I hope I did these photos right. I’m on mobile.
What did I do wrong?
Here’s my story, briefly. I decide I want to build a new computer. I start out thinking I just need something newer than my current 14 yr old laptop to do computer science homework on. Over time, researching, I do what I do and decide it’s got to be the best. Spend shitloads of money. But I’m not a millionaire. Or even a thousandaire at the moment. (Back in school). So I hunt for parts on eBay. Find some good deals. Ok.
I get a video card. Great deal. Should have been suspicious. It mostly works. After literally hours of troubleshooting I figure out that the computer repeatedly crashes because the GPU. It’s after about a new power supply and replace motor parts. Card mostly worked, so It was tough to diagnose. I make arrangements to sell it back, I mean send it back to the seller but then out of curiosity and desperation to if I take the top off and this random part is clearly out of place- 0.33uH power choke.
https://goo.gl/images/PsKEzy
https://goo.gl/images/bH26ps
That one. Marked r33. Now I have soldering gear, a fairly Full set up but it should be noted I’m terrible at shouldering. 90% a time I get cold solder joints. I can’t seem to figure out what I’m doing wrong. Anyway, I take my soldering gun, and solder the piece back down. Mine didn’t explode like the pictures it was just half hanging off. Apparently it’s a common problem and the MOSFET next to it actually is typically fine for whatever reason.
I sold her back down and beat it up and run the car through a benchmark/stress test and it worked perfectly for the first time ever. I do a little jig. Went to double check and read it a second time this time it crashed like before. The next four times it crashes again. So this point I know I had fixed it but then it unfixed. I take the piece of 20 double check and read it a second time this time it crashed like before. The next four times it crashes again. So this point I know I had fixed it but then it unfixed. I take the card apart again and that damn little piece just straight up falls out when I take the top off. OK. Like I said my soldering sucks.
So! Then I think o need to reflow it. I order some paste. I’ve never done it before. But I’ve seen it done. And I have a heat gun. The real kind for this purpose. No oven but it’s a focal problem.
I get two kinds. Low temp silver bearing lead free (~165c) and regular tin lead. I can get the exact types of needed.
So I practice on some other boards. It goes swimmingly. I am thinking to myself solder paste where have you been all my life. I make several practice joints, solder some SMT stuff down on protoboards things like that.
Then I turn to the real piece. At this point I know that the component works despite my abuse with the soldering gun the day prior. Importantly I had a very hard time getting solder to melt. Even on my max temp of 480. There is a lot of copper/ heat management underneath this piece I think. So with the gun I tried for a few hours before deciding smt reflow was the way to go.
I had done it perfectly on the protoboards no problem. I decided the low temp stuff melted too fast- and because I was worried that it was heat that destroyed this component in the first place - I decided on the high temp.
—process starts here—
Put it down put the component down, get the small heat gun tip, and go to it. This. Solder. Will. Not. Melt.
I have the heat on it- for probably way too long. I’ve tried to shield the surrounding bits - but the whole card- several inches away is getting hot.
I cut my losses. Scrape up the now dry powder crud. Dry and not melted. I try again with less. Same thing. Ok so I can’t get it hot enough. Try again this time with lots of flux and the solder paste -which I think has its own.
Now I try the low temp stuff. This stuff transitioned almost too fast before. At like 160C. I know operating temp of this card can be like 100C so I didn’t want to use it. But I did.
It. Will. Not. Melt. Again. Heat gun at recommended temp. Then up to 365C. +100. I finally, FINALLY get it to transition and join buy holding the heat gun over in AND touching the solder iron at 480c (set, actually cooler unless you hit it right at peak cycle). But yea it’s a LOT of heat. And it melts and joins.
That’s all well and good, but resistance is now 0 across the part. Should be low- couple ohms , but 0 is bad.
I think maybe solder bridge between pads. Take part off. Nothing. Looks good. Try the part. It’s in continuity. Pretty sure I’ve killed it. I’m not surprised.
I measure some of the nearby inductors, R22 in the pics. They’re now in continuity too. Oops. I think I damaged several with heat.
I can still clean up my mess and send it back to the guy. I have 0 guilt- because I should mention it was sold to me as perfectly working, always well cared for etc. and I should mention when I took it apart it looked like this part that was hanging and been previously attempted to be repaired. So he lied to me in the first place. If I was able to fix it I would still have negotiated some money back.
Anyway- a few questions for the tech gurus out there. What did I do wrong?? It should have had more than enough heat, even the low temp stuff wouldn’t melt. It melts fine on the protoboards as soon as you get a foot away. I was right on top of it.
I know it likely has to do with the thermal management of the board, the heat was being spread quickly elsewhere I guess. Should I have used paste and a tip instead of air?
Can this card be fixed?
I’m not sure what’s been destroyed. Certainly some of those inductors. I have some of the r33 on the way, but not the r22. I can get some but it means waiting another weak for mouser. I don’t think tanner has these SMT parts. Maybe the space has some or one of you guys.
Do you think more than r22 and r33 are bad now- some of the surrounding parts? I ordered some more of the mosfets they were cheap. But I will likely
Have sent it back before it gets here.
How can I test some of these other components (with only a basic multimeter, a resistor, or capacitor, a paper clip and bubble gum?)
I put the card in and it’s more dead than before, confirming suspicions. No image at all now. I might have killed it by running it- as I know that inductor chokes the power to the DRAM and could have cooked it.
This was supposed to be an easy repair! Just a little heat gun rejoin the parts and rock and roll. And somehow, which seems like it always happens when I get a soldering gun- things have ended up so much worse.
I didn’t want to try and oven it, because a, it was a focal problem. And b, I don’t have one. Sorry this is too long.
at Issue, was just unexpected behavior of the paste. I still need to figure out if the temp readout on the station is bad. Or if it really was the ground plane. I need a third temp device I trust. That paste still behaves weird sometimes, so maybe my technique is still off. But I can handle that too, thanks. I don’t do this for a living, no. But my desire is to learn how to do it. Not just do it. Money is an issue in that I’m frugal. Not in the sense that if I blow $300 on a video card I won’t make rent. I bought an $800 one while I was sorting this out (to bring my score up- see below).
). And for me, it’s almost always on talk. By the time i start doing, I’m neurotically research enough that I’m pretty confident in what I’m doing.