Soldering IC's and Transistors

[knutolai]

I've always used sockets for my IC's and transistors. Mainly I've thought of this as a precaution so as not to overheat the semiconductors when connecting them to the PCB boards as I see them as sensitive components. Recently I've had problems with some IC's falling out of their sockets so I thought I'd start soldering them directly onto the board for a more solid build quality.
As I've never actually done this before and don't want to burn my components; is there any heat precautions to take when doing this type of soldering in particular?

thx

[rutabaga bob]

Alternately, you might consider just using some hot glue to secure them in place before boxing up the circuit.  I have done this lately with some transistors.

[newperson]

If you have been soldering for a while there really is not issue.  Some people put a little heat sink on the legs to suck up stray heat going toward the transistor.  Sockets are useful to try different parts, or put in an IC that might be super sensitive to heat/static.  For example I would not solider in a SAD1024.  They cost too much for error.  

I would be more concerned about what type of sockets you are using or why your stuff is getting banged up enough for something to fall out of a socket.  Are they milled sockets?  Those single and some double wiper things are of poor quality.  

[digi2t]

From my experience, spring sockets vary in quality. Milled socket are better, but even they tend to get a little loose if you swap out too many IC's. Ideally, a good quality socket should hold your IC just fine. My Mutron Flanger used a spring socket for the SAD1024, and in over thirty years of service, the IC never moved. Even when I removed it, to test an MN3007 adapter board, the springs felt nice and tight when I reinstalled the SAD1024. As in all things, you get what you pay for. Sockets are no different.

As insurance, I always place two blobs of hot glue at each end of the IC. I start at the bottom, where the socket meets the board, and come up, just over the edge of the IC. When the glue cools, it forms two hook-like grips on the IC. They're easy to remove, if ever you wish to change the IC. You don't need to encapsulate the whole IC, it's not necessary, and the heat of the hot glue might damage the IC.

[knutolai]

Thanks for the tips!
I'm using pretty standard sockets from Tayda (http://www.taydaelectronics.com/connectors-sockets/sockets/dip-sockets/14-pin-dip-ic-socket-adaptor-solder-type.html). Not sure if they are spring/milled/something else.

I would be more concerned about what type of sockets you are using or why your stuff is getting banged up enough for something to fall out of a socket.

I might be a pedal-beater of sorts 

Could superglue be a viable option too or is it conductive? I really dislike handling hot-glue guns

[alanp]

Superglue is a bit permanent.

[tommycataus]

+1 on the heat sinks. Before I found out about sockets, I got hold of some copper crocodile clips. Put one of these on the terminals and it will safeguard against any heat damage.

[R.G.]

There are really only three issues to consider:
(1) As mentioned, how well can you solder? If you can't reliably solder a lead in less than three seconds of the iron on the joint, use a socket. Most semiconductors are specified for three seconds of all leads at soldering temps, because they have to be wave soldered. You should use an iron and temperature and flux that will heat the joint to molten and wetted in less than three seconds. Frankly, one to one and half seconds is where you and your iron should be. If you can do that reliably, you won't damage parts soldering them in. I haven't used heat sink clips for decades, and I quit using sockets except in special cases back in the later 80s. If your soldering is bad, use sockets to make up for remedial soldering until you're skilled.
(2) Considering your soldering skills, consider the part. Silicon transistors can usually be overheated until you can't hold the part in your fingertips without damage. Silicon ICs are more delicate, but I basically ignore that. There are exceptions: germanium transistors and high-dollar ICs. Germanium devices I solder in, but I make sure the joint is clean and fluxed first, and get in, then out fast before the temperature change can be conducted up the leads to the pellet inside. High-dollar ICs, which to me means SAD1024 and a few others, I'd use a socket. Haven't done that in years.
(3) Consider whether you will want to replace the part a lot. If you're going to easter-egg parts through to see if one opamp sounds like angels and another sounds like devils, use a socket. In this case, the socket is not just a part holder, it's a defined part of the circuit function, just like input jacks are.

Using sockets makes the circuit less, not more reliable. It adds an electrical contact in series with multiple leads where even one mechanical loosening or some corrosion will make the circuit fail.

Using sockets is expensive. For common parts, the socket costs as much or more than the part, at least if it's a good one. I'd rather clip off the leads and install a new opamp than use a socket unless I knew I would be easter-egging opamps.

[midwayfair]

I've been unable to solder any IC directly to a PCB without destroying it since I started using a soldering station (with temperature control) no matter how careful I am. I actually think the station is creating static discharges. However, I think my problem is unusual, and I'm able to add a little dab of solder on one or two legs of a transistor or IC in a socket to hold it in place without destroying the part.

Those Tayda sockets suck. This kind has a much tighter grip on the IC: http://www.taydaelectronics.com/connectors-sockets/sockets/dip-sockets/8-pin-dip-ic-socket-machine-tooled.html

Using sockets is expensive. For common parts, the socket costs as much or more than the part, at least if it's a good one.

To do this, you need a good way of removing the op amp and clearing solder from the board, which could very well mean a desoldering iron. I did the math when I was trying to figure out whether to get one and you can get thousands of sockets for the cost of a desoldering iron.

[guitarpedalparts]

Thanks for the tips!
I'm using pretty standard sockets from Tayda (http://www.taydaelectronics.com/connectors-sockets/sockets/dip-sockets/14-pin-dip-ic-socket-adaptor-solder-type.html). Not sure if they are spring/milled/something else.

I would be more concerned about what type of sockets you are using or why your stuff is getting banged up enough for something to fall out of a socket.

I might be a pedal-beater of sorts 

Could superglue be a viable option too or is it conductive? I really dislike handling hot-glue guns

Those would be the spring sockets where the part is held in by...springs.  I prefer the Mill-Max sockets and have never had an IC come loose.

I used to also use sockets for all ICs (probably because I read somewhere it was "dangerous" to solder them to the board).  But keep in mind many of these parts go through wave soldering and can withstand quite a lot of heat during soldering.  Many spec sheets will show the heat a component can withstand during soldering.  In my experience, for regular IC chips such as in overdrive pedals, soldering direct to the PCB is fine if you can get a good joint in 1-2 seconds per pin with a temp of about 625F.  As a precautionary measure if you're soldering an 8-pin IC, for example, solder one side of 4 pins, go off to solder some other components, then after it's cooled solder the other side.  Of course sockets can be great in overdrives if you want to swap the chips out for experimentation.

[deadastronaut]

wow ive never had a chip fall out...

i push down on one side to get it to push in...but i leave enough ''spring ' in it to hold...they are'nt coming out easily..

[davent]

author=midwayfair link=topic=106506.msg964009#msg964009 date=1394905171

Using sockets is expensive. For common parts, the socket costs as much or more than the part, at least if it's a good one.

To do this, you need a good way of removing the op amp and clearing solder from the board, which could very well mean a desoldering iron.

I think R.G.'s sacrificing the cheap opamp in question, chop the body off and extract the legs one at a time.

[poppyman]

very interesting read. I'm sorry if it's a little bit off topic:

I recently soldered a big muff transistor to its socket because I was modding it (feedback mod).
My concern was that the switch might come loose and the wire on the transistor's lug could pull on the transistor's lug.


I made me wonder if just soldering ONE lug of an IC or a transistor to its socket could actually be a good solution. One solder joint could solidely keep it in place but also would make swapping easier (one joint to remove to get your part back).

silly?

[R.G.]

I've been unable to solder any IC directly to a PCB without destroying it since I started using a soldering station (with temperature control) no matter how careful I am. I actually think the station is creating static discharges. However, I think my problem is unusual, and I'm able to add a little dab of solder on one or two legs of a transistor or IC in a socket to hold it in place without destroying the part.

If that happened to me, I would do some really quick testing on the soldering station. I can't think of any good way for that to happen that does not involve a defective (and possibly dangerous!) soldering station.

To do this, you need a good way of removing the op amp and clearing solder from the board, which could very well mean a desoldering iron. I did the math when I was trying to figure out whether to get one and you can get thousands of sockets for the cost of a desoldering iron.

There are two ways. First, get a spool of good desoldering braid and practice. This works best with temperatures of over 650F/343C and lead bearing solders. Lead free solder is essentially useless and permanently destroys any electronic parts it touches    er, I mean, lead free solder is much, much, much more difficult to work with. I flood the joint of lead free solders with lead bearing solder and alloys it down so the desoldering braid works.

The comments on solder and temperature go extra for a fancy desoldering station, of which I have two, one from ebay for $30, another given to me by someone who was disgusted with it. I'm disgusted with both of them, and use desoldering braid or fix #2, which is a manual solder sucker. These tend to cause lifted pads, though.

The best way to do it is to decide that an IC is disposable if you solder it in, and then if you want to remove it, cut all the legs off at the body, lift them out one at a time, then clear the holes with desoldering braid or poke a tiny splint of wood through the hole as you melt the solder. This clears the hole FAST.

I did a re-look up at the parts cost. The cheapest 8-pin DIP socket Mouser stocks is $0.12. The cheapest good one, useful for more than one session of swapping stuff is $0.42. A TL072 is $0.62. So I am incorrect - the sockets are not more expensive than the parts. They are slightly cheaper for good ones, and about 20% of the price for cheap ones. I find that what has fooled me is that the price of DIP ICs has moved up as their quantities have gone down. SMD parts are getting cheaper, but through hole is getting worse.

But I will still skip the socket. The economics are not much changed.

[midwayfair]

If that happened to me, I would do some really quick testing on the soldering station. I can't think of any good way for that to happen that does not involve a defective (and possibly dangerous!) soldering station.

Two prong adapter for a three prong plug, and no three prong sockets near the workspace. A friend of mine is a fire department investigator. I was giving him a soldering lesson and he chewed me out for it already.

[R.G.]

That's a little scary. It's bad enough that you're getting bad results, but we don't want you to die in the cause of making effects. Get that checked out. Haul the thing off to a three prong plug and see if it pops breakers - or itself! - when plugged into a properly grounded outlet.

If it were me, I would not use it until it was proven safe.

One step you can take right now is to use your ohmmeter. Unplug the solder station and check for resistance between the [UNPLUGGED!] ground prong and all exposed metal surfaces you can touch. If any of them do not show continuity, it's potentially dangerous.

If they do, then the problem may be to the circuit parts more than you. If, for instance, it has some non-negligible capacitance between the AC line and the iron tip, there's hundreds of volts on one side of that cap, and some leakage across. This may well be enough to pop ICs.

[midwayfair]

That's a little scary. It's bad enough that you're getting bad results, but we don't want you to die in the cause of making effects. Get that checked out. Haul the thing off to a three prong plug and see if it pops breakers - or itself! - when plugged into a properly grounded outlet.

If it were me, I would not use it until it was proven safe.

One step you can take right now is to use your ohmmeter. Unplug the solder station and check for resistance between the [UNPLUGGED!] ground prong and all exposed metal surfaces you can touch. If any of them do not show continuity, it's potentially dangerous.

If they do, then the problem may be to the circuit parts more than you. If, for instance, it has some non-negligible capacitance between the AC line and the iron tip, there's hundreds of volts on one side of that cap, and some leakage across. This may well be enough to pop ICs.

RG, thanks for looking out. Ground checks out as you described and it doesn't pop a breaker when I plug it into a three prong.

knutolai, sorry about hijacking your thread.

[knutolai]

no problem at all. Interesting stuff really 

[Mark Hammer]

1) The 40-pin milled socket strips are very handy, and generally cheaper than buying milled 8, 14, or 16-pin sockets.  Buy a bunch of those and you can simply snap off as many pins as you need; 2 x 4, 2 x 7, 2 x 8, etc.  The only downside is that they don't indicate where pin 1 is supposed to go, so the burden of remembering that part falls on the user.

2) If you're like me, then many of the semiconductors in your parts drawers will have pins that are NOT gleaming shiny by any stretch.  At least part of the problems people have in soldering come from the obstacle to rapid soldering that such tarnishing presents.  These days, I tend to inspect the legs of the ICs and transistors, and if I think they're a little too dull, I scrape 'em shiny with an X-acto knife.  A clean surface will let you form a good joint much faster.  If you want to expedite joint formation even more, make sure the pad on the board is every but as shiny and clean, and maybe dab a smidgen of liquid flux that on the pins.

3) When it comes to ICs, part of the "heat challenge" is cumulative heat buildup.  The two IC pins you just soldered may have been done quickly, but the IC is still quite warm; the heat having been conducted up the pin and across the body of the chip.  Continuing to solder only adds more and more heat to the chip, effectively lowering the critical threshold for frying the last few pins.  Leave a little time for the device to cool down before soldering any more pins.  It's also not unreasonable to space the pins out as you solder, so that any heat is spread across as broad an area (and dissipated) as much as possible.  So, for example, on a 14-pin DIP, do pins 1 and 11, wait a bit, then 2 and 10, wait, 3 and 9, etc.

[pinkjimiphoton]

to be fair, i don't like the 40 hole milled ones. they are way lower quality than most of the sockets i've gotten, and they wear out REALLY quick if ya swap components in and out. they may work better with ic's but with transistors, i find them to totally suck... stuff will fall out of them sometimes with very little provocation.

for transistors, i REALLY like the 6pin milled ic sockets. since nobody seems to use 6 pin ic's for anything... cut 'em in half, and ya have two solid transistor sockets where the pins WILL stay stuck. using those i've had zero failures, i've had multiple failures with them 40 pin deals. they're easy to work with, but the only way i trust 'em is to use oldschoolanalog's trick of cracking off the plastic after the fact, and quickly soldering one or more pins of the transistor. but... it kinda defeats the purpose of using a socket.

the first od i built wags died en route... the dang transistor had fallen loose from the socket (cut down 40 pin one like everyone uses) and since then i won't use 'em... have had zero failures like that since i stopped using them.

of course, ymmv i guess.