I'm currently working on bringing a manual coating process/workcenter online and my QA team is giving me some grief.
For background, we're using Humiseal 1B31 thinned with thinner 521, mixed at 1:1 by volume. Spraying through a Binks HVLP with 10 PSI inlet pressure at the gun. Below is a rough process outline:
- Wash (Aquastorm with chemistry, can't remember what we're using off the top of my head) - Bake (about 100c for a couple hours) - Plasma Clean (confirmed that adhesion is greatly increased, waiting for dyne pens to confirm, contact angles on DI water drops go from near 60 degrees, to almost nothing at all, so surface energy shouldn't be a problem) - Mask (ATS-667 tape for masking, known compatible with AR coatings, no silicone adhesives) - Plasma Clean again if too much time has passed during mask stage - Coat, 3 coats per side, resulting in 2-2.5 mil finished thickness as measured on an aluminum coupon
On some smaller chip resistors (0201/0402) the coating almost appears to be missing (images attached), but from the images alone and looking at the assembly I'm not seeing any signs of de-wetting on the resistors, just a darker part due to the resistors already being black. I also used a plastic probe to scrape coating off the top of one of these resistors and could verify that the thickness is about the same as on chip caps.
Mostly looking for opinions based on the pics (the top of some larger parts are intentionally mask, ignore those).
Images... http://imgur.com/a/c1cYo
Thanks! -Phil
reply »