Even if the boards look clean, you should do a Control test to investigate the surface energy of the boards by using dyne pens or a contact angle test, where you measure the contact angle of a droplet to see how much the droplet beads up or wets out onto the surface.
I suspect when you do the droplet test, you'll find the beads are quite round and do not wet, indicating low surface energy and poor bondability.
Next, as an experiment you should do a suitable gas plasma treatment on some boards (usually using argon and/or oxygen gas), and repeat the surface energy tests. I think you'll see the drops wet out quite a bit at that point.
That's because the plasma treatment activates the surface of the board and improves bondability and adhesion. (Using this technique, one can even make non-stick surfaces like Teflon wettable and bondable.)
Finally, you can bond actual parts to both Control and plasma treated boards to confirm the improvement in adhesion.
If you have access to a plasma system in your facility, we can advise you on how to set up your DOE. Once you confirm increasing surface energy with plasma resolves your bonding issue, the final step is to figure out how to implement the treatment into your production environment.
