I Always use sockets now Re: [sdiy] IC socket question

[harrybissell]

Hah... now I have to disagree with Magnus

(Hej Magnus ;^)

The sockets that grip from the thin edge of the IC pin can be a
problem. Yes... the contact area is smaller and there is more chance
of the tiny contact point being gas tight...

... but the .1" spacing of the pins does not leave much room for a decent
spring arrangement. These are usually a torsion (tuning fork style) spring
that has a limited travel. Some ICs are fine, some fit badly. The insertion
holes in the socket are usually small and alignment can be hard. They are
usually loaded from the bottom of the socket and can trap flux.

The "holds on the flat side" sockets usually have a ramp leading into the
contact area formed by the outer leaf. There is some mechanical wiping
action that scrapes the contact area clean...  The inner spring contact is rolled
behind itself (its had to describe - have to cut them open)... which makes
for excellent retention of a wide variation in pin thickness.

And (as Magnus correctly pointed out) makes the insertion force VERY high
on large pin count packages.  A 40 pin chip takes POUNDS of insertion force...
better hold from behind or you break the board.

These sockets are really good for 8 - maybe 20 pin DIP. I never like the sockets
for 6 pin dip no matter which kind... there's no retention at all...

ZIF sockets anyone ???

H^) harry

Magnus Danielson wrote:

> From: Cynthia Webster <cynthia.webster at gte.net>
> Subject: I Always use sockets now Re: [sdiy] IC socket question
> Date: Thu, 12 Dec 2002 16:54:23 -0800
>
> Hi Cynthia,
>
> > This is interesting to me because I'd been taught to insist on machined pin
> > sockets over all others.  They call them "High Reliability" yet many folks
> > here say otherwise!  (I know that they can be pure hell sometimes
> > -just getting the chips into them without bending pins!)
>
> There is a real issue with ICs and sockets. DIPs are made so that when they
> are put onto the PCB, they will hopefully stay there until you've soldered.
> This is done by letting the pins be slight bent outwards, thus allowing for
> them to work as a kind of spring holding them put in their holes.
>
> Naturally that makes the actual placement of holes being slightly denser than
> where the lower ends of the pins actually hit. If you have a socket with
> propper dimensions you will experience this as a problem. There are however two
> solutions to this problem:
>
> 1) Slightly bend the pins by a small plier. Personally I use my favorite tool,
>    a small Stirex to do the work. You better work pin-for-pin in order to get
>    a good result. While this is a bit labour intensive (doesn't take that much
>    time once you've got the hang of it) it will make the insertion process
>    much more painless.
>
> 2) Use a professional inserter. There are specially designed inserters. They
>    will grab the chip and bend all the pins flat down for you. You only need to
>    aim and release to get it right. A good one surely cost money, but I am
>    quite sure that this is the way to go for serious buissness folks. A good
>    tool in the right hand both makes the job quicker and better and is worth
>    it's money just cause you get less hassel and better quality.
>
> Since what I do normally fall into DIY, I choose option 1, but you might want
> to consider option 2. I'd at least recommend you to consider the option.
>
> > Personally I always use military grade machined pin sockets on every chip
> > everywhere now, (Gold if possible) because at the rate things are going with
> > surface mount components and miniaturization... even lowly 741 DIP Opamps
> > may sell for well over ten-dollars each sooner than you think!
>
> Indeed. However, SMD is getting a hole lot easier to do. It's a matter of
> training. Maybe it's time to make that "SMD DIY" site that a friend of mine
> (who's well beyond DIY level into advanced prototyping) dreamed of doing
> since he feels many people have an unecessary fear of SMD.
>
> > For example I have some CA3094s that have turned out to be quite valuable
> > and because I'd been using sockets, I was able to easily salvage them as
> > practically good as new.
>
> Good point.
>
> > I'd be curious though to hear if anyone thinks the use of sockets like this
> > actually ~causes~ more reliability problems than they fix?
>
> IMHO will a good socket only lower the reliability (MTBF) slightly, but with
> many other benefits. A bad socket is however a problem.
>
> There is however an issue with all sockets you must recall, and that is how
> they handle the tin on the IC components leg. All (but a few) DILs and SILs
> have their pins covered with tin. This is to make them "wet" quickly in the
> soldering process. The downside to this is that tin will oxidice which also
> makes ICs sensitive to long storage times, they are intended to be used quite
> quickly after manufacturing. When you solder on those legs, you burn through
> what ever oxide is there with your solder's flux and embedd the pin in solder
> and if you've done that process right and doesn't have made any other huge
> errors (like mechanical stress, hazardous environment, thermically induced
> mechanic stress etc) it will survive the lifetime. I've seen so many cases of
> solderings that does not survive the lifetime of the box itself.
>
> Anyway, most sockets only mechanically press onto the pin so that it touches
> the tin. If you have a hazardous environment, or you only make a narrow
> point-contact, or you experience much movements of the contact, you run into
> the possibilty of that tin sheat to go oxidizing on you anyway. You could rub
> of oxide by mechanical exercising.
>
> You could imagine sockets grabbing the pin so hard that it cuts through the
> tin layer for sure (it probably happends anyway to some degree for certain
> holders), but the insertion force of those I don't want to think about.
>
> The sockets I've seen (and now this is from memory) either spring to the flat
> side (cheap once) or grab on the corner sides of the pins (expensive onces).
> The later onces is able to cut through I think. The solder can glide away from
> the corner, and there's 4 points of contact which is pair-wise pressing against
> each other. I rarely use sockets, but prefer the later onces for obvious
> reasons.
>
> I haven't studdied the actual details of sockets as of lately, but at least you
> got some of the knowledge that I've got on the issue. Possibly it makes it
> easier to understand what the manufactures actually say and mean.
>
> > I usually wait until a PCB board is totally wired to the pots and the
> > switches and everything else ~before~ stuffing the chips into the sockets
> > for the first time.
> >
> > Does anyone think this is being too paranoid?
>
> Well, no... I would do it like that without being paranoid, and so should you!
>
> There's no reason to stuff them prior to that point anyway. When you've got
> sockets stuffing them is what comes on about the last thing for a board.
> So relax and just think it as "my habbit".
>
> Cheers,
> Magnus