On Mon, 29 Jan 1996 11:43:34 PST Michael Barmuta wrote:

   From: Michael Barmuta <[log in to unmask]>
  Date: Mon, 29 Jan 1996 11:43:34 PST
  Subject: Re: SOLDERMASK DETERIORATION
  To: [log in to unmask]
  Cc: [log in to unmask]
  	Ron: I do believe what you are seeing is a result of your cleaning 
cycle.The question is,is this an unreasonable expectation of the mask? I 
don't think so.	
  								
  From the information you provideed the other three PCB fabricators 
apparently are not causing you a problem(I'm unsure if they are using Tayio 
or another LPI).The fourth fabricator supplied "acceptable" boards for a 9 
month time period after some process correction and now the problem is 
back,same PCB supplier.I don't think the last batch of boards you received 
were fully cured.							
  If I were to quess as to what your seeing I would say it's this: The heat 
cycles from your relow and flowsolder are starting to breakdown the resin 
system of the mask.Then when you wash the board with hot(165F)alkaline 
chemistry this further attacks the surface of the mask exposing the filler	
(tycically silica or talc)which absorbs moisture during the wash/rinse/dry 
  cycles.This then shows up as white spots.A quick way to check,is to use a 
heatgun	applied to the area with white spots.If they go away or lessen you   
 can be pretty sure it was  moisture.					
		
  So how do you make it better?	First of all I'm assuming you have verified 
that your cleaning process time/temp/conc.are all correct and nothing has 
changed	in your soldering operation.You could backoff a little in the wash 
 time and temp as stated earlier by other replys, which should help some but 
I think the mask is the main culprit here.	
  								
  								
  I don't think it's the type of mask,the Tayio PSR-4000 family has a long 
proven track record.IMHO the mask is not fully crosslinking during the cure 
cycle.Here's the problem,when suppliers market an LPI mask to the PCB 
  fabricators they usually tell them a UV bump is not required and thermal 
bake times can be done in 30-60 mins.Why;because time is the enemy of board 
fabricators.The more process steps and longer process times equate to higher 
board costs which can result in a non-competive board quote.So the fabricator 
 is always looking for ways to reduce board processing and it's associated 
costs,soldermask final cure however should not be one of them.When LPI 
suppliers are pressed on what cure cycle provides the best physical/elec-  
trical properties they typically call out a longer bake time 60-90 mins.AT 
TEMP.(150c) and a UV bump(3-4joules).Find out what your fabricator is 
actually doing for final cure. If your not getting something close to the 
above cycle see if they can process the next run using it and verify it.	
	 Also I'm not sure why YOU are doing a "cure bake" on your end.Is    
 this a moisture bake out or is this to compensate for an insufficient cure 
from the PCB supplier?  Good luck and no you are not alone with this problem. 
                                                                             
                                					
					Regards				
					Michael Barmuta			
					Staff Engineer			
					Fluke Corp.			
					Everett Wa.			
					(206)356-6076			
									
				
> 			   
> On Thu, 25 Jan 1996 15:48:12 -0600 rawlyk@SEDSys	 tems.ca wrote:
> 
> > From: [log in to unmask]> Date: Thu, 25 Jan 1996 15:48:12 -0600
> > Subject: SOLDERMASK DETERIORATION
> > To: [log in to unmask]
> > 
> > This message is directed to circuit card assembly shops and printed 
> wiring 
> > board manufacturers.
> > 
> > We are a medium sized company that designs and manufactures CCAs for 
> commercial
> > and military applications.  We currently have four different 
> manufacturers of 
> > PWBs.  We are experiencing a problem with one type of soldermask that 
> is 
> > provided by our vendor of choice.  The soldermask is liquid photo 
> imagable 
> > and is IPC-SM-840 class 3 certified.  The name of the soldermask is 
> PSR 4000 
> > manufactured by Taiyo America.
> > 
> > Our vendor feels that the soldermask is deteriorating after our semi 
> automated 
> > saponified aqueous cleaning process.  Over copper substrates the 
> soldermask 
> > develops small white spots that range in size approximately .010 to 
> .020 inches
> > in diameter.  Soldermask covering copper traces.010 to .015 inches 
> wide has 
> > turned white over 50 to 80% of the trace length.  This occurs both 
> over the 
> > solder side and component side of the PWB.  The deteriorated 
> soldermask on the 
> > traces then looses adhesion to the copper traces leaving bare copper 
> exposed.
> > 
> > Our assembly process is as follows: 
> > 1. Cure bake PWBs ( to recommended parameters set by the vendor to 
> "condition" 
> > the soldermask.)
> > 2. Install SMT components by hand.  ( RMA flux.)
> > 3. Install the through hole components.
> > 4. Wave solder CCAs. ( Or hand solder if wave soldering is not 
> feasible. 
> > RMA flux.)
> > 5. Clean CCAs using the batch washer.  ( Semi-aqueous method using 
> De-Ionized 
> > water and Armakleen E-2001A saponifier.)
> > 
> > The washer recipe is as follows:  ( CCAs are washed in a Unit Design 
> model 
> > ABC 2001 machine.)
> > 1. Wash cycle. Duration 4 min., temperature 165 F, 4% saponifier 
> solution added.
> > 2. Quantity 3 rinse cycles.  Duration 1 min., temperature 165 F, 0% 
> saponifier
> > solution added.
> > 3. Dry cycle.  Circulating fan. 15 minutes at 150 F.
> > 
> > The pH value for the washing solution is 10.8.
> > 
> > This problem was first noticed about 9 months ago and only on one 
> vendors PWBs.
> > At that point discussions with the PWB vendor and soldermask 
> manufacturer 
> > resulted in adjustments to our curing parameters and wash recipes.  
> The problem
> > did not totally disappear but dissipated to a level that was 
> acceptable.  The 
> > problem has resurfaced this week with a vengeance.
> > 
> > I am interested in any feedback on occurrences of this problem and 
> solutions. 
> > Specifically these questions.
> > 
> > 1. Are we the only people in this industry that are experiencing this 
> problem?
> > 2. Is our cleaning process indeed causing the damage, or is it a 
> process 
> > problem with our vendor?
> > 3. Is our saponified aqueous cleaning harmful for IPC-SM-840 class 3 
> LPI's 
> > in general?
> > 
> > Thank you for your anticipated responses.
> > 
> > Ron Rawlyk
> > Manufacturing Engineering
> > SED Systems Inc.
> > 306 933-1688
> > rawlyk @ sedsystems.ca
> > 
> > 
> > 
> 
>