Help, Can't Desolder Capacitor Even With 2 Irons

[DROBNJAK]

I need to remove few electrolytic capacitors that leaked, but can't desolder them.

I am using 2 soldering irons at once. One is Hakko FX-888 65w, set to 450 ºC and other one is 'sucking' iron 30W. I get both of them to touch the connection for maximum heat transfer, yet solder stays solid. I placed plenty of flux on the junction and yet I can't even start to melt the old solder around the cap's leg.

I do not want to try to drill capacitor's legs out for fear of rendering the PCB hole disconnected from a circuit. PCB in question is a multilayer PC motherboard and all connections are rally tight, with minimum of a solder applied.

Any ideas?

[eliocor]

put/melt some MORE solder on the pcb/contacts of the capacitor: this will help to "freshen" the solder point.
Once done, it will be easier to remove/clean the component from the PCB.

[peter.mitchell]

Get an iron,crank it all the way up, put some solder on the tip, then add a bunch of solder to the joints, then use a pair of pliers to pull the component gently and rock it as you heat up the pins. More solder = good for desoldering, not enough solder makes it hard to keep the solder molten, and makes it hard to get good thermal contact.

[jeroen74]

Try preheating the board to 100C or so. There are probably a solid ground plane on a layer or two that suck away the heat.

[DROBNJAK]

thanks

[TerraHertz]

Second that on preheat board due to groundplane sucking heat away.

Another possibility is that you have a motherboard that uses ROHS solder, aka-lead-free solder, aka *&@#$!COME ON, MELT YOU MOTHER**CKER!
It has a much higher melting point, that's actually below ABOVE the temp most regulated irons are set at.

If the ROHS movement had been intentionally planned to make circuit boards unrepairable, and put small local board assemblers out of business, it could not have done a better job.

[DROBNJAK]

It does look something like high temp. solder. I have iron set to 450 deg C, and it won't budge. And yes, there is a big ground plane.

[T4P]

I personally find ROHS solder a pain in the butt in every single way
1) Tips designed for lead-free still lasts shorter compared to tips for lead solder only due to having to crank the temp up (more ions stripped still)
2) Because the melting temp is high, it sets very quickly like you have no time to look away
3) You won't really know whether it's a dry joint or not
4) You need lots of thermal horsepower, something like a QUICK 201 or 303B (120W) or a Hakko FX-951 or any JBC stations (Metcal too)
5) More aggressive flux, the only things that will boil in solder is flux and more aggresive flux means having to clean thoroughly as well as requiring better fume extraction. Why more aggressive flux? Tin of such purity in most lead-free (unless it's a Sn42Bi58) have problems adhering to copper surfaces so again more copper ions stripped from the tip

[WorldPowerLabs]

I've had good luck with Chip Quik when removing stubborn parts from boards I really don't want to damage (I usually don't bother using it because it's a bit expensive, but if the board is valuable it's worth it).  Use lots of flux, and melt their special solder into the joints.  That leaves you with an alloy of metals that melts at a very, very low temperature.  The part will then come free with much less effort.  Clean the special solder out of the joints, remove excess flux, and install a new part.  (see http://www.chipquik.com/)

I have replaced lots of leaked electrolytics, and what I've noticed is that once the electrolyte gets onto the solder joints and corrodes them, it takes a bit of effort to re-melt that damaged solder and remove the part.

[rsjsouza]

I used to have trouble with some stubborn components and their large pads, until I got one of these for me: a Weller 8200PK.



Obviously you need to be extra careful in not ripping everything else from the board with overheat, but it works like a charm.

[TerraHertz]

ROHS solder is more than a pain in the butt to PCB assemblers, it's a serious threat to their business. I was told why by a manufacturer:
It costs several thousand dollars to charge a wave soldering machine with lead free solder, due to the high cost of the solder.
To maintain their ROHS certification, board manufacturers have to have their solder baths regularly tested (expensively) for purity.
But here's the killer. If they run a loaded PCB through the wave solderer and it turns out the pre-tinning on the PCB or components was lead-based solder, their entire bath of ROHS solder is contaminated, won't pass purity test, and has to be replaced.

Now, how would you like to be the one responsible for inspecting all incoming components and bare pcbs, to ensure they don't contain any lead? Knowing one mistake was going to cost many thousands, and a few mistakes could kill your business. How would you do it? It's a near-impossible task.

Naturally, Asian board assemblers don't have to deal with this shit or even care about it, and hence operate much more profitably.

[AlphZeta]

Like others suggested,  add some solder on the tip to help heat transfer. Also not sure what kind of iron tip you have, but for this kind of job you want to use the larger chisel tip.

[TerraHertz]

Here's an idea. Something I did to remove a heavy connector with pins in-line from a PCB.
My soldering irons and sucker couldn't budge it.
So I found a bit of old copper bus-bar, hack sawed a bit off, put a bevel on it with a radial saw (but a file would have done the job too.) Now it was a suitable shape to heat all the pins at once.
Added a 'handle/stand' to my old style soldering iron. The vertical rods are stainless steel for low heat conductivity. They are just hammered into press-fit holes drilled in the copper. Ditto with the bit of wood.

Heated it up with a blow torch while tinning the working end (using metalworking soldering fluid and ordinary solder), added a fresh solder blob to each pin of the connector so they'd all form a good heat transfer contact with my new 'iron', and presto! It worked like a charm.

You just have to guess the temp of the copper block, by the way it and the solder on it are oxidizing. You can also get an idea from the way the solder is running on the copper block. Don't want to overheat it or you'll ruin the PCB.

[SeanB]

I used a hot air gun to gently warm up the whole back area around after applying a new solder coat to the pins, then when it was almost at melting point I took the gun away, put the hot soldering iron on the leads and then pulled the cap out quickly, which left each hole filled partially with solder. 0.8mm stainless steel wire then poked through each hole after adding fresh solder and left to cool, then pull wire out to leave a clean hole. 316 SS wire does not solder easily.

[ptricks]

Something else you can run into with capacitors and sometimes transformers is leads that are installed using through hole rivets. Rivet connections can be very difficult to repair because the rivet holds the part along with the solder. We used a tool that can cut a capacitor cleanly in half which basically allows you to remove the capacitor leads from the top of the board, clipping them away until nothing was left and heating the connection allowed the excess lead to be removed.  I haven't seen one of those tools sold anywhere, it was custom designed , basically a round device that slid over the top of a capacitor and with the turning of a thumbscrew on the top would cleanly cut off a capacitor about 5 mm above the board.

If a ground plane is removing heat faster than you can apply it then pull out the dremel or other multi-tool and trim the copper around the pins. After replacing the part use solder to re-join the cut away sections of the plane.  If you are heating up an area 2 to 3x that of the part you are using the wrong technique.

I see a lot of board damage from people using way too much heat for too long.  The rule is do not apply  heat to an area larger than what would be needed to solder the part onto the board.

[TerraHertz]

The trouble with 'isolating the ground plane' is that on many boards the ground plane is an internal layer. A good PCB designer will include thermal reliefs (having the through hole copper only connected to the ground plane by short, narrow traces) but careless/novice designers will just have the internal ground plane butting right up to the through hole.
This makes desoldering almost impossible, since you're trying to get enough heat conducted via the via, to still melt solder at the point where it contacts a very effective heat sink.

So you just have to apply a LOT of heat fast.

I've never had to deal with desoldering a contiguous inner layer ground plane *and* ROHS solder. I think it would be impossible. Better to just drill out the leads then use copper foil or something to improvise new through hole conductors if needed.

[Fraser]

I have seen official modifications where a through hole component is cut off of the PCB and a new component soldered to the old lead stubs. I wasn't impressed at the time but I can now see why such action may be necessary. The PCB's were very expensive multilayer types, and butt jointing is still a reasonable connection if done well.

Fraser

[jeroen74]

I wonder whether the OP has successfully removed the cap by now

[SeanB]

We did that, mostly because the pins were impossible to get to. Take a board with CERDIP all on one side with a 0.1 spacing, then invert it and place more CERDIP's on the other side in those 2 rows left. Makes a board that fits in your hand, but is over 2kg in mass. If a chip is faulty ( and you better make sure, like the top has popped or the outputs are totally wrong even with static inputs) then cut the leads off at the case and take the new one and cut the leadframe flush with the bottom of the body. Tin both sides and place in position then touch each pin join with a hot soldering iron to reflow. Then test. BTW these were 16 layer boards and were really fragile.

[Fraser]

SeanB,

That is really interesting. I have not seen that done with IC's , but I can think of several occasions when such a technique would have been better than the ripped out through-holes resulting from interference fit pins and through hole sleeves ! Thanks for sharing. I will file that away in my mind for those awful situations where an IC removal represents an unacceptable risk to the PCB. Another manufacturer trick that I have suffered for is the stacking of one IC on top of another. In my case an Ceramic DIL EPROM soldered over the top of another Eprom, with certain pins splayed away for chip select etc. A very messy job when you want to remove the lower EPROM, especially if they have managed to force the upper EPROM pins into the same holes as the lower EPROM pins !

Fraser

[SeanB]

So you have suffered at the hands of French designers then........... 400 tools on the list, and all marked "Tool, Special". This included a spanner with a bend to get at a certain bolt, a hex key with an extra long shaft and a screwdriver with a narrow tip. One set was a whole pile of stainless steel plates, with board sockets bolted to them, used to align the pin sockets on the backplane. Not that fixing a pin on the backplane was easy, all were wire wrap sockets, and all were wired in white. Luckily you could brute force the pin out without breaking the socket, and break a shell to get a pin as a replacement. No way was I going to unwrap 100 pins after unsoldering them then rewrap them again. Luckily I only replaced one..... The test board was worse, they must have looked hard to find such unreliable a socket as they could, and then used it for everything, including resistors and supply decoupling capacitors. In the end I soldered most of the IC's to the sockets to keep from having intermittent faults. Only one that got a new socket soldered to the old one was a fusible link prom that was hard to program, not having any other method than using a breadboard and wire links along with filament displays to read the data afterwards. We had plenty of blank ones though........ and more than plenty of dead ones.