We manufacture pcbs and have used Taiyo LPI for over 7 years. I have no
experience using clean cycles or fluxes, but I offer the following practical
observations.

It is possible to get white spots on LPI soldermask when certain assembly
fluxes react with under-cured soldermask. This is true of Taiyo as well. We
believe that a simple test for this condition is to apply a hot soldering
iron to the white spots. If they become clear, then this is the case. 

At the time we experienced problems, Taiyo LPI was basically a
"thermal-cured" mask. That is, the original process specifications from
Taiyo said to simply bake the boards after development to achieve final
cure. We found that mask/flux interactions (white spots) were totally
eliminated if we gave the boards a "UV bump" after the bake cycle (similar
to UV cure cycles required by other masks on the market).

I offer no explanation for peeling/deteriorating mask. I have no way of
knowing if your vendor currently UV bumps the pcbs, nor have I read Taiyo's
latest process specifications to see if they now recommend the additional UV
bump. I offer this only because the soldering iron is such a quick and low
cost test that you couldn't lose by trying it.

Glynn Shaw, President
Proto Engineering Corp.
181 Commercial St.
Sunnyvale, Ca. 94086

email: [log in to unmask]
phone: (408)738-0693
  fax: (408)738-1290
modem: (408)738-1388