About capacitors for PC motherboards...

[~misfit~]

Hi folks, new poster, electronically challenged....

I hope you will be kind enough to help me, I face the following problem:

I have a motherboard in one of my PCs that has gone all flakey. It has bad
cap syndrome. The low ESR electrolytic aluminium can caps are bulging and
leaking. I know that this problem was rife around 2000/2001 but thought it
fixed by the time I bought this board, 2003. Sadly not true. The board
itself has been, up until the caps failed, an excellent, stable and very
useful bit of kit.

I have replaced capacitors on motherboards before with success so am not
worried about that side of things. My problem lies in the fact that I live
in New Zealand. It seems that it's just about impossible to get a handful of
caps the same specs as the ones on the board anywhere. I've found suppliers
in the US but not any who will ship *outside* of the US. As I assemble
computers for a hobby and used this same model mobo in a few different
builds for friends I'm expecting to see others with the same problem soon
too.

Ok, so the main caps I need are low ESR, 3,300uF, 6.3v, 10mm can, 5mm
lead-spacing (25mm high but that's unimportant, there's room to go higher)
electrolytic caps. I can't seem to source these anywhere in this wonderful
country of mine. I've been told I can get them in 15mm diameter cans which
is unacceptable as there is a row of five tightly satcked together. I've
tried for Nichcon and Rubycon first as I hear that these are good caps. The
local Nichicon agent said they don't have them ex-stock but could order them
in. A minimum order of 2K pieces and a 90-day turn-around! I need 5 for a
start, maybe another 20 later. You wouldn't believe how difficult a
seemingly easy thing like this has turned out to be.

So, to the advice I'd like: A year or two back, when I first struck
motherboards with bad caps a guy I met on a NZ newsgroup offered to get me
some suitable caps as part of his next order (which lead me to assume he was
"in the trade"). The caps needed for those three mobos I was hoping to fix
were the ame spec, 3,300uF 6.2v.... Anyway, the caps he sent me were 2,200
10v and he said the should be fine. He couldn't get the exact match. I went
ahead and fitted the caps and the motherboards worked fine. I still have one
of them running the PC I use for a stereo/mp3 jukebox that runs 24/7.

The board I'm currently looking to fix is (was) running an Athlon XP3200+ so
is still reasonably useful, I was expecting to get a few more years out of
it. The boards I repaired before using different spec caps were less
valuable so I wasn't so worried if it didn't work. However, I've been told
in a local newgroup that, often, caps on a mobo are specced +/- 50 to 100%,
that they are 'smoothing' caps. (They are right next to the CPU and the
first symptoms was unstable CPU core voltage).

My question to the collective here is: Do you folks think it would be
relatively safe to use 2,200 uF caps in this motherboard? It's worth quite a
bit more than the others I fixed and I can't afford to make a mistake with
it.

Kind regards, and thanks in advance,
--
Shaun.

 

[Eeyore]

~misfit~ wrote:

Hi folks, new poster, electronically challenged....

I hope you will be kind enough to help me, I face the following problem:

http://search.ebay.co.uk/capacitor-motherboard_W0QQfromZR40

Graham

 

[Eeyore]

~misfit~ wrote:

My question to the collective here is: Do you folks think it would be
relatively safe to use 2,200 uF caps in this motherboard? It's worth quite a
bit more than the others I fixed and I can't afford to make a mistake with
it.

It's *very* likely that the capacitance is chosen for ripple current capability
and not the absolute value. Subbing a smaller capacitance from a top-grade
vendor is very likely to do the job just fine.

Graham

 

[Homer J Simpson]

"~misfit~" <misfit61nz@yahookers.com.au> wrote in message
news:ect4jg$2po2$1@registered.motzarella.de...

I have replaced capacitors on motherboards before with success so am not
worried about that side of things. My problem lies in the fact that I live
in New Zealand. It seems that it's just about impossible to get a handful
of caps the same specs as the ones on the board anywhere. I've found
suppliers in the US but not any who will ship *outside* of the US. As I
assemble computers for a hobby and used this same model mobo in a few
different builds for friends I'm expecting to see others with the same
problem soon too.

Try R P Electronics http://www.rpelectronics.com/

 

[Franc Zabkar]

On Mon, 28 Aug 2006 09:56:23 +1200, "~misfit~"
<misfit61nz@yahookers.com.au> put finger to keyboard and composed:

Ok, so the main caps I need are low ESR, 3,300uF, 6.3v, 10mm can, 5mm
lead-spacing (25mm high but that's unimportant, there's room to go higher)
electrolytic caps. I can't seem to source these anywhere in this wonderful
country of mine. I've been told I can get them in 15mm diameter cans which
is unacceptable as there is a row of five tightly satcked together.

Farnell have 10V, 12.5mm dia caps at AU$3.67 + GST:
http://au.farnell.com/jsp/endecaSearch/partDetail.jsp?SKU=1144687

Alternatively, WES Components, Ashfield NSW, have 10V, 13mm dia, low
ESR, Hitano EXR series caps, AU$1.15 including GST (trade).
http://www.wescomponents.com/datasheets/EXR.pdf
http://www.wagner.net.au/Catalogue/02_09.pdf

The same series has a 6.3V, 3300uF, 10mm dia cap, but WES don't list
it in their 2003 catalogue, or their Nov 2005 supplement.

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

 

[Ancient_Hacker]

Try www.digi-key.com, they're inexpensive, have the biggest stock in
the world, and ship internationally. Search for electrolytics, low
esr.

BTW you really don't need to get those tiny 10mm size-- you can mount
them with up to 1 inch leads no problem and thereby spread them out a
bit.

 

[Eeyore]

Ancient_Hacker wrote:

Try www.digi-key.com, they're inexpensive, have the biggest stock in
the world, and ship internationally. Search for electrolytics, low
esr.

BTW you really don't need to get those tiny 10mm size-- you can mount
them with up to 1 inch leads no problem and thereby spread them out a
bit.

Not very practical to be honest.

Graham

 

[~misfit~]

Franc Zabkar wrote:

On Mon, 28 Aug 2006 09:56:23 +1200, "~misfit~"
misfit61nz@yahookers.com.au> put finger to keyboard and composed:

Ok, so the main caps I need are low ESR, 3,300uF, 6.3v, 10mm can,
5mm lead-spacing (25mm high but that's unimportant, there's room to
go higher) electrolytic caps. I can't seem to source these anywhere
in this wonderful country of mine. I've been told I can get them in
15mm diameter cans which is unacceptable as there is a row of five
tightly satcked together.

Farnell have 10V, 12.5mm dia caps at AU$3.67 + GST:
http://au.farnell.com/jsp/endecaSearch/partDetail.jsp?SKU=1144687

Alternatively, WES Components, Ashfield NSW, have 10V, 13mm dia, low
ESR, Hitano EXR series caps, AU$1.15 including GST (trade).
http://www.wescomponents.com/datasheets/EXR.pdf
http://www.wagner.net.au/Catalogue/02_09.pdf

The same series has a 6.3V, 3300uF, 10mm dia cap, but WES don't list
it in their 2003 catalogue, or their Nov 2005 supplement.

- Franc Zabkar

Thanks Franc. However, these caps are really bunched up, I *need* them to be
10mm. :-(
--
Shaun.

 

[~misfit~]

Eeyore wrote:

~misfit~ wrote:

Hi folks, new poster, electronically challenged....

I hope you will be kind enough to help me, I face the following
problem:

http://search.ebay.co.uk/capacitor-motherboard_W0QQfromZR40

Graham

"These are recovered from almost new motherboards"

No thanks. The board that has bad caps has others around it of the same
brand that look just fine but are probably no better, just under less load.
--
Shaun.

 

[~misfit~]

Eeyore wrote:

~misfit~ wrote:

My question to the collective here is: Do you folks think it would
be relatively safe to use 2,200 uF caps in this motherboard? It's
worth quite a bit more than the others I fixed and I can't afford
to make a mistake with it.

It's *very* likely that the capacitance is chosen for ripple current
capability and not the absolute value. Subbing a smaller capacitance
from a top-grade vendor is very likely to do the job just fine.

Thanks. That's what I'd hoped to hear.
--
Shaun.

 

[~misfit~]

Eeyore wrote:

Ancient_Hacker wrote:

Try www.digi-key.com, they're inexpensive, have the biggest stock in
the world, and ship internationally. Search for electrolytics, low
esr.

BTW you really don't need to get those tiny 10mm size-- you can
mount them with up to 1 inch leads no problem and thereby spread
them out a bit.

Not very practical to be honest.

Not at all. I've tried that before on a different mobo, insulating the long
leads. Still didn't work, I couldn't get them to fit. There are a bunch of
caps (three rows) between the CPU socket and the backplane. No room for big
caps on stalks.
--
Shaun.

 

[~misfit~]

Ancient_Hacker wrote:

Try www.digi-key.com, they're inexpensive, have the biggest stock in
the world, and ship internationally. Search for electrolytics, low
esr.

Ok, trying it now. What's with the 'series' column?

BTW you really don't need to get those tiny 10mm size-- you can mount
them with up to 1 inch leads no problem and thereby spread them out a
bit.

Umm, nah. I've tried that before on a board less crowded than this one. Not
an option really.

Cheers,
--
Shaun.

 

[~misfit~]

Homer J Simpson wrote:

"~misfit~" <misfit61nz@yahookers.com.au> wrote in message
news:ect4jg$2po2$1@registered.motzarella.de...

I have replaced capacitors on motherboards before with success so
am not worried about that side of things. My problem lies in the
fact that I live in New Zealand. It seems that it's just about
impossible to get a handful of caps the same specs as the ones on
the board anywhere. I've found suppliers in the US but not any who
will ship *outside* of the US. As I assemble computers for a hobby
and used this same model mobo in a few different builds for friends
I'm expecting to see others with the same problem soon too.

Try R P Electronics http://www.rpelectronics.com/

Thanks, I've checked the site out but seem unable to search for 'low ESR' or
find such a category.

Cheers,
--
Shaun.

 

[Michael A. Terrell]

Eeyore wrote:

~misfit~ wrote:

My question to the collective here is: Do you folks think it would be
relatively safe to use 2,200 uF caps in this motherboard? It's worth quite a
bit more than the others I fixed and I can't afford to make a mistake with
it.

It's *very* likely that the capacitance is chosen for ripple current capability
and not the absolute value. Subbing a smaller capacitance from a top-grade
vendor is very likely to do the job just fine.

Graham

Bullshit. Lower capacitance will reduce the filtering, and cause
problems.


--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

 

[Eeyore]

"Michael A. Terrell" wrote:

Eeyore wrote:

~misfit~ wrote:

My question to the collective here is: Do you folks think it would be
relatively safe to use 2,200 uF caps in this motherboard? It's worth quite a
bit more than the others I fixed and I can't afford to make a mistake with
it.

It's *very* likely that the capacitance is chosen for ripple current capability
and not the absolute value. Subbing a smaller capacitance from a top-grade
vendor is very likely to do the job just fine.

Graham

Bullshit. Lower capacitance will reduce the filtering, and cause
problems.

At smps frequencies ? You're simply babbling.

Graham

 

[Michael A. Terrell]

Eeyore wrote:

"Michael A. Terrell" wrote:

Eeyore wrote:

~misfit~ wrote:

My question to the collective here is: Do you folks think it would be
relatively safe to use 2,200 uF caps in this motherboard? It's worth quite a
bit more than the others I fixed and I can't afford to make a mistake with
it.

It's *very* likely that the capacitance is chosen for ripple current capability
and not the absolute value. Subbing a smaller capacitance from a top-grade
vendor is very likely to do the job just fine.

Graham

Bullshit. Lower capacitance will reduce the filtering, and cause
problems.

At smps frequencies ? You're simply babbling.

Graham

How many motherboards have you repaired? 0? How many switch mode
power supplies have you repaired? 0?

You have a minimum capacitance requirement for a given output
current, no matter what the frequency. Yes, it varies with the
switching frequency, but all it takes is a little ripple to make the
switcher shut down or make the CPU's operation flaky. I had what looked
like a bad CD burner. I found a couple failing electrolytics on the
motherboard. When the CPU was loafing along, everything was fine, but
when it tied to do a resource intensive task that required more current,
it would become erratic. A sign of failing PS electrolytics is a system
crash when you save a large file to your hard drive.

Do you really believe that the OEM would use larger caps than they
need? You claim to be a designer, so you should know that OEM
motherboard manufacturing is the most penny pinching segment of the
market. You are either a fucking idiot or you get a big kick out of
misleading the beginners.

Go back to your crappy audio mixing boards.


--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

 

[Peter Dettmann]

On Thu, 31 Aug 2006 16:19:15 GMT, "Michael A. Terrell"
<mike.terrell@earthlink.net> wrote:

Eeyore wrote:

"Michael A. Terrell" wrote:

Eeyore wrote:

~misfit~ wrote:

My question to the collective here is: Do you folks think it would be
relatively safe to use 2,200 uF caps in this motherboard? It's worth quite a
bit more than the others I fixed and I can't afford to make a mistake with
it.

It's *very* likely that the capacitance is chosen for ripple current capability
and not the absolute value. Subbing a smaller capacitance from a top-grade
vendor is very likely to do the job just fine.

Graham

Bullshit. Lower capacitance will reduce the filtering, and cause
problems.

At smps frequencies ? You're simply babbling.

Graham


How many motherboards have you repaired? 0? How many switch mode
power supplies have you repaired? 0?

You have a minimum capacitance requirement for a given output
current, no matter what the frequency. Yes, it varies with the
switching frequency, but all it takes is a little ripple to make the
switcher shut down or make the CPU's operation flaky. I had what looked
like a bad CD burner. I found a couple failing electrolytics on the
motherboard. When the CPU was loafing along, everything was fine, but
when it tied to do a resource intensive task that required more current,
it would become erratic. A sign of failing PS electrolytics is a system
crash when you save a large file to your hard drive.

Do you really believe that the OEM would use larger caps than they
need?

Of course they would. The realities of life are that tey would
reasonably use what was available, and preferrably cheap instead of
buying in a new capacitor rather than just using a big pile of
leftovers.

However having said that and being aware that the precise capacitance
(over a large range) is not strictly important for a simple filtering
function. It depends on largely arbitary specifications we set for
allowable ripple, and then a big fudge factor to allow for loss of
capacitance with ageing, and then bring it to the nearest standard and
cheaply available size. No it is not a precise science at all.

You claim to be a designer, so you should know that OEM
motherboard manufacturing is the most penny pinching segment of the
market. You are either a fucking idiot or you get a big kick out of
misleading the beginners.

Go back to your crappy audio mixing boards.

Surely a good designer would use logic and actual facts rather than
generalise, and then use snide childish vitriol?

Peter Dettmann

 

[Eeyore]

Peter Dettmann wrote:

On Thu, 31 Aug 2006 16:19:15 GMT, "Michael A. Terrell"
mike.terrell@earthlink.net> wrote:

Do you really believe that the OEM would use larger caps than they
need?

Of course they would. The realities of life are that tey would
reasonably use what was available, and preferrably cheap instead of
buying in a new capacitor rather than just using a big pile of
leftovers.

Exactly.

However having said that and being aware that the precise capacitance
(over a large range) is not strictly important for a simple filtering
function. It depends on largely arbitary specifications we set for
allowable ripple, and then a big fudge factor to allow for loss of
capacitance with ageing, and then bring it to the nearest standard and
cheaply available size. No it is not a precise science at all.

Quite so. In every PSU I've ever designed the value is primarily set by ripple current
rating ( and ESR ) which is anything but repeatable from brand to brand, so speccing a
slightly larger one allows the use of more suppliers as alternative sources.

You claim to be a designer, so you should know that OEM
motherboard manufacturing is the most penny pinching segment of the
market. You are either a fucking idiot or you get a big kick out of
misleading the beginners.

Go back to your crappy audio mixing boards.

Surely a good designer would use logic and actual facts rather than
generalise, and then use snide childish vitriol?

He's not a designer. He's a 'tech'.

Graham

 

[Franc Zabkar]

On Wed, 30 Aug 2006 00:25:56 +1200, "~misfit~"
<misfit61nz@yahookers.com.au> put finger to keyboard and composed:

Franc Zabkar wrote:
On Mon, 28 Aug 2006 09:56:23 +1200, "~misfit~"
misfit61nz@yahookers.com.au> put finger to keyboard and composed:

Ok, so the main caps I need are low ESR, 3,300uF, 6.3v, 10mm can,
5mm lead-spacing (25mm high but that's unimportant, there's room to
go higher) electrolytic caps. I can't seem to source these anywhere
in this wonderful country of mine. I've been told I can get them in
15mm diameter cans which is unacceptable as there is a row of five
tightly satcked together.

Farnell have 10V, 12.5mm dia caps at AU$3.67 + GST:
http://au.farnell.com/jsp/endecaSearch/partDetail.jsp?SKU=1144687

Alternatively, WES Components, Ashfield NSW, have 10V, 13mm dia, low
ESR, Hitano EXR series caps, AU$1.15 including GST (trade).
http://www.wescomponents.com/datasheets/EXR.pdf
http://www.wagner.net.au/Catalogue/02_09.pdf

The same series has a 6.3V, 3300uF, 10mm dia cap, but WES don't list
it in their 2003 catalogue, or their Nov 2005 supplement.

- Franc Zabkar

Thanks Franc. However, these caps are really bunched up, I *need* them to be
10mm. :-(

Search for EXR3300/6.3 at http://www.x-on.com.au/ (Hitano agents)

You can get 10 for $40, or 5 for $36, including $15 freight, GST, and
a $15 penalty for ordering less than $200 worth of goods. :-(

You may be able to avoid the freight cost if you live near one of
their offices:
http://www.x-on.com.au/contact.asp

- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.

 

[Michael A. Terrell]

Peter Dettmann wrote:

On Thu, 31 Aug 2006 16:19:15 GMT, "Michael A. Terrell"
mike.terrell@earthlink.net> wrote:

Eeyore wrote:

"Michael A. Terrell" wrote:

Eeyore wrote:

~misfit~ wrote:

My question to the collective here is: Do you folks think it would be
relatively safe to use 2,200 uF caps in this motherboard? It's worth quite a
bit more than the others I fixed and I can't afford to make a mistake with
it.

It's *very* likely that the capacitance is chosen for ripple current capability
and not the absolute value. Subbing a smaller capacitance from a top-grade
vendor is very likely to do the job just fine.

Graham

Bullshit. Lower capacitance will reduce the filtering, and cause
problems.

At smps frequencies ? You're simply babbling.

Graham


How many motherboards have you repaired? 0? How many switch mode
power supplies have you repaired? 0?

You have a minimum capacitance requirement for a given output
current, no matter what the frequency. Yes, it varies with the
switching frequency, but all it takes is a little ripple to make the
switcher shut down or make the CPU's operation flaky. I had what looked
like a bad CD burner. I found a couple failing electrolytics on the
motherboard. When the CPU was loafing along, everything was fine, but
when it tied to do a resource intensive task that required more current,
it would become erratic. A sign of failing PS electrolytics is a system
crash when you save a large file to your hard drive.

Do you really believe that the OEM would use larger caps than they
need?

Of course they would. The realities of life are that tey would
reasonably use what was available, and preferrably cheap instead of
buying in a new capacitor rather than just using a big pile of
leftovers.

They have anything they want available. They are made in China,
where most of the electrolytics and other components are made. They are
built in huge batches, so if they can shave 2 cents or more off each cap
it can be a dollar or more savings per board, times the total run of
boards. In some cases it can hit a million dollars per year on just
electrolytics. That cost cutting can make a difference between a profit
and bankruptcy. If the board is a buck higher than a similar board HP,
gateway and similar OEM PC companies will buy the cheaper board.

However having said that and being aware that the precise capacitance
(over a large range) is not strictly important for a simple filtering
function. It depends on largely arbitary specifications we set for
allowable ripple, and then a big fudge factor to allow for loss of
capacitance with ageing, and then bring it to the nearest standard and
cheaply available size. No it is not a precise science at all.

You claim to be a designer, so you should know that OEM
motherboard manufacturing is the most penny pinching segment of the
market. You are either a fucking idiot or you get a big kick out of
misleading the beginners.

Go back to your crappy audio mixing boards.

Surely a good designer would use logic and actual facts rather than
generalise, and then use snide childish vitriol?

Peter Dettmann

Believe whatever pleases you. I worked in the real world. Very few
companies can completely ignore component costs. I have worked at a
few, but there aren't a lot of customers for radios that are over
$50,000 US each. One I built is aboard the ISS. It used a purchased
Switching power supply. We started with Vicor, but their failure rate
was so bad that we went through samples from over 20 vendors before we
found something that Engineering and manufacturing could agree on. I
rejected a lot of them outright for poor mechanical design. Cheap fans,
PC boards that didn't have enough support, and ones that didn't shut
down on overload. The sad thing was the few remaining domestic sources
had too much noise to be acceptable. the 100 KHz switching frequency
had harmonics well past 70 MHz, and would leak into the IF stages.


--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida