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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

bad connections, or not?


bad connections, or not? | 12 January, 2001

I've got a question:

Can a joint look good externally, even under 45X magnification, and not be mechanically and/or electrically sound? I thought that if the wetting angles were sufficient, there's no discoloration from extreme thermal distress, or unusual surface anomolies, then the joint should be good.

We've sent some samples (this is a large 4 leaded oscillator with J-leads, can't remember the package style) out for sectioning and the preliminary results are that the connections are fine, but I'm not sure even this information is going to placate design. They don't wanna be at fault because they can't find a flaw. They can't even I.D. the specific component (the UART or the oscillator) let alone which trace the failures are occurring on. I've done everything I can in-house and then some and can't come up with any process issues.

Any clues? If not, feel free to just shoot me. I've just about had it with the politics here.

reply »


bad connections, or not? | 15 January, 2001

Let me make several points on this issue �

POINT 1: Let�s be basic � Soldering and welding are examples of process that the quality types call "special processes".

Special process. The results of special processes cannot be verified fully by subsequent nondestructive inspections.

Soldering is a special process because it is difficult to test the quality of solder connections without damaging the connection. Companies control and perform special processes in accordance with written procedures.

POINT 2: Design types sometimes get very conservative near the end of design projects, possibly "butt covering" because: * Compound effect of all the "simplifying assumptions" they�ve made could come to roost. * Turning control of the project over to someone else provides an opening to broader criticism. * Want to do a good job * Don�t know what they�re doing.

POINT 3: Pushing forward, apart from � * Solder connections appear nice * You�re frustrated

� I don�t really understand much about the situation [but I�ve never let not that stop me from rattling on about a topic or my perception of what was the topic]. And as far as being on frustrated, you�ve come to the correct place. Since we don�t know what you�re talking about, we can only offer platitudes that tend to make you more frustrated [while not the intent]. Cool, eh?

When I�m frustrated, I really like to pull-out the ol� "Bozzo The Clown" blow-up punching bag and give it a cuppla upper cuts to the jaw. Unfortunately, the good feeling I get from smacking the clown is always so transitory, since Bozzo is always smiling and always pops back-up for an other shot to the snout. I wish I could do that!!! Anywho, I digress.

POINT 4: When you couldn�t figure-out what was wrong, you went with the gut � "Let�s bust the thing and see if we want figure-out what�s wrong with it." [A bit like medieval scientists carving-up people to try and figure-out how the human body worked.] Consider continuing with the approach by performing the following: * Design qualification. This highly accelerated stress test literally beats the living stuff out of products to identify design, component, process, etc weaknesses in order to correct these weaknesses before exposing customers the product. For an overview, checkout "Accelerated Reliability Engineering: HALT And HASS" Hobbs, J Wiley & Sons 2000 [047197966X] * Accelerated reliability testing. This accelerated test strips about a year from the life of a product before shipment to provide confidence that the design and manufacturing process are capable of meeting the goals for product performance. For recommendations and strategies, look at "Guidelines For Accelerated Reliability Testing Of Surface Mount Solder Connections" [IPC-SM-785]

reply »


bad connections, or not? | 19 January, 2001

Thanks Dave. Talked to a few people at APEX this week and if nothing else, had my frustration reduced some. Of course I made sure I talked to mfg. people and avoided the design guys like the plague. ;)

I actually picked up a copy of 785 a while back. When we come up with the cash for the chamber I plan to do some testing. In the mean time, the lab that did the x-sections for us found nothing wrong as far as solderability or any indication of anything out of the ordinary thermally, but they DID find some fine cracks (don't know how to quantify "fine" yet, but I'll get 200x pics. with the hard copy of the report. That's just one sample and I don't believe that the cracks are anything that would open the connection, but it's something we gotta look at. The proximity of the parts in question to the depanelng score is a possible (probable, from my biased perspective) cause here. Design generated problem, but up to the process people to deal with.

Anyway we're working our end, and I've received some messages from the guys in the other building that indicate they've begun to at least consider this to be a design related issue, so maybe all is not as bad as it seems.

reply »


Re: bad connections, or not? | 19 January, 2001

Frankly I have no clue about whether or not the parts are good because our receiving inspection team is equipped with nothing more than a couple of caliper sets and a partial gauge pin set. There ARE things I miss about the medical instruments business.

The part is used elsewhere, however, and we don't see the same problems there. As I mentioned in my reply to DaveF, we may have some problems down stream from the SMT process (now it's MY turn to throw it over the wall) that we can attribute some joint damage to, but there's nothing conclusive as to whether or not this damage is extensive enough to cause the failures we're seeing.

The story continues.........

reply »

Software for SMT

One stop service for all SMT and PCB needs