 on: January 30, 2019, 02:58:27 pm »
Nice project.
But 4µm accuracy in a repeatable way? Good luck.
 

Offline Domagoj T

  • Frequent Contributor
  • **
  • Posts: 505
  • Country: hr
Re: Industrial-Grade Pick'n'Place-Machine DIY
« Reply #3 on: January 30, 2019, 05:14:14 pm »
Nice project.
But 4µm accuracy in a repeatable way? Good luck.
Not only difficult, but of questionable necessity. 0201 is 0.6 x 0.3 mm. Wouldn't 0.03 mm resolution be significantly more than enough to plop the component on the pad and allow the reflow do the rest?
 

Offline TynogcTopic starter

  • Newbie
  • Posts: 2
  • Country: de
Re: Industrial-Grade Pick'n'Place-Machine DIY
« Reply #4 on: February 02, 2019, 09:42:48 am »
Not only difficult, but of questionable necessity. 0201 is 0.6 x 0.3 mm. Wouldn't 0.03 mm resolution be significantly more than enough to plop the component on the pad and allow the reflow do the rest?

Might have been a bit inaccurate on my Part: 4um is the Positioning-Accuracy, so the overall minimum in resolution for movement.
The Repeating-Accuracy will be a bit worse like 10um or so, but there is no way to tell how this is going to be until I get a prototype done ;) So I better aim high on positioning.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf