Need opinion about CTE mismatch on PCB

Thread Starter

aq_blues

Joined Jul 20, 2011
14
Hi guys,
Recently I encountered a solder crack issue on my PCB and was wondering if anyone else has ever encountered such (I am sure someone here would have).
The problem is caused by an IC which has 48pin TSOP package, possibly with an alloy42 lead frame (CTE = 5 ppm/degC). Cause of high-temperature requirements, my PCB is polyimide (Arlon85: CTE = ~19 ppm/degC). I was testing temperature cycles (-20degC to +125degC, ramp = 3degC/min) on this PCB and after a few thousand cycles, the solder started developing cracks. I've tried SAC305, SN63, and HMP(Pb93.5) so far and now testing Au80 as a last resort.
One solution might be to change PCB material, but that would need extra qualifications. I was reading about CIC and CBM, but CIC is magnetic and such. Any recommendations here?
Another solution might be to mount this IC on a custom package (to match CTEs appropriately). Keep leads to be Copper (CTE = 17 ppm/degC) and match package material to alloy42. But again, extensive qualification would pursue.
Thanks for the ears :)
 

andrewmm

Joined Feb 25, 2011
574
In my experience, the IC manufacturer is the expert at this sort of question,
even things such as changing the pad shape and process / thermal soldering profile alter the crystalline structure of the solder joint,

One thought, lead.
parts with lead solder on the leads can have this problem when soldered lead free.
These are lead free parts ?
Are you allowed to use solder with lead in it ?
 

Thread Starter

aq_blues

Joined Jul 20, 2011
14
Thanks Andrewmm for the reply.
I actually tested with a variety of solders. Some were better than other (lead-free yields better than leaded solder actually)
I tested SN63Pb37, SN96Ag3.5Cu0.5, Pb93.5Sn5Ag1.5 so far.
AQ
 
Top