PLCC socket problems

So, the socket wasn't mechanically stiff enough (?). Interesting. Most of the PLCC sockets I've used were large and bulky. "Flimsy" wouldn't be a term I would have applied to them.

So, the pins nearest the corners probably exhibited the least number of problems while those midway along each side would be most vulnerable (?)

Agreed. Sockets/Connectors and Caps are the root of all evil. But, in the event you *need* a socket in the future, you'll just look for something "beefier"?

Yep, mid way along the sides showed the biggest problems. Could have been "fretting corrosion" which someone else mentioned. Re-seating the ic a number of times would fix the problem temporarily.

Preferably dont use sockets! Same for any interconnects, get rid of as many as possible.

an

sides

at

I have seen PLCC sockets that flex due to hairline cracks in the corners. Likely caused by customers monkeying around or inserting backwards as mentioned in other posts.

I have seen PLCC sockets that flex due to hairline cracks in the corners. Likely caused by customers monkeying around or inserting backwards as mentioned in other posts.

=============================================

Also happens when the "base plate" is picked out to make acess for hand soldering easier on prototypes.

[attributions elided]

Of course, that's not always possible (e.g., connecting to the Field, etc.).

I find my biggest use of sockets is when using SMT devices on boards that must be thru-hole (e.g., many devices intended for 2nd and 3rd world countries where "service facilities" are often scant and "behind the times").

The biggest problem I see with socketed components is that it encourages (hmmm... perhaps too strong a word? maybe "tempts"?) folks to remove the socketed device -- if only to immediately reseat it. As if they are thinking, "Hmmm, maybe this component is 'broken'?" even though they don't really *do* anything with it once removed.

Often, they don't use the right sort of extractor to remove the device -- it seems a tiny "standard-tip" screwdriver is the extraction tool of choice -- and the socket gets beat up in the process (even gouging contacts on the device!)

I had tried gluing devices into their sockets (this can be a real crapshoot if you're sloppy with the adhesive). But, this just aggravates the above behavior -- they sit there

*digging* at the component, wondering why it isn't coming out.

Most recently, I tried "encapsulating" the device so that it is obvious that the only way to remove it is to unsolder the socket. We'll see how that works... :>

Memory modules are a real PITA. They (SIMM, DIMM, etc.) are an economical and versatile way of handling memory configuration options. But, where a socketed "IC" might "tempt" someone to remove it, socketed memory devices seem to SCREAM: "Remove Me!!!"

And, they tend to be easier to mis-install thereafter (put in the module slightly crooked so it *looks* like it is in place but really isn't). Making matters worse, many of these sockets have really low insertion cycle ratings which leads to flakey long-term behavior.

I haven't, yet, found a way around this that leverages the economy and ubiquity of the modules... :< Maybe pot the whole damn device!?!

But for exactly those same reasons it's at least equally bad an idea to use a socket. You've just laid out the explanation yourself.

Those people who can't properly repair SMT boards for lack of tools or skills will, because of that same lack, damage your sockets trying to do the one thing (they think) they can: take out the socketed device, and put it back in.

So, no, sockets are not the solution to that problem. Daughter boards, OTOH, might be.

But still, it will appear to them to be the easiest thing they can do to the whole device. By putting it into a socket, you've just turned that CPU into the lowest-hanging fruit around, so don't blame people for picking that first.

Say the only symptoms they've been able to find up to that point were "power consumption higher than specified" and "CPU rather hot to the touch". So unplug the CPU and see if that brings power consumption down to what they usually get for a CPU-deprived board. If so, exchange CPU. If not, make measurements right there at the empty socket pins until a difference shows up.

If you're seriously worried about badly unqualified people tampering with it: absolutely.

Well, either that or put a dye pack into the enclosure that will explode into the perpetrator's face, plus in the complementary color to the dye, a _huge_ sticker stating an acceptable equivalent of "Hey moron! We wrote 'No user-serviceable parts inside! Seriously!' on the cover you just removed, and we bloody well meant it!"

Some parts just aren't *available* in a thru-hole configuration. Looking for some *other* component often leaves you with a tougher implementation (if you can even meet the goals at all!)

Exactly! Just like folks pulling "tubes" out of TV sets (ages ago) in the hope that this would magically "fix" a broken set...

Any sort of connector invites trouble. People mate things "backwards" (regardless of the presence of a key), bend connector pins, etc.

That;s why I've resorted to making the socket+chip a single assembly. In a sense, converting a SMT device into a thru-hole device. This makes the particular device one of the *least* desirable things to tamper with -- because the consequences of "breaking it" are very obvious ("How will we ever repair *that*??")

The problem is making devices so that they *can* be serviced by "local (um) 'professionals'" with the sorts of tools and skills that they have available. Shipping something back to the US for service gets expensive, fast!

Often, a little knowledge is dangerous. I.e., if they are skilled enough to use a soldering iron, they *can* remove a thru-hole device and replace it. But, since this represents "work"/effort, they will look for easier things to try -- without necessarily being aware of the consequences of those other actions.

I've looked at *soldering* SIMMs to PCB's as a way of "discouraging" their casual removal/replacement, etc.

I've seen people plug power connectors into ATA drives UPSIDE DOWN! When I exclaimed, "Didn't it *feel* wrong to be doing this when you did it? I means, *really* FEEL WRONG???", they just shrugged. (try it, some time. I can't see how you could *easily* do this without hands the size of ham hocks *or* a pair of pliers!)

We use spray electrical contact cleaner, applied to a bush, then to the PLCC devices edges. There was a different design PLCC socket from Tyco ( now TE CONNECTIVITY ) , I think the 'current' number is 1571541-x - not cheap, but worth a try if you have the slightly larger holes it needs.