Thinking outside the box. Disassemble capacitor without removing the leads?

[gregr]

I searched the FAQ. I read all kinds of warnings about overheating a motherboard with the soldering iron, mechanical desoldering devices, warning this, warning that, etc, etc, etc.

If (and that's a big IF ) it was safe to disassemble the capacitor and leave just the leads poking out the top of the motherboard, why not just parallel solder the new capacitor to the old leads. If I remember my electrical engineering 101 class in college, the length of the lead shouldn't change the properties of the capacitor. Right?

Tell me where I'm going wrong in this line of thinking...

Greg

[PCBONEZ]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

Length of leads affects ESR big time.

To test that I measured ESR with the test leads right at the bung and then again with the test leads ~ 3/4" from the bung.

ESR about doubled.

Of course depending on what caps you try this with your mileage may vary.
I was using an MBZ if I remember correctly.
.

[gregr]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

PCBONEZ, that is amazing that you've tested that. Brilliant.

Ok... That begs the question, wouldn't it then be easy to screw up how tight to pull the new capacitor to the board when you are recapping? I'm facing a recapping job on the wife's computer, and failure means many more years of pain (vs. just getting fired from my job if it were at work). I've done a lot of small PCB work, but never a motherboard.

If each lead DID have a mark on it indicating "pull this lead level with the PCB to here", how many people know what that mark really means? Based on your test results, the length plays a MUCH bigger role than I would have EVER thought.

[PCBONEZ]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

Most caps sit all the way down on the board. The bung on good brands usually has a protrusion that bottoms on the board before the can does anyway. That's more to allow it to dry under there after a cleaning than anything else though.
Tyan is the only brand I know that seems to consistently space caps off the board. I'm not sure if space is underneath is their goal or it's just a result of their process. If you ask me it just makes their boards look sloppy because caps end up leaning every which way.

There should be no pulling force on the leads when you solder them in. Some people bend the the leads against the hole to get them to stay put for soldering. I do that sometimes too but usually I use an assortment of small blocks and random items [coins, bottle caps, sockets] and sort of stack the block, then the cap, then the mobo insofar as weight distribution goes. [That's functionally backwards. I flip the board solder side up, prop the board, then slide the blocks in underneath the cap as needed. Depending on the layout I can sometimes prop 6-8 caps at a time that way.] Some folks will use tape on top the caps or even glue under them. I'll pass on the glue method but other than that whatever works.

If you know how to solder already the only thing that might take some getting used to is how thick the boards are. They'll suck the heat right out of the tip and if you don't have enough watts behind it to reheat it quickly the result is you slowly heat the whole area of the board instead of quickly heating just the joint.
You want to use the fact that PCB material transfers heat slower than the materials in the joint.
Think like: lots of heat, on and off it quick.
Too long a time due to not enough heat in the iron allows the heat transfer rate of the PCB material to catch and keep up with the joint's temp before the joint's temp is high enough to melt solder.

Without using the fancy air sucking solder gear what works best for me is:
a 60 watt iron + rosined solder wick for 'off', use the wick before you pull the cap, then alternately heat the leads and 'walk' the cap out, then use wick again if needed, then clear the hole [if need be] with an appropriate sized sewing needle [needle on one side, iron on the other], then I usually use a 40 watt iron and .022" diameter or smaller solder for soldering the new ones in.

Mucho big deal!! - Make a cap map BEFORE you pull any caps.
Make sure you mark the polarity in addition to uF and volts.
Some board silk-screens have errors so if you don't do this before you pull caps you can end up with a cap in backwards and it will take forever to figure it out.

~
Lead length might not matter if ESR isn't a concern.
ESR is kind of like Capacitive Reactance only if you plot both on a curve against frequency they mirror each other's shape.
When ESR is high Xc is low, and vice versa.
On motherboards the frequency of concern is <~200 kHz and is due to the SMPS's switching action. Down there ESR is the biggie.

.

[ratdude747]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

just desolder the dumb thing... ive never overheated one... didnt pentium have a thread about his friend doing that on a gigabit switch and how it didn't work.

[gregr]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

ratdude,

Thanks for the perspective. I'm definitely going to desolder the dumb thing and recap it. I'd rather throw it out the window across the yard. I'd feel a whole lot better, but it for sure wouldn't post after that.

I'll post the results.

BTW, it's an ASUS T2-R Barebones combo, purchased about 3-4 years ago. It uses nichicon HD(M) 1500uF 6.3V 20mm tall capacitors. Here's the obligatory pre-surgery pic:

.

[Toasty]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

>>purchased about 3-4 years ago<<
It was likely 2-3 years old by then, which puts it squarely in the midst of the crap cap production era.

Manual was out in late 2003, last update to the BIOS file was 2005.

I presume they're 10mm diameter caps, so get the Rubycon MCZ series from badcaps.net:

https://www.badcaps.net/store/produc...products_id=40

Toast

[gregr]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

Ok... Time flies when you're having fun (and kids). Maybe I did buy it in 2004... Thanks for the link, Toast. They are 10mm diameter, but 20mm tall, not 13mm. The 20mm Rubycons are the ZL series. I get the feeling from reading the boards that the MCZ series is the better choice, even if they are only 13mm tall. I have plenty of vertical clearance so it really doesn't matter to me.

[PCBONEZ]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

Are you sure the ones in Vcore aren't HM(M) ?
.

[gregr]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

Here's a pic. They are definitely HD(M)'s. They are 8mm in diameter. As you can see by the date code, they were in the 2003 timeframe.

[kc8adu]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

hd are not low enough esr to last long term in a p4/a64 vrm.
i have seen them before blown.cant lay the blame on the caps as they are not the right choice for this app.esp seeing this is a sff that can use a 3ghz prescott.

[gregr]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

That's really unfortunate for my opinion of ASUS design. For years I have trusted ASUS for my personal builds. So, I picked them for the wife's board so there would never be a complaint (you know the cobbler's children story). It would be easier to forgive them if they had a good electrical design and just happened to be victim of a bad supplier.

[PCBONEZ]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

Asus was better in the '90's.
After ~2000/2001 they quickly eroded to average and stayed there.
.
Even though they've probably had over 100 affected models, Asus has STILL never even admitted there was a problem with bum electrolyte or to the HN/HM problem.
- Deliberately ignoring the existence of a serious problem is the same as lying to their customers. PARTICULARLY when they ran customers with models with known caps issues through extra whoops to minimize RMA's.
- Asus turned into: "You bought it, you got it, now go away."
.

[berniedd]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

Gregr, that's exactly what I did to my Asus P5V800-MX motherboard. It would often hang and I quickly traced it to be many bulging 820 uf/6.3 v caps. They were all over the board. I've recapped many many boards over the past few years, but this is the first time I've had problems melting the solder on the leads. Then I noticed that many of these caps were installed in a slightly raised position, not flush with the board. So I just rocked the caps side to side till they snapped off on the leads, and I had 2 mm or so of bare lead sticking out of the board. To these I soldered my replacement caps. I ended up with a total lead length of about 8 mm between each cap and the board. A total of 9 capacitors were replaced.

I did it this way only because I couldn't desolder the old caps well enough, for some reason I can't identify. I cheated a little, because I used 1000 uf as replacements. The board is now back in service, and performs flawlessly.

To give you a glimpse back in the old days (1960's and 1970's), replacing resistors and capacitors on printed circuit boards routinely meant crushing the body of the old part with cutting pliers and soldering the new part on the leads left after removing the remnants of the old part.

[PCBONEZ]

Re: Thinking outside the box. Disassemble capacitor without removing the leads?

That's okay as long as you realize the when you mount caps with long leads like that you raise the ESR, so you need caps with even lower ESR to compensate.

To test that I measured some 16v 1000uF MBZ right at the bung and then again about 3/4" from the bung.
ESR went up 0.01 ohms from 0.02 to 0.03 and it was consistent.
[I wish I had a 3 decimal point ESR meter, but I don't.]
.