Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Latent shorts on QFN package

Views: 6273

#58238

Latent shorts on QFN package | 3 March, 2009

I'm having shorts appear on my QFN packages between pins, usually only after a week or so of power-on operation. The shorts are 1K and slowly down to hard shorts over time. Typically a defective part will have failures on multiple pins. I tried having my assembly house switch to a "no-clean" process, however it hasn't eliminated the problem. I'm using Lead-free solder.

Does this problem seem familiar to anyone? I need minimally a test to tell me whether any given assembly is defective, and ideally a defect-free assembly method.

DP

reply »

#58249

Latent shorts on QFN package | 4 March, 2009

It seems familiar, but nothing that I've actually experienced.

I've read some articles about tin-whiskering in lead-free solder. I think NASA has some pretty good analysis (or, at least they did a couple of years ago). The summary is that without lead in the solder, the tin grows whiskers from the solder joint, potentially shorting out fine-pitch devices, etc.

That said, we haven't experienced anything of that nature in any of our lead-free processes, nor have any of my customers reported anything along those lines. As I said, the article I read was a number of years ago, and it's likely that there have been advances in the solder chemical formulations that prevent or minimize tin-whiskering in the time since I read that article.

cheers ..rob

reply »

#58250

Latent shorts on QFN package | 4 March, 2009

I don't think anyone can give you an answer until you've had some failure analysis done. Silver migration? Dendritic growth? Tin whiskers? Lots of possibilities, but you won't know without failure analysis.

reply »

#58252

Latent shorts on QFN package | 4 March, 2009

From your description of the problem first reduced SIR to dead short is exactly what happens with dendrite growth. What is needed to form them is low bias voltage (10 volt range is ideal with low current), a surface, 2 conductors (containing tin), ionic's and humidity.

When you see the SIR reduction to 1K that means a fern like conductive filament has been formed. Once the filament can carry a low current pure tin is transferred eventually resulting in a dead short.

In my previous studies of dendritic growth I found that the higher the tin content was the easier it was to grow dendrites (without blaming the phenomenon on lead free).

You can easily see them with a 40X magnification.

I included some pic's

Attachments:

reply »

#58253

Latent shorts on QFN package | 4 March, 2009

I have to correct something from my previous post (before someone else corrects me). You can grow dendrites from any conductive surface (Cu, Ni etc.) but the final stage would be lower SIR. In my experience other metals except for tin do not lead to dead shorts at least not in the time frame described in your post.

reply »

#58287

Latent shorts on QFN package | 6 March, 2009

Yes, We've already experienced this same problem. 1) What kind of flux are you using? No Clean? Rosin? OA? 2) If it's OA, What kind of water wash do you do? Do you use DI water? How Hot? Who's machine? What are the pressures? 3) Have you performed a cleanliness test? Do it! It's an eye opener. 4) Do you do cleanliness testing on your bare boards before assembly? Do this too! 5) What kind of surface finish is on the bare PCB? HASL, ENIG, TIN, OSP or SILVER? Everything above is something to consider. I seriously doubt your dealing with a tin wisker problem. Our industry has done a great job of resolving this issue. Due to the tight geometry for the QFN's they are very difficult to clean. They are quickly becomming a part that everyone hates due to this very problem. Pad layout, soldermask layout all play a significant part in being successful with QFN's and DFN'S in the smaller packages. I was just in Chicago at the IPC cleanliness siminar and Lockheed in Florida has been doing a lot of work on this very issue. They hate the parts. They use rosin so they have to wash the boards. the contamination gets under the part and they can't wash it out. They are going so far as to design a wash system specific to clean these little rascals. Good luck Feel free to contact me if you need anything further. Keith@winsystems.com

reply »

#58295

Latent shorts on QFN package | 6 March, 2009

I would imagine it is lead free HASL finish board from the sounds of it am I right. Cheers Greg

reply »

#58296

Latent shorts on QFN package | 6 March, 2009

ENIG

reply »

#58301

Latent shorts on QFN package | 9 March, 2009

wow enig is usually not associated with this failure

reply »

Boundary Scan

PCB Soldering Tools