Desoldering electrolytics with bent leads on old boards?

[mmagin]

I was beginning to re-cap an old Keithley 616, but I'm close to giving up in frustration.

The existing electrolytics are flush to the board, so you can't clip the leads and do one lead at a time.  Also, the original assemblers seem to have bent the leads to make them hold in place before soldering.  Any attempt to heat and unbend the leads, or just heating and tug on the capacitor usually leads to some of the pads delaminating from the board by the time the capactor comes out.  I thought that was just the domain of cheap old consumer electronics phenolic boards, but it seems to be true for this too.

I'm using a Metcal STTC-036 tip, so I'm certainly not using too much heat, and I've been liberal with 63/37 solder and rosin flux (and alternately tried removing excess solder with solder wick), but I cannot seem to find a technique that doesn't involve a fairly high chance of screwing things up in the process.

[DTJ]

It's a PITA.

>Solder wick most of the solder away.

>Heat the joint and use a blade to stand the lead up vertical.

>Use solderwick to clean around the hole

>Remove the part (hopefully)

[MarvinTheMartian]

You could try a solder sucker instead of solder wick to try to get most of the solder out from around the leads, then heat lead and bend straight if necessary before removal. 

[RobertHolcombe]

You could cut the component leads off flush with the board which may require oblique flush cutters if clearance is an issue, then heat both pads simultaneously with the board upside down and let gravity do the rest.

[vk6zgo]

Initially,try to desolder only the bit where the lead comes through the hole on the copper side of the PCB.
Using sidecutters,cut at the point where it just comes through,leaving the rest of the bent part soldered.
You can then remove the capacitor.

Now,gently desolder the cut off ends---you may have to do it in sections.

[Ian.M]

ChipQuik low melting point alloy.  Wick off as much old solder as you can quickly, add ChipQuik alloy and the join will stay molten to well below 100 deg C.  If you can keep the board preheated to around 80-90 deg C, the joint will stay liquid and you will have as much working time as you need to coax the leads straight with a stainless or oiled steel scalpel blade or V scraper.

I wouldn't try cutting the lead with side-cutters where it emerges through the pad, there is too much chance of the shock and/or stress loosening the pad to substrate bond that has already been heated considerably to desolder the joint enough to expose the lead. 

For radial caps, its possible to cook the cap body till the seal softens, pull the can off, cut the leads flush with the end-plug, pull the end-plug off the lead stubs, cut them flush while maintaining tension oand then desolder individually, but its a right messy business that will require thorough cleanup of the electrolyte afterwards, and should only be done as a last resort.  Not all cap seals let go easily enough to make it viable.

[vk6zgo]

Ian,I've done it plenty of times----the secret is to be gentle.
Maybe even just "nibble" the lead rather than a straight "snip",if you are really concerned about the pad to substrate bond..

[Ian.M]

Maybe I've worked on too many really crappy boards with barely enough pad round the hole and hair thin tracks? 

Some of them, the pads would lift if you merely looked at them cross-eyed, which made changing DIP ICs extremely challenging.   Even Dremmeling the legs off was pretty dicey as if the cutoff wheel snagged the leg, it usually lifted the pad, so one had to cut into the body just enough to steady the leg while still letting it be pulled free when you heated the joint.

[vk6zgo]

Maybe I've worked on too many really crappy boards with barely enough pad round the hole and hair thin tracks? 

Some of them, the pads would lift if you merely looked at the cross-eyed, which made changing DIP ICs extremely challenging.   Even Dremmeling the legs off was pretty dicey as if the cutoff wheel snagged the leg, it usually lifted the pad, so one had to cut into the body just enough to steady the leg while still letting it be pulled free when you heated the joint.

I must admit I have ran into some like that.

Barco & Bosch  were the worst offenders.
Their boards looked beautiful,compared to,say,Sony,but they were sods to work on.

For some unknown reason,they often made the holes so tight that component leads were an interference fit.
That made it easy to wreck the hole (especially plated through ones).

Sony used good sized holes & generous pads,allowing for easy replacement of parts,in both their Studio & Domestic stuff,

NEC were also  mostly good that way,although in one RF board they had a relay with about 20 pins,15 of which were to secure it to the board,& connect to the ground plane.
Braid didn't work,a Pace Desoldering Station didn't work.

Eventually,we found there was enough space to run a "Stanley knife" blade between the relay & the PCB,cutting off all the pins.
With the relay removed,it was easy to remove the pins.
It wouldn't have been possible if we had needed to salvage the relay!

[mmagin]

Thanks for all the advice.  Yeah, in this case, Keithley saw fit to use tiny little traces and pads IN THE POWER SUPPLY.  Admittedly it's low current stuff, but it seems just ridiculous.

I'm going to definitely have to get some of the (leaded!) ChipQuik stuff before my next adventure.

[SeanB]

If it is a plain electrolytic simply use a big set of side cutters, or large end nippers, and cut the capacitor in half just above the bung. That leaves a small aluminium ring and a rubber bung with 2 legs, which then you remove the ring and bung, and can then do each leg as a separate item.  Otherwise wimply wriggle it until the leg fatigues and breaks off, or you can pull the aluminium sleeve out of the bung, leaving either a broken off lead flush with the board which you heat with a soldering iron and push through with a stainless steel needle, or can desolder like the other method.

Fatal to the capacitor, and you will get electrolyte on the board so will need to wash it well with cold deionised water at least twice before putting the new caps in, but does not damage fragile traces and vias.  Works well in those 1970's made GRP boards which tend to peel traces at the slightest application of a soldering iron.