Vlog #68: OpenPNP DIY open source pick and place machine (complete run) 

Indeed they seem to be quite nice feeders, even better than some store brought ones (so I have been told)

Very clever. Like a surgeon doing microsurgery with a manipulator that reduces movement.

Yes, the noise is quite annoying after a while. The repeated whirrrr-thunk of the head is quite relaxing, but the constant buzzing of the air pumps drives me nuts. In this video I pulled the sound level down a lot: it's even noiser in real life.

Yes, that's easy to do in OpenPNP. However, it means that all the feeders sit with one part exposed between boards. That's fine if you're doing a lot of boards, but I only do a few at a time so I've chosen to take the performance hit and have the feeders keep their parts covered until required.

I could also do it by changing the feeder controller to return "ok" immediately, and trusting that the feed will be completed before the head gets there. That would avoid the exposed-parts problem, but it means the feeder doesn't have a chance to report a fault before the head tries the pick.

I could still tune a bit more speed out of it by changing the acceleration settings. It doesn't reach top speed in most of those short movements. You can hear that the longer movements to the extreme right hand feeder are a higher pitch because it gets up more speed. Reducing the move mass would definitely help. The motors start missing steps if I push it much harder.

I have a T-962C reflow oven, which I showed a few years ago: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-cP2SYn7XFl4.html

@@SuperHouseTV awesome!! I will need to watch all of this. So you probably figured out I've been around SMT assembly lines :-)

Perhaps a project on robot arm to pick the finished board to oven and reload a new empty board for PnP.

Miklmaekit is just my joke name for this machine, which is a combination of various things from different sources. The feeders are the popular MGRL0816 feeders (docs.mgrl.de/maschine:pickandplace:feeder:0816feeder), the frame is a local version of the X-Carve, the software is OpenPNP, the feeder controller shield is in my GitHub account. It's evolved over the years. I did a talk about it 4 years ago, when you can see it looked quite different even though it still had the same frame: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-dvwZeb2uwkg.html

My only thought would be if you do it right before, it peals the cover off, so at the end of a run the last component would be covered where if you did it after, the last part would be uncovered and couldn't fall out upon storage. That's just my thoughts though.

Yea that does make sense, might be a thing I will change and do in the future