many mobo's will benefit from polymer cap upgrades

**PCBONEZ** · 12-07-2007, 02:42 PM

Re: many mobo's will benefit from polymer cap upgrades

. . . . . . . . . deltaI * deltaT
Cb > = . . --------------------------
. . . . . . . . (deltaV - (deltaI * ESR))

I may be confused but I don't follow your thinking.

If deltaV, deltaI, and deltaT remain the same and the ESR goes down:

Then: (deltaI * ESR) -> goes down.

And: (deltaV - (deltaI * ESR)) -> goes up.

..... which means Cb goes up.

.

**PCBONEZ** · 12-07-2007, 02:52 PM

Re: many mobo's will benefit from polymer cap upgrades

Assuming random values

. . . . . . . . . 3 * 3
Cb > = . . --------------- = 1.5
. . . . . . . . (3 - (3 * 3))

Lowering the ESR

. . . . . . . . . 3 * 3
Cb > = . . --------------- = 3.0
. . . . . . . . (3 - (3 * 2))

Cb went up.
~~~ Or am I missing something?
.

**gdement** · 12-08-2007, 05:46 AM

Re: many mobo's will benefit from polymer cap upgrades

Originally posted by PCBONEZ

Assuming random values

. . . . . . . . . 3 * 3
Cb > = . . --------------- = 1.5
. . . . . . . . (3 - (3 * 3))

Lowering the ESR

. . . . . . . . . 3 * 3
Cb > = . . --------------- = 3.0
. . . . . . . . (3 - (3 * 2))

Cb went up.
~~~ Or am I missing something?
.

Your denominators are negative, which isn't allowed for this formula. Your first result is actually -1.5, and the second is -3. This issue imposes a limit on how high the ESR can be, but I don't know if that limit is a real thing or just a limitation of the formula.

. . . . . . . . . deltaI * deltaT
Cb > = . . --------------------------
. . . . . . . . (deltaV - (deltaI * ESR))

I may be confused but I don't follow your thinking.

If deltaV, deltaI, and deltaT remain the same and the ESR goes down:

Then: (deltaI * ESR) -> goes down.

And: (deltaV - (deltaI * ESR)) -> goes up.

Correct

..... which means Cb goes up.

no, because the denominator got larger, so Cb will get smaller.

**Logistics** · 12-08-2007, 02:16 PM

Re: many mobo's will benefit from polymer cap upgrades

What's everyones take on what will now be (with the polymers) an extremely high leakage current compared to before?

**PCBONEZ** · 12-08-2007, 03:50 PM

Re: many mobo's will benefit from polymer cap upgrades

Damn. I think I forgot math... LOL

Thank you.

I was ignoring the sign (+/-) intentionally. I was confused by a note in this document:

>>>>>
Note: The denominator of the Cout equation cannot be a negative value, which implies a negative capacitor (such as a battery).
<<<<<
[AMD uses "Cout" to Intel's "Cb".]
-
I took that to mean the sign of the demoninator was simply ignored to make the equation work.

.

**gonzo0815** · 12-09-2007, 06:45 AM

Re: many mobo's will benefit from polymer cap upgrades

Leakage current of polymers are in the same range then aluminum electrolytic ones.
E.g. a Samxon GA 3300uF 6.3v (7mR) capacitor would have a leakage current about 624uA.
An Oscon SEPC 820uF 4v (7mR) leakage current is 656uA.
Sure, the higher capacitance polymers from Oscon do have larger ripple current, but it depends on the series to.

I would say, that the increase is not that significant for most applications.

**Logistics** · 12-09-2007, 05:03 PM

Re: many mobo's will benefit from polymer cap upgrades

ಠ_ಠ Ah, it's your application. I wasn't paying attention. In the more common sizes up to 16V, I have never seen a polymer cap that didn't have nearly twice or more the leakage current than an electrolytic counterpart... at least, not in the lines that I can readily order from Digi-Key.

**Super Nade** · 12-12-2007, 10:38 AM

Re: many mobo's will benefit from polymer cap upgrades

My contribution.

http://www.overclockers.com/articles1494/index02.asp

I think I may have made a mistake in calculating the ripple current distribution. Any thoughts would be much appreciated!

**kc8adu** · 12-12-2007, 01:22 PM

Re: many mobo's will benefit from polymer cap upgrades

going to post a few i did too.
several other techs i sold these caps to are going to post their results.
these made a big difference on a bp6 for overclocking.
my neighbor was able to push another 10mhz on fsb!
yes i know several folks that still use bp6's

**Super Nade** · 12-12-2007, 03:08 PM

Re: many mobo's will benefit from polymer cap upgrades

Cool! Looking forward to picking up a few tips from ya bro.

**kc8adu** · 12-16-2007, 04:16 PM

Re: many mobo's will benefit from polymer cap upgrades

going to use the 820@2.5 fujitsu's on vrm out on a friends beloved abit bh6.
its still running ok but those crapcaps have to be suffering after all those years of overclocking 24/7/365!
i flogged it 4 years myself and he traded me out of it when i got my vp6.
i think it has tayeh and jackcons in it.
welcome to extreme makeover motherboard edition!

**Super Nade** · 12-17-2007, 07:42 AM

Re: many mobo's will benefit from polymer cap upgrades

LOL! cool!

I'm working on an Albatron board, which seems to be decent (Panny FJ, OS-CON, WG and crap GSC).

I'm keeping track of the info here:-
http://super-nade.blogspot.com/

And man, I'm gonna take you up on your offer of the 820 uF OSCONs. I'll let you know how many I need. I'm gonna use these for an experiment with a Gigabyte board I have laying around.

**kc8adu** · 12-17-2007, 08:40 AM

Re: many mobo's will benefit from polymer cap upgrades

dont see many albatron boards around here.
we will be watching for updates.
as for the 820@2.5 pm me when ready.
i posted a bunch in my google store.

**MD Willington** · 12-18-2007, 03:56 PM

Re: many mobo's will benefit from polymer cap upgrades

Originally posted by Super Nade

LOL! cool!

I'm working on an Albatron board, which seems to be decent (Panny FJ, OS-CON, WG and crap GSC).

I'm keeping track of the info here:-
http://super-nade.blogspot.com/

And man, I'm gonna take you up on your offer of the 820 uF OSCONs. I'll let you know how many I need. I'm gonna use these for an experiment with a Gigabyte board I have laying around.

COOL!, nice blog!

**gonzo0815** · 12-18-2007, 06:04 PM

Re: many mobo's will benefit from polymer cap upgrades

Regarding the ripple current calculation, AFAIK the ripple current depends on the VRM circuit, e.g. the number of phases and the way they are interleaved.
From my understanding, the ripple current is nearly constant, but the lower the ESR of the caps the less ripple will show at the O/P.

I would lock into the data sheet of the VRM controler chip, usually there are some formulas or rough numbers of the ripple current.

**PCBONEZ** · 12-18-2007, 06:37 PM

Re: many mobo's will benefit from polymer cap upgrades

Yes.
The lower the ESR is the easier it is for the capacitor to dump the ripple to ground.

For calculations the best thing to do is to use the ATX power supply specification and use the max allowed ripple as the number for the equation. (By doing this you are assuming the cap has to deal with the ripple from a worst case (crappy) power supply.)

.

**kc8adu** · 12-23-2007, 06:18 PM

Re: many mobo's will benefit from polymer cap upgrades

at last.
the bh6 has been hotrodded!
i used chemicon psa 820@2.5v for vrm out ,sanyo sepc 560@4v for most of the 1000@6.3v.the ones on +5 got rubycon mbz.
this board was running ok but cranky at times at stock speed with a 700/100 cumine.
it used to run ok at 112 but never at 124.
even stock recently the owner reported w2k lockups.
testing with mprime first showed ok stock,occasional illegal sumout at 112, and 124 it would instantly give about 100 illegal sumouts and reboot.133 nopost.
after.
133 post but lock.
124 has been running mprime about 4 hours without error.
will look tomorrow on the way home.
btw pics coming.
i replaced the 100@16's on vdimm with sanyo sepc 560@4 also.