About capacitors for PC motherboards...

[Michael A. Terrell]

Eeyore wrote:

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

A good tech is better than a crappy designer any day. A number of
the EEs I've worked with came to the production floor with questions
about our older products. They would make written requests for me to
test new designs and to qualify new components, or do failure analysis
and track failure trends. I had more databooks in my corner of the
production floor than most of the engineers kept in their offices. A
common question from them was why I didn't have a degree, because I
could do their work, and often caught errors in their preliminary work.


--
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:

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

A good tech is better than a crappy designer any day.

Fine by me.

A number of
the EEs I've worked with came to the production floor with questions
about our older products. They would make written requests for me to
test new designs and to qualify new components, or do failure analysis
and track failure trends. I had more databooks in my corner of the
production floor than most of the engineers kept in their offices. A
common question from them was why I didn't have a degree, because I
could do their work, and often caught errors in their preliminary work.

Designers do know stuff that techs don't though.

Graham

 

[Eeyore]

"Michael A. Terrell" wrote:

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.

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.

" Shuttle FV24 and FV25 Kit Update!!!!!!
I ran out of stock on the 10v caps that were needed to make the kits. 3300uF~4700uF 10v
caps in a 12.5mm diameter are IMPOSSIBLE to find, and anything larger will not fit the
board. I had to temporarily discontinue the kit because I didn't have a FV24 or FV25
motherboard to conduct experiments on. However, early last month, I got one of each, and
did some extensive testing and research on it, and discovered the 10v caps were NOT needed.
The positions those caps are in only have 2v MAX (VCORE voltage only) at any given time, and
the voltage does NOT fluctuate. Why Shuttle used 10v caps is beyond me, but manufacturers
have been known to do this in the past.

I recapped them both with Rubycon and Nichicon 3300uF 6.3v, and they've been running 100%
reliable under heavy load, and in brutally hot conditions for over a month now. "

http://www.badcaps.net/kits/

So, Michael, why did they use 10V caps ?

Graham

 

[Michael A. Terrell]

Eeyore wrote:

"Michael A. Terrell" wrote:

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.

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.

" Shuttle FV24 and FV25 Kit Update!!!!!!
I ran out of stock on the 10v caps that were needed to make the kits. 3300uF~4700uF 10v
caps in a 12.5mm diameter are IMPOSSIBLE to find, and anything larger will not fit the
board. I had to temporarily discontinue the kit because I didn't have a FV24 or FV25
motherboard to conduct experiments on. However, early last month, I got one of each, and
did some extensive testing and research on it, and discovered the 10v caps were NOT needed.
The positions those caps are in only have 2v MAX (VCORE voltage only) at any given time, and
the voltage does NOT fluctuate. Why Shuttle used 10v caps is beyond me, but manufacturers
have been known to do this in the past.

I recapped them both with Rubycon and Nichicon 3300uF 6.3v, and they've been running 100%
reliable under heavy load, and in brutally hot conditions for over a month now. "

http://www.badcaps.net/kits/

So, Michael, why did they use 10V caps ?

Graham

Shuttle is a high end board maker. Using the 10 volt caps will ensure
their boards live through their warranty period. They aren't trying to
make boards for under $500 systemms, so they can spend a couple more
cents per electrolytic. I haven't seen a shuttle board with bad
electrolytics, so far. I know some do fail, but none have crossed my
bench. Hell, I have a Sun workstation in the shop that had all but one
of the CPU PS electrolytics fail, and it isn't a cheap computer.




--
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:

"Michael A. Terrell" wrote:

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.

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.

" Shuttle FV24 and FV25 Kit Update!!!!!!
I ran out of stock on the 10v caps that were needed to make the kits. 3300uF~4700uF 10v
caps in a 12.5mm diameter are IMPOSSIBLE to find, and anything larger will not fit the
board. I had to temporarily discontinue the kit because I didn't have a FV24 or FV25
motherboard to conduct experiments on. However, early last month, I got one of each, and
did some extensive testing and research on it, and discovered the 10v caps were NOT needed.
The positions those caps are in only have 2v MAX (VCORE voltage only) at any given time, and
the voltage does NOT fluctuate. Why Shuttle used 10v caps is beyond me, but manufacturers
have been known to do this in the past.

I recapped them both with Rubycon and Nichicon 3300uF 6.3v, and they've been running 100%
reliable under heavy load, and in brutally hot conditions for over a month now. "

http://www.badcaps.net/kits/

So, Michael, why did they use 10V caps ?

Graham

Shuttle is a high end board maker. Using the 10 volt caps will ensure
their boards live through their warranty period.

How would the voltage rating influence the lifetime ?

Graham

 

[Mark Fortune]

Eeyore wrote:

"Michael A. Terrell" wrote:


Eeyore wrote:

"Michael A. Terrell" wrote:


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.

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.

" Shuttle FV24 and FV25 Kit Update!!!!!!
I ran out of stock on the 10v caps that were needed to make the kits. 3300uF~4700uF 10v
caps in a 12.5mm diameter are IMPOSSIBLE to find, and anything larger will not fit the
board. I had to temporarily discontinue the kit because I didn't have a FV24 or FV25
motherboard to conduct experiments on. However, early last month, I got one of each, and
did some extensive testing and research on it, and discovered the 10v caps were NOT needed.
The positions those caps are in only have 2v MAX (VCORE voltage only) at any given time, and
the voltage does NOT fluctuate. Why Shuttle used 10v caps is beyond me, but manufacturers
have been known to do this in the past.

I recapped them both with Rubycon and Nichicon 3300uF 6.3v, and they've been running 100%
reliable under heavy load, and in brutally hot conditions for over a month now. "

http://www.badcaps.net/kits/

So, Michael, why did they use 10V caps ?

Graham

Shuttle is a high end board maker. Using the 10 volt caps will ensure
their boards live through their warranty period.


How would the voltage rating influence the lifetime ?

Graham

It makes them better able to handle voltage spikes, and they handle
smaller loads in relation to their maximum.

Mark

 

[Eeyore]

Mark Fortune wrote:

Eeyore wrote:

"Michael A. Terrell" wrote:


Eeyore wrote:

"Michael A. Terrell" wrote:


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.

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.

" Shuttle FV24 and FV25 Kit Update!!!!!!
I ran out of stock on the 10v caps that were needed to make the kits. 3300uF~4700uF 10v
caps in a 12.5mm diameter are IMPOSSIBLE to find, and anything larger will not fit the
board. I had to temporarily discontinue the kit because I didn't have a FV24 or FV25
motherboard to conduct experiments on. However, early last month, I got one of each, and
did some extensive testing and research on it, and discovered the 10v caps were NOT needed.
The positions those caps are in only have 2v MAX (VCORE voltage only) at any given time, and
the voltage does NOT fluctuate. Why Shuttle used 10v caps is beyond me, but manufacturers
have been known to do this in the past.

I recapped them both with Rubycon and Nichicon 3300uF 6.3v, and they've been running 100%
reliable under heavy load, and in brutally hot conditions for over a month now. "

http://www.badcaps.net/kits/

So, Michael, why did they use 10V caps ?

Graham

Shuttle is a high end board maker. Using the 10 volt caps will ensure
their boards live through their warranty period.


How would the voltage rating influence the lifetime ?

Graham

It makes them better able to handle voltage spikes,

What voltage spikes would those be btw ? Read the quoted text.

and they handle
smaller loads in relation to their maximum.

What's that supposed to mean ?

Graham

 

[Michael A. Terrell]

Eeyore wrote:

How would the voltage rating influence the lifetime ?

Are you really that clueless?


--
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:

How would the voltage rating influence the lifetime ?

Are you really that clueless?

Come on tell me. I want to know what part you think voltage rating plays in
lifetime for electrolytic caps.

Graham

 

[Eeyore]

Spehro Pefhany wrote:

On Sat, 02 Sep 2006 17:10:09 +0100, the renowned Eeyore
rabbitsfriendsandrelations@REMOVETHIS.hotmail.com> wrote:

How would the voltage rating influence the lifetime ?

Graham

One rule of thumb is "proportionally", up to 2:1 derating. So, you'll
get 58% more lifetime out of a 10V elect. cap operated at 5V compared
to a 6.3V cap, all other things being equal. Whether the formula holds
for the low-Z caps, I don't know.

Where do you get this rule of thumb from ? I suspect it's a myth.

Graham

 

[Spehro Pefhany]

On Sat, 02 Sep 2006 17:10:09 +0100, the renowned Eeyore
<rabbitsfriendsandrelations@REMOVETHIS.hotmail.com> wrote:

How would the voltage rating influence the lifetime ?

Graham

One rule of thumb is "proportionally", up to 2:1 derating. So, you'll
get 58% more lifetime out of a 10V elect. cap operated at 5V compared
to a 6.3V cap, all other things being equal. Whether the formula holds
for the low-Z caps, I don't know.


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[Spehro Pefhany]

On Sat, 02 Sep 2006 20:14:02 +0100, the renowned Eeyore
<rabbitsfriendsandrelations@REMOVETHIS.hotmail.com> wrote:

Spehro Pefhany wrote:

On Sat, 02 Sep 2006 17:10:09 +0100, the renowned Eeyore
rabbitsfriendsandrelations@REMOVETHIS.hotmail.com> wrote:

How would the voltage rating influence the lifetime ?

Graham

One rule of thumb is "proportionally", up to 2:1 derating. So, you'll
get 58% more lifetime out of a 10V elect. cap operated at 5V compared
to a 6.3V cap, all other things being equal. Whether the formula holds
for the low-Z caps, I don't know.

Where do you get this rule of thumb from ? I suspect it's a myth.

Graham

Check electrolytic capacitor manufacturer's data.

Here, for example, is Cornell Dublier which gives a somewhat more
complex formula:

k = 4.3 - 3.3 * Va/Vr , which works out to 1.0 at Va = Vr, and 2.65 at
Va = Vr/2, and 1.68 at Va = 5/6.3, which would yield 63% more life for
a 10V cap compared to a 6.3V cap, not so far from the +58% calculated
by my rule of thumb.

http://electrochem.cwru.edu/ed/encycl/misc/c04-appguide.pdf

Of course temperature is a greater factor in the lifetime, but there
is often little you can do about that.


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[Eeyore]

Spehro Pefhany wrote:

On Sat, 02 Sep 2006 20:14:02 +0100, the renowned Eeyore
rabbitsfriendsandrelations@REMOVETHIS.hotmail.com> wrote:

Spehro Pefhany wrote:

On Sat, 02 Sep 2006 17:10:09 +0100, the renowned Eeyore
rabbitsfriendsandrelations@REMOVETHIS.hotmail.com> wrote:

How would the voltage rating influence the lifetime ?

Graham

One rule of thumb is "proportionally", up to 2:1 derating. So, you'll
get 58% more lifetime out of a 10V elect. cap operated at 5V compared
to a 6.3V cap, all other things being equal. Whether the formula holds
for the low-Z caps, I don't know.

Where do you get this rule of thumb from ? I suspect it's a myth.

Graham

Check electrolytic capacitor manufacturer's data.

I very much have done !

Here, for example, is Cornell Dublier which gives a somewhat more
complex formula:

k = 4.3 - 3.3 * Va/Vr , which works out to 1.0 at Va = Vr, and 2.65 at
Va = Vr/2, and 1.68 at Va = 5/6.3, which would yield 63% more life for
a 10V cap compared to a 6.3V cap, not so far from the +58% calculated
by my rule of thumb.

At what ripple current ?

http://electrochem.cwru.edu/ed/encycl/misc/c04-appguide.pdf

Of course temperature is a greater factor in the lifetime, but there
is often little you can do about that.

Don't give Michael too many clues. I'm waiting for his reply.

Graham

 

[Michael A. Terrell]

Eeyore wrote:

Don't give Michael too many clues. I'm waiting for his reply.

You won't get it. I replied directly to Spehro, and you are being
plonked on this newsgroup, too.


--
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:

Don't give Michael too many clues. I'm waiting for his reply.

You won't get it. I replied directly to Spehro, and you are being
plonked on this newsgroup, too.

By you ? I'm devatated. FYI applied voltage is a relatively insignificant factor
in capacitor lifetime.

So, what really causes electrolytic caps to 'wear out' ? I know. Do you ?

Graham

 

[Joel Kolstad]

"Eeyore" <rabbitsfriendsandrelations@REMOVETHIS.hotmail.com> wrote in message
news:44F92D67.CC93418D@REMOVETHIS.hotmail.com...

Designers do know stuff that techs don't though.

Seems like that would depend on the designer and tech in question... there are
plenty of freshly-minted EEs out there today who couldn't bias a transistor to
save their lives... although if that isn't their job, I suppose it doesn't
really matter...

 

[Michael A. Terrell]

Joel Kolstad wrote:

"Eeyore" <rabbitsfriendsandrelations@REMOVETHIS.hotmail.com> wrote in message
news:44F92D67.CC93418D@REMOVETHIS.hotmail.com...
Designers do know stuff that techs don't though.

Seems like that would depend on the designer and tech in question... there are
plenty of freshly-minted EEs out there today who couldn't bias a transistor to
save their lives... although if that isn't their job, I suppose it doesn't
really matter...

Aren't they the EEs who end up in sales? ;-)


--
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

 

[Joel Kolstad]

"Michael A. Terrell" <mike.terrell@earthlink.net> wrote in message
news:44FDDBFA.2384A1F@earthlink.net...

Aren't they the EEs who end up in sales? ;-)

I was surprised to find that these days it's generally *required* to have a EE
degree if you want to sell scopes, etc. for Agilent. I honestly think this is
more a reflection of how dumbed-down the average EE curriculum is today rather
than how "sophisticated" modern test equipment has become.

 

[Eeyore]

Joel Kolstad wrote:

"Michael A. Terrell" <mike.terrell@earthlink.net> wrote in message
news:44FDDBFA.2384A1F@earthlink.net...
Aren't they the EEs who end up in sales? ;-)

I was surprised to find that these days it's generally *required* to have a EE
degree if you want to sell scopes, etc. for Agilent. I honestly think this is
more a reflection of how dumbed-down the average EE curriculum is today rather
than how "sophisticated" modern test equipment has become.

More than likely I reckon.

Graham

 

[Michael A. Terrell]

Joel Kolstad wrote:

"Michael A. Terrell" <mike.terrell@earthlink.net> wrote in message
news:44FDDBFA.2384A1F@earthlink.net...
Aren't they the EEs who end up in sales? ;-)

I was surprised to find that these days it's generally *required* to have a EE
degree if you want to sell scopes, etc. for Agilent. I honestly think this is
more a reflection of how dumbed-down the average EE curriculum is today rather
than how "sophisticated" modern test equipment has become.

Yes, that's sad, and it seems the less a new engineer knows, the
more they brag about their degrees. The older engineers didn't even have
their diplomas on the wall, but some of the younger ones had copies on
every wall of their cubical to try to impress everyone else. The real
engineer could look at a schematics and the notes, make a decision and
have the ECO done before my modifications made it through test. The
kids took months.

I was amazed that I knew more about a digital H-P spectrum analyzer
I'd only had for a few hours, than the engineer who had it for two
years. I was configuring them and teaching the other techs long before
one was freed up for my bench(es). He referred to the advanced menus as
a "Bunch of useless garbage", but you had to use them daily. He
couldn't even change it from the IEEE-488 printer it was set for when it
was purchased, and four others didn't know where to find the settings.


--
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