1
\$\begingroup\$

I have a graphics card with a dead cooler. I have a replacement cooler for it, but the graphics card is a bit non-standard and the cooler won't fit --- the heat pipes hit some components on the board.

As said components are a big, chunky capacitor and a two-pin inductor, all connected via wires through the board, it should be easy enough to move them to the other side of the board to make room for the cooler.

Unfortunately I seem to be totally unable to make the solder on the board melt.

I have a 70W gas soldering iron. It melts my solder fine. I've heated the joints up as much as I dare and they don't even show signs of softening. If I add some of my own solder to the joint, I can then melt that, but the original joint stays solid.

Is there such a thing as high-temperature solder? Would they have been likely to have used it on a graphics card? I'm aware that this kind of PCB is likely to have a thick copper ground plane, which will suck away heat from the joint, but that doesn't explain how I can melt my solder but not their solder in the same joint.

The soldering iron does have a hot-air-gun mode, but I have no experience of that and would rather not set fire to my graphics card...

(I want to get the components off intact because I need to attach them again. The capacitor is easily replaceable, but the inductor is just a black cube with 1R0 written on it, and I'm not confident of my ability to find an exact duplicate. If I cut the pins it will then be too short to reattach.)

\$\endgroup\$
4
\$\begingroup\$

I'm suspicious that the ground and power planes that your capacitor is connected to simply have a lot of thermal conductivity to ground. I've desoldered caps from graphics cards before, and my 95-W temperature-controlled iron, set on 400°C/750°F takes quite a long time to heat it up. I'm not sure that a gas-powered iron is up to the job of graphics card rework.

One thing you should probably do is to preheat your graphics card before attempting rework. Warm the whole thing up to about 300°F with a hot-air gun or oven, and then it should be much easier to get the point that you're trying to melt hotter than that. When the rest of the card is at room temp, you're likely to put a lot of thermal stress on the board.

If your solder is melting, and actually mixing with the solder in the component, then the melting point of the mixture will be approximately the average of the relative volumes in the mixture. This article explains how preheating can be used with a low-temperature solder to do rework. I would guess that you simply had a ball of your solder on the solder joint, and didn't actually bring the joint to the point where your solder would melt if the two were mixed together.

As a sidenote, you should probably replace the capacitor when you remove it. Electrolytic capacitors can be damaged by heat, and are the components most likely to go bad on your card. Inductors are likely to be among the most robust. This brings me to another point: The cap is built with two pins stuck into a roll of paper-like substance, and stuffed into the can. You can pull the can and paper off of the pins without too much effort, but this will destroy the cap. This does have the advantage of allowing you to work from both sides and on one pin at a time, which is likely to be much easier.

When resoldering, remember that you'll have to get the solder joints just as hot as when you removed the components, or your solder will not form a good joint. Don't settle for balls of solder on the component leads; you need a nice conical joint.

\$\endgroup\$
  • \$\begingroup\$ The idea of heating the entire board up makes me extremely uneasy: it's well out of my comfort zone when it comes to soldering (plus I don't have the tools to do it). My solder is not mixing with the board's native solder. It's melting, and forming a layer on top of the original joint, and then I can let it cool again, heat it up only my solder melts; the original joint is untouched. Which is what's confusing me. Is the hot air attachment of the gas iron likely to be more effective at heating the relevant bit of board than the soldering tip? \$\endgroup\$ – David Given Aug 12 '11 at 21:20
  • \$\begingroup\$ @David - The problem is that the 'relevant bit of board' is thermally connected to large heat sinks. Whatever gets you the most heat is what you'll need to do. This may mean borrowing a better iron. \$\endgroup\$ – Kevin Vermeer Aug 13 '11 at 16:57
  • \$\begingroup\$ But I can add solder to the joint and then melt that solder, but not the solder in the original joint. This implies that my soldering iron is perfectly capable of getting the joint to temperature, even with the heat sinking effect; at 70W it's not exactly small. Why, then, is the board's own solder not melting? I was under the impression that wave soldering used low temperature solder, not high... \$\endgroup\$ – David Given Aug 13 '11 at 17:56
4
\$\begingroup\$

As part of your problem (along with ground and power planes sucking heat away from the joint), your solder may well be lead based, and the solder on the graphics card is almost certainly lead-free (everything in the last few years has moved to lead-free solders). The lead-free solders require significantly higher temps to melt than older lead-based solders.

Unfortunately, I think the basic problem is that your iron is just not up to the task. The pre-heat trick might help, but I'd recommend just getting an iron that's got more punch.

\$\endgroup\$
4
\$\begingroup\$

Three suggestions I have for you:

  1. Get a bigger tip (if you iron supports interchangeable tips).
  2. Take some solder and melt it on the tip of your iron before you touch the part's pin. Sometimes it takes a little melted solder to get it all flowing
  3. Solder wick - Pick some up some and use that to wick up whatever you melt that way you don't have to regain all the ground you spent reheating this thing each time your hand gets tired and have to remove the iron. Get a smaller wick so that you're not fighting heating up that too along with the pads. Sometimes you can wick off enough solder that a small xacto blade can be used to score through the remaining solder.

It sounds like it's just thermally connected to GND which is more than likely a huge heat-sink

EDIT

If you can't get a bigger tip, make sure you're turning the tip you have sideways to maximize the surface area that comes in contact with the solder-joint. I see too many people think that by just putting a needle point tip onto the solder joint it should melt without actually taking into account what is going on thermally.

\$\endgroup\$
  • 2
    \$\begingroup\$ +1 for #2: Pb-containing solder dilutes the Pb-free one, and the latter melts on a much lower temperature. \$\endgroup\$ – whitequark Aug 12 '11 at 18:56
  • \$\begingroup\$ In addition to diluting the lead-free solder and lowering the melting point, a bit of solder on your tip will increase the thermal conductivity of your iron from one point to the whole area that the liquid solder is contacting. Liquid solder is molten metal, and conducts heat much, much better than air. \$\endgroup\$ – Kevin Vermeer Aug 12 '11 at 19:05
  • \$\begingroup\$ Also, I have only ever used the xacto knife trick on SMD parts. Inductors and big capacitors are likely to be through-hole parts, so that won't help. \$\endgroup\$ – Kevin Vermeer Aug 12 '11 at 19:07
  • 1
    \$\begingroup\$ I've tried #2, and #3 is currently irrelevant because I'm not melting any of the board's native solder. I can add my own solder to the joint, and then that melts and I can remove it again, but the native solder is untouched (and unscratched). This is what makes me think there's something funny about their solder. \$\endgroup\$ – David Given Aug 12 '11 at 21:10
  • \$\begingroup\$ Fair enough. Then I would suggest a bigger tip/bigger (hotter) iron. \$\endgroup\$ – Joel B Aug 14 '11 at 17:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.