the lowest end lowesr electrolytic will give that board more life than you will be able to use up 😀 For me personally its a waste of perfectly good Polymer caps, but its your board 😀
GA-5AX is an atx version of GA-5AA, AA diagram Re: 430VX or 580VPX schematics
Yours uses RC5051M (19A) instead of 5AA RC5057M (16A) DCDC controller https://www.alldatasheet.com/datasheet-pdf/vi … ILD/RC5051.html https://www.alldatasheet.com/datasheet-pdf/vi … LD/RC5057M.html
They recommend 8 44mohm 1500uf caps for Max load Coppermine designs. 5AA uses only 5 1000uF caps and calls it good, yours went with 8x 330uF 😀

bytesaber wrote on 2026-03-24, 03:05:

Are hybrid replacement caps a bad idea?

something rich&beauty 😀
And excessively, but... A cheaper alternative would be the 470/560 at 6.3V.

25V on 1–5V power lines is an old developer lifehack.
If you open any datasheet for electrolytic capacitors, you’ll notice that at equal capacitance, the ESR drops to somewhere between 63 and 100V, and then rise again.
Thus, this is simply an attempt to replace low-ESR capacitors with generic use. This is somewhat excessive for polymer/hybrid capacitors.

PS. The black capacitors on the yellow circuit board look great. Everything will be fine.

rasz_pl wrote on 2026-03-24, 06:07:

44mohm 1500uf caps

For those years, such a capacitor will be able to capable a maximum ripple current of 1.5 amps.
The specified 330 μF hybrid capacitor is rated at 2 amps per unit.
Everything will be fine.

First off, my first concern is if using hybrids could damage or make things worse for the board. I'll admit there's likely no way that I would ever perceive the difference of using new electrolytic, compared to something like hybrids or polys. I'm simply going for "fun" and good board health. At the same time, trying to learn a bit as well from comments here. I have minimal knowledge on circuit or DC power design.

rasz_pl wrote on 2026-03-24, 06:07:

the lowest end lowesr electrolytic will give that board more life than you will be able to use up 😀

Ok! 😀

rasz_pl wrote on 2026-03-24, 06:07:

GA-5AX is an atx version of GA-5AA
Yours uses RC5051M (19A) instead of 5AA RC5057M (16A) DCDC controller
They recommend 8 44mohm 1500uf caps for Max load Coppermine designs. 5AA uses only 5 1000uF caps and calls it good, yours went with 8x 330uF 😀

Does this mean the datasheet suggests Gigabyte did not use large enough capacitance originally, for the case of a newer (perhaps not yet released) CPUs at higher loads?
Therefore... if I use the 330uf hybrids it would possibly provide something more stable?

shevalier wrote on 2026-03-24, 07:34:

something rich&beauty 😀

Absolutely!

shevalier wrote on 2026-03-24, 07:34:

And excessively, but... A cheaper alternative would be the 470/560 at 6.3V.

Can you help me understand what is meant by this? Are you describing 6.3V 470uf / 560uf Hybrid caps as cheaper alternative to the 330uf 25v Hybrid caps I have? I did not want to go to a lower voltage than what was already used. So I am a bit uncertain why 6.3v would be a save alternative. Is it because the power line is only at most 5v?

shevalier wrote on 2026-03-24, 07:34:

25V on 1–5V power lines is an old developer lifehack.

25v, as in the factory caps? So the power lines for all these caps is only 1v - 5v? So using a 25v rated electrolytic is some kind of a hack that did what back then?

shevalier wrote on 2026-03-24, 07:34:

If you open any datasheet for electrolytic capacitors, you’ll notice that at equal capacitance, the ESR drops to somewhere between 63 and 100V, and then rise again.
Thus, this is simply an attempt to replace low-ESR capacitors with generic use. This is somewhat excessive for polymer/hybrid capacitors.

Does this tie into what you are describing with the "old hack"?

shevalier wrote on 2026-03-24, 07:34:

PS. The black capacitors on the yellow circuit board look great. Everything will be fine.

Yes... It is true that you've caught me with "it's not broken, don't fix it." I admit the board has been working. I decided to pull it apart from my retro system and recap it because... I have no reason other than I would like to.

So far from both of you, I'm hearing that it is safe to proceed. Is this true? I am absolutely no holding anyone to this. Just conversing. I may not even replace them. All learning here.

bytesaber wrote on 2026-03-24, 08:16:

Does this mean the datasheet suggests Gigabyte did not use large enough capacitance originally, for the case of a newer (perhaps not yet released) CPUs at higher loads?

the part Gigabyte used was designed with Intel Coppermine in mind as thats where most of the market and money was. SS7 was low margin low end non sexy business niche.
AMD-K6-III datasheet only lists some sparse data for 400 and 450 MHz parts calling for 10-12A 2.2 V and 12-13.5Amp 2.4 V.
2.2V ____________| 2.4V
400 MHz| 450 MHz| 400 MHz| 450 MHz
18.10 W| 20.20 W| 26.80 W| 29.50 W
older one lists 3.3V 233 MHz AMD-K6 at ~10A 30W

bytesaber wrote on 2026-03-24, 08:16:

Therefore... if I use the 330uf hybrids it would possibly provide something more stable?

there is only stable or unstable, cant be more pregnant 😀
looks like AMD parts are less demanding than any Coppermine design would be

bytesaber wrote on 2026-03-24, 08:16:

So using a 25v rated electrolytic is some kind of a hack that did what back then?

not hack, just wasting money. 6.3V caps are perfectly fine for 5V line for couple of service life years of a computer, manufacturers werent in a business of wasting money. Premium offerings advertising long life "military grade capacitors" and number of PCB layers came out ~6-8 years later.

bytesaber wrote on 2026-03-24, 08:16:

So far from both of you, I'm hearing that it is safe to proceed. Is this true?

Yes as long as you know what you are doind. Desoldering THT leads from big ground planes might be a pita and source of frustration for someone with no soldering experience (judging by desk on pictures thats not the case) and bad cheap tools depending on quality of pcb. If its your first time grab a scrap multi layer board and experiment a little.

bytesaber wrote on 2026-03-24, 08:16:

Can you help me understand what is meant by this? Are you describing 6.3V 470uf / 560uf Hybrid caps as cheaper alternative to the 330uf 25v Hybrid caps I have? I did not want to go to a lower voltage than what was already used. So I am a bit uncertain why 6.3v would be a save alternative. Is it because the power line is only at most 5v?

Not hybrid - "old school general use polymer" cap.
On motherboards of this type, the voltage across the ‘power’ capacitors does not exceed 5V.
Even third-rate manufacturers’ general use polymer capacitors will have an ESR of 10–15 mΩ at a capacitance of 470/560/680 μF × 6.3 V.
That is half the value of your hybrid capacitors.
8-10mm Diameter.
https://www.capxongroup.com/files/PS_Series.pdf

bytesaber wrote on 2026-03-24, 08:16:

So using a 25v rated electrolytic is some kind of a hack that did what back then?

Yes, this is an attempt to replace the missing low-ESR capacitors.
The General use 330μF 25V capacitors will have a similar ESR and ripple current to the 330μF 6.3V low-ESR capacitors. Moreover, they have the same case size (which is why they have the same current rating, incidentally).

Your hybrids have several advantages
1. You already have them on your desk
2. They’re black.
.
.
.

333. They’re more similar to the originals cap in terms of capacity and ESR

rasz_pl wrote on 2026-03-24, 08:56:

6.3V caps are perfectly fine for 5V line for couple of service life years of a computer

https://images.100y.com.tw/pdf_file/05-Samwha-WL.pdf
470u from 6.3V to 400V Try to compare ESR&Ripple.

bytesaber wrote on 2026-03-24, 03:05:

Majority of them are all the same VENT brand 330uf capacitor. They appear to be rather generic. This includes any near the CPU.

They are not "VENT" branded, they have a vent on top as a pressure relief in case of failure. They are Jenpo CE-WR 25V 330µF - rated for 105°C. I'd agree that these were not low ESR caps, but just normal ones. Using a higher voltage rating will result in a lower ESR. So they apparently just went with fitting capacity and higher voltage general purpose caps to save a buck by not using specifically low ESR ones. Nothing too uncommon.

shevalier wrote on 2026-03-24, 09:16:

https://images.100y.com.tw/pdf_file/05-Samwha-WL.pdf
470u from 6.3V to 400V Try to compare ESR&Ripple.

yes more better. But its not a race car where more wins, this is a product where sufficient wins.

>tehsiggi

Oh now I get the 25V talk, thats hilarious. Proves the point even stronger any caps will do in this design if you fit enough of them 😀

I'm fairly certain they got the Jenpo ones in bulk for a pretty good deal.. and if the ESR of them is good enough, they have saved a buck in the production. So they didn't care what voltage rating.. it would probably have been more expensive to get the 6.3V low ESR variants

rasz_pl wrote on 2026-03-24, 09:23:

shevalier wrote on 2026-03-24, 09:16:

https://images.100y.com.tw/pdf_file/05-Samwha-WL.pdf
470u from 6.3V to 400V Try to compare ESR&Ripple.

yes more better. But its not a race car where more wins, this is a product where sufficient wins.

sufficient wins.

Yes. Except this table is for "Extremely Low Impedance" caps and here we have Gigabyte plopping random ones and it still worked fine for probably at least 5 years before shelved.

rasz_pl wrote on 2026-03-24, 10:28:

Yes. Except this table is for "Extremely Low Impedance" caps and here we have Gigabyte plopping random ones and it still worked fine for probably at least 5 years before shelved.

The table covers behavior of virtually all electrolytic capacitors.
It’s just that this datasheet happened to be the first one I came across.
If you fancy having a moan, you could check out a few other datasheets. 😀

rasz_pl wrote on 2026-03-24, 08:56:

Desoldering THT leads from big ground planes might be a pita and source of frustration for someone with no soldering experience (judging by desk on pictures thats not the case) and bad cheap tools depending on quality of pcb. If its your first time grab a scrap multi layer board and experiment a little.

Definitely agree! It's easy to do a bad job and damage VIAs. I'm certainly capable of mistakes, but I can carefully remove them. I use some nearby hot air to help soak the ground. Add some some leaded solder and use a desoldering gun. Gently remove. Repeat a bit here and there. That's about all I can do.

tehsiggi wrote on 2026-03-24, 09:19:

They are not "VENT" branded, they have a vent on top as a pressure relief in case of failure. They are Jenpo CE-WR

Ah I didn't realize that. Now I know! Thank you.

To everyone,
This has been a really helpful thread for me. I really appreciate how much is being taught. Thanks for this! It's been rather enlightening on selecting capacitors of different types, what manufacturers did to bring down ESR, and how this all can be considered when doing something like a recap. There's another thread here that didn't go quite the same. So I appreciate the guidance. I'll see how this weekend goes with a recap.