Interesting idea. I've always done all of those parts with tweezers and eyeballs, but some of the TSSOP packages can be a bit of a challenge to line up well enough for reflow.<p>If you used FR4 instead of acrylic or whatever, you might be able to leave the aligner stencil in place for reflow. That might let you do some of the smaller parts.
There are still dirt-cheap gadgets [1] that you can make before going for a liteplacer.<p>[1] <a href="http://vpapanik.blogspot.com/2012/11/low-budget-manual-pick-place.html?view=classic" rel="nofollow">http://vpapanik.blogspot.com/2012/11/low-budget-manual-pick-...</a>
DirtyPCBs has never done me wrong, and they offer a cheap laser acrylic service, hmmm!<p><a href="http://dirtypcbs.com/store/lasercut" rel="nofollow">http://dirtypcbs.com/store/lasercut</a>
A simple improvement. add pins (even temporarily) to the PCB. Add pin holes to the top acrylic template. No need for the lower (outer) acrylic part. Plus the pin holes force very good alignment.
This is really cool. I wonder if it would make hand assembly worth it for small runs of products? PCBs are already so cheap that assembly is by far the biggest cost of making anything on a circuit board.
<i>"While this setup has been working really well for large parts, small components haven't worked out"</i>.<p>Too bad. This is a good idea. I have some boards with lots of small components where this would help, and I have access to laser cutters. Maybe with a thinner plastic layer on top... Also, with small components, getting the placement stencil off without taking the tiny components with it may be tough.<p>Board edge shearing is less precise than the pad placement, and this thing uses a base frame to align to the board edge. Pick and place machines align to reference marks on the board, using cameras. You might need some kind of fine screw adjustment to move the board very slightly.<p>I've seen videos of the Liteplacer, which is a slow pick and place machine for prototypes. It's useful for when you are only making a very small number of boards, because it can work from components laid out in wells or on tape that's not on reels. The production oriented machines require that you have enough components on reels to get the feeders started.
Unrelated, but on the termdriver page, CSI codes C and D are labeled as both being "cursor down", when they should be "cursor forward" and "cursor back".
What a great idea! Might be tricky to make this work with fine-pitch QFPs, but for QFN and wide-pitch ICs (e.g. SOIC) I can see this working really well.
I don't get it. If you need 0.1mm accuracy, all the machining, on plastic, mind you, must be done to say 0.05mm accuracy, both in position and size. And, it appears that one of the parts registers to the edge of the pc board, which is inaccurate. Since you got it to work, what am I missing here?
Clever! I'm thinking a similar design with a range of hole sizes and wider borders could be a handy generic tool - might have to design and order one next time I'm making boards with QFNs.<p>Does anyone here know what the smallest practical hole size is in acrylic?
Nice, after doing a bit of electronics, I realize how much todays pcb work is unfit for hand soldering (unless you want to enjoy frustration). Having geometric help is never a waste.