After installing the nozzle changer and setting it up in the software we started to do a test placement session to test the placement using automated nozzle change. One big problem was, that it was not possible to find proper filter setting for nozzle calibration. With only two different filter sets but six nozzles this is really hard. But that's an other story!
After setting up the Tapes and job data we started to place the parts. But already with the first part 0603 0.1uF we run into a problem. The part was picked (nozzle N045) up and placed on the board, but after vacuum off and moving up the nozzle to our surprise the part was not on the board but still on the nozzle!
Even moving to pick up the next part did not cause the part to fall off. We tried this with a couple of parts e.g 0805 10uF, SOT23-5, also we switched to N08 nozzle, but always the same result. Cleaning the nozzles did not change anything.
The only reason we can imaging is that the nozzles were magnetized by the magnets in the nozzle changer. Since many SMD parts are magnetic and the weight of the parts goes to nearly zero, magnetic nozzles are not really good.
Ok, one could argue that with solder paste the parts may stick on the board, but for setting up a new board we need to test the placement without solder paste.
Should we think about a nozzle changer which uses some spring mechanism instead of magnets ?
The magnets do a nice job of "pulling" it into place. With a purely mechanical solution, I think we would need a special spring or a milled part with a small ramp which slides nozzle into place... Maybe you have a brilliant idea on this?