Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

desoldering SMD DIP IC with a heat pad....

Status
Not open for further replies.
OK.
The part is dead. Grind off the plastic and get to the metal. You can heat that!
BUT
If you can't heat the part to melt the solder underneath then you will never be able to solder the part back on. If the solder on the heat sink is not right the part will get hot and die again.
 
Preheating the board will not involve melting the solder on the chip components on the rear of the board. Stick the whole assembly in the oven for a half hour at 150 degrees. Take it out and immediately remove the chip with your hot air wand. That way you are only requiring to raise the chip by the delta temperature in order to melt the solder. You may find that there are some glue dots below the chip too, if there are, you will need to twist it slightly back and forth before lifting it off.
 
Success! Indeed, the preheat was the answer. I had already determined my issue was the fact that the big Aluminum case must be sinking all heat so I need to get the case hot. I took it to 150F (Hotter prob would be even better), and with the hot air at 480C and my 40w iron at 480C, it took about 30sec to release. Hardly any solder there at all. Looks like I escaped pad damage too and I do not see compromise of other components. However, I know I got a few things pretty hot! Specifically the caps to the right of the chip. How hot is really the question but they are still meter testing decent.

Also, not sure if this trick I figured out is common knowledge but I dipped my solder wick in flux paste (just the tip) and it seems to do 10x better. I can run a sweeping motion and sops things right up with much less heat.


So.... to reinstall a chip, this will be really tricky. I know if I get too much solder, it will bridge some pins. I want to use a lower temp solder for the sink. I am hoping I can figure out a way to get it done. I think I will need to solder the sink first, then move over to the pins and probably manually solder each one by tinning the iron only.

Thoughts on the process?
 

Attachments

  • SDC10082.JPG
    SDC10082.JPG
    494.9 KB · Views: 168
Method 1, my first choice:

Meticulously clean the board, pads etc of any glue, old solder or flux residue and liberally flux it with a flux pen once clean. Run a couple of fine lines of solder paste towards the middle of the heatsink pad, but away from the edges and not too close together to allow it to spead out under pressure. Clean and flux the mountnig tab on the device. Again, pre-heat the assembly and drop your device in place, put a little pressure on it and hit it with your hot air wand until you feel it seat and settle. Keep the hot air on it for a little longer than you think is too much, to ensure that capillary action pulls the solder all around the pad. Let it cool and then solder up your pins after giving them a rub over with a flux pen. Clean the assembly afterwards of any and all residues left over.

Method 2, not my first choice:

Failing that, and this method assumes that you will not be attempting this again in the future, you could use a tiny amount of Chip Bond (Silver loaded epoxy), again spread around towards the middle of the pad to allow for some expansion under pressure, to bond the device back to the heatsink pad. If you do this, forget about ever removing the device again, it's very unlikely to come back off without wrecking the board in the process. Also, don't get it anywhere other than exactly where you want it, it conducts quite well and will be virtually impossible to remove from unwanted areas. Leave it to set up and cure over night with a little piece of tape holding it in place so that surface tension can't move it, then flux your pins and solder them up. You could also just solder tack two diagonally opposite pins to hold it. Again, clean the board of all residues afterwards.

Choice of solder would be application dependant. If really high working temperatures are involved, like 100 degree C and above, you will need an appropriate specialised solder to do it justice. If it's not getting above an average of 40-50 degree C, then I would just hit it with some good old 60/40 Tin/ Lead and be done with it. You could also go the lead free route, but that will raise your melting point.
 
Last edited:
I've done similar with a micro torch, you need one with a very small flame and reasonably powerfull.
 
Because of the pads on the bottom, you;ll have to use paste there and re-flow with hot air.
tunedwolf said:
Choice of solder would be application dependant. If really high working temperatures are involved, like 100 degree C and above, you will need an appropriate specialised solder to do it justice. If it's not getting above an average of 40-50 degree C, then I would just hit it with some good old 60/40 Tin/ Lead and be done with it. You could also go the lead free route, but that will raise your melting point.

I'd never use 60/40 for SMT stuff. 63/37 wire solder works great for SMT stuff. I'd Use a paste for the bottom pads.

MY FIRST removal/desolder was a processor from a car radio. No heat sink pads. Probably before ROHS. I cut the pins and removed them one at at a time primarily with solder wick and/or adding solder and vacuum desoldering.

The technique I used was to tack 2 opposing corners with 63/37 flux core solder. Then I turned the IC so it was effectively over my head. You can then WIPE the 4 sides with the solder quickly. then it was turned over, and inspected for bridges. Just heating would using remove the bridge.

63/37 has the same solidus and liquidus points, so it melts and solidifies at the same temperature making bridges less likely and easy to remove.
 
I'd never use 60/40 for SMT stuff. 63/37 wire solder works great for SMT stuff. I'd Use a paste for the bottom pads.

I wouldn't use 60/40 either, well not for a new build done in a reflow anyway. For repair work, where low device working temperatures are expected, I see no advantage in using anything special, other than perhaps a fine diameter, quality cored solder. About the only real advantage I see of 63/37 would be plasticity, being Eutectic it goes from liquid to solid state almost instantly, which might be handy, if you are working with miniscule chip parts, placing and hand soldering them. In hand soldering where joint temperature is a hit and miss affair anyway with anythng but a top notch station and a highly skilled operator, there's no great advantage of the lower melting point of 63/37 either, in fact for components that run pretty hot, it may even become a disadvantage further up the line due to joint fracturing. 60/40 will also be a smidge better at holding larger components. I'm all for using specialised alloys when the need is there for it, but I don't see it in this case :)
 
You would be surprised at what hot air temp you can get away with. The balance is temperature vs time. The chip and board will need time to get up to temperature, enough to melt the solder on the heatsink pad, and the higher the temperature, the faster this will happen. But it will also mean the temp of the chip will rise faster than the board. I tend to use around 420-450C hot air, with a nozzle about half the size of the body of the IC. I do not know if this is the preferred way, but I have reused many parts and they function fine.

It is also a good idea to mask off the rest of the board if you can with thin metal sheeting (which of course will get hot..). For parts that have heatsinks, by the time you've got the board/chip up to temperature to gentle 'nudge' the chip out of place, the surrounded parts will be sitting on molten solder - only held on by surface tension.

For replacing parts (that are dead) then sacrificing that old part can be an advantage. Although I have done a fair bit of rework, for awkward PCB's, where you don't have good access, I tend to sacrifice either the chip, or the board. The former is fine for repairs, the latter for salvaging.

And yes, dipping your solder wick (or just thin copper shielding from audio cable) before use makes it much more effective for cleaning up pads afterwards - or for re-tinning corroded pads after cleaning. I haven't had continuous access to the best equipment, but my $40 chinese hot air station seems to work just fine, even allowing me to rework smart phones, and PC graphics cards (where many of the parts have great thermal conduction to the board).
 
Well, again, the sink is primarily what I am stumped on. Since I had to grind out the center to get better access to the pad, I would get to do that with the new one. One option I discussed with my bro is silver paste rather than solder? I realize this would not be a true bond but would not require heat. However, in most cases where that is used, there is a mechanical means to keep contact pressure.

I would not expect the board to operate anything over 100C normally.


I might like to get a few more thoughts to recap ideas for soldering the pins. I originally considered just lightly tinning each pad, then sit the chip on after fluxing the feet, and just touch each foot with a little heat to sit them down. However, I can see where that might present a few issues. Once I have attached the heat pad, I don't want to have to remove the chip again due to bridges. Remember too that I have tons of tiny components very close to the feet of the chip so too much heat will probably cause issues.
 
There are lots of unknowns, so that leaves lots of possibilities.

So far mentioned:
1) Thermal conductive epoxy. LocTite also makes something similar. I used the LocTite product (McMaster Carr) for some high-power mosfets. I also just soldered the same to copper bus bars (0.625" x 0.025" x >6" long) with no problems.
2) Low-temperature solder. Bismuth alloys are available with with melting points from 58°C (ChipQuik removal alloy) to 138°C (Sn42/Bi57/Ag1), as well as other temperatures.
3) A paste, either silver or diamond.

With any of the solder options, you could rig up a low-voltage source and simply heat the tab as a resistance element.

I would tend to go with the low-temp solder and flux, since you already know you can pre-heat the whole board to 150°C. I would then use regular tin/lead solder for the pins and not fuss with resistance heating of the tab.

John
 
I probably won't use the epoxy just due to the no rework problems. I don't plan to do this again but I also am not yet sure how the chip failed in the first place.

The paste I was considering is not a solder paste but just a conductive paste that is common to components such as the TO-220 package. I am not sure without a mechanical pressured connection, if that is wise.

I am leaning towards a lower temp solder for the heat pad. Can you recommend something? I think most solders today are around the 180C melting point. I would like to see something around the 125C range?

For the pins, I intened to just use my Kelser #44 solder which is rosin core, 60/40, .20" just because I have used it for yrs and pretty familiar with its performance. I am not yet sure of the pros/cons of the rosin core but I know when commonly tinning stranded wires, non flux solder does not seem to work nearly as nice as this stuff. I also have 62/36 with 2% Silver that I bought specifically for audio electronics. That is the one that did not seem to like tinning stranded wires as much but is supposed to be targeting SMD components.
 
Last edited:
Indium Corp has a lot of information about low-temp alloys. Here is a link to its bismuth alloys: https://www.indium.com/solders/bismuth/

There are lower melting alloys, such as Wood's metal. I would not recommend experimenting with various alloys. Some alloys are quite brittle (e.g., Wood's). Also, alloys with indium will attack gold and may make unstable joints. You also want an alloy with good wetting properties. That's why I would stay with a commercial product. Kester used to sell a low-temp bismuth solder. I have a spool, but couldn't find a link to it after a little searching. The spool is 42 miles away from me right now. I won't be back to that location for several days. As I recall, it was about 17% Bi.

A search for bismuth solders will give you several sources. Mouser and DigiKey also carry common brands.

I didn't mention before, but I would not try doing this with hot air. Hot air can heat a large area and can be quite hot, but it is still subject to the relatively poor thermal conductivity of air. Thus, when trying to heat something on a heatsink, it can be defeated. A nice soldering iron with a bit of molten solder will give better localized heat. Get yourself an 80W with chisel tip. Mine is ancient and rarely used, but when needed, it's the right tool. I run mine from a Variac so I can keep the tip from overly oxidizing.

John
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top