Just do a forum search for recapping - it's a well-covered topic. Here's some to get you started:

Is it OK/safe to change electrolytic caps for solid-states?

Capacitor replacement

http://www.hardwareinsights.com/wp/poly-moddi … -actually-work/

Thanks bud

I have a question about caps, I might aswell ask here.

As Im not that keen on paying 1 euro each for top grade Nichicon/Sanyo caps I did some searching at Mouser.

Panasonic caps seems cheaper and have a good reputaion. The FM series seem to fit the bill but I find two simliar caps for each one I need, are there any diffrences I miss? Edit I just realized that the only difference is the packaging? /Edit

10V. 1000uF. 105C. low ESR. (I need 100+ of these) Panasinic FM or some other brand?
http://www.mouser.se/Panasonic/Passive-Compon … 1yx4aw4Z1z0t6dh

10V. 1500uF. 105C. low ESR. (I need 100+ of these) Panasonic FM or some other brand?
http://www.mouser.se/Panasonic/Passive-Compon … 1yx4aw4Z1z0wfbt

I think I can use 6.3V caps in most cases but as many boards are equipped with 10V ones I rather go for a larger number of 10V caps to get a lower price. I need some 16V caps for some newer boards and PSUs and also some low voltage ones but I do not need huge numbers of those. Up until now I have mostly recapped bulging caps using used caps from scrap boards with more or less the exact same values as the original but if I'm going to do full recaps I'm not going bother with using old scavenged caps. When ordering new caps price is really an issue for me as I have ... many boards in need of new caps. I know about the issue with too fat caps but in many cases it can be solved with some creativity.

I'm in no hurry to order caps as Im in the middle of moving my workspace from one room to another which will take another month or two before everything is setup as I want it but it never hurts to start making lists of components I need to buy.

What are you re-capping exactly?

You can get away with using Panasonic FM or FR in anywhere BUT the VRM low and high...

On P3 boards, FM is also good for the VRM. Anything P4 and later, either use UESR caps like Nichicon HM/HN/HZ, or do a poly-mod.

mockingbird wrote:

What are you re-capping exactly?

You can get away with using Panasonic FM or FR in anywhere BUT the VRM low and high...

On P3 boards, FM is also good for the VRM. Anything P4 and later, either use UESR caps like Nichicon HM/HN/HZ, or do a poly-mod.

We are talking about mostly i440BX Slot-1 boards and some late Socket-7 boards at this stage. Also my Supermicro Super S2DG2 dual Slot-2 board and my Supermicro P6DGU dual Slot-1 board, perhaps also my Tyan S1834 dual Slot-1 board if I can find some use for it.

I will get to the P4 and newer boards later but as it is I have too many spare boards of the newer sockets so I will not do any recaps on those that will cost me more than 5 euro each. I can just aswell pick up more boards with ok caps with minimum bids on the local Swedish auction site, Socket-478 and Socket-A boards seem to be worth nothing around here except for the top models like the Asus P4C800-E Deluxe.

Yes, FM is fine for those. You can also use Nichicon HD/HV, Panasonic FR, UCC KZE/KZN, Sanyo/Suncon WX, Illinois Capacitor KXM, etc...

mockingbird wrote:

Yes, FM is fine for those. You can also use Nichicon HD/HV, Panasonic FR, UCC KZE/KZN, Sanyo/Suncon WX, Illinois Capacitor KXM, etc...

Good, I will probably use Panasonic FM or other cheap but high quality caps where ever I can get away with it. On a motherboard like the Supermicro Super S2DG2 with 41 caps to replace there is a huge difference between paying 0.2 euro and paying 0.5 euro for each cap.

I did find Nichicon 6.3V HZ for a similar price compared to the 10V Panasonic FM but then again I could aswell go with 6.3V Panasonic FM which are even cheaper. It would not hurt to order a few Nichicon HZ both 6.3V and 16V though in case I should need them for a newer board.

http://www.mouser.se/Nichicon/Passive-Compone … 1z0vhz5Z1z0ynq8

The minimum amount you need to buy of the Nichicon HM series makes them less attractive.

http://www.mouser.se/Nichicon/Passive-Compone … 1z0vhz5Z1z0ynqb

And the Nichicon HN series seems too expensive.

http://www.mouser.se/Nichicon/Passive-Compone … 1z0vhz5Z1z0ynqa

Some Panasonic FM 6.3V/10V/16V 1000uF and 1500uF caps of interest.

http://www.mouser.se/Panasonic/Passive-Compon … 1z0wrjwZ1z0wrk7

The links are mostly so I do not have to use Mousers search again.

Yes and keep in mind that FM-class caps have a better endurance than UESR electrolytic series caps.

They are lower ESR because they are more aqueous, the tradeoff is that they're all rated for 2000 hours (the exception is Sanyo/Suncon WG which is rated for 3000 hours).

So only use them when you have to.

Wait, 2000 hours? As in two thousand? So for a computer that is on 24 hours a day the caps will only last 83 days? Or am I totally misunderstanding what you are saying?

2000 hours at 105C... So there's a significant gain in longevity the lower the temperature.

Rule of thumb: 2x the longevity every -10 degrees C

Oh I see, that makes a lot more sense.

Skyscraper wrote:

We are talking about mostly i440BX Slot-1 boards and some late Socket-7 boards at this stage. Also my Supermicro Super S2DG2 dual Slot-2 board and my Supermicro P6DGU dual Slot-1 board, perhaps also my Tyan S1834 dual Slot-1 board if I can find some use for it ...

^ For that stuff's VRM caps, Panny FM will do. Nichicon HM (lower ESR) would be even better.

. . .

Nichicon HN, even better. Nichicon HZ, even better (lower ESR).

Once fully recapped with good jap lythic caps, and provided good cooling and decent PSU, you can expect a decade or more of longevity.

TELVM wrote:

^ For that stuff's VRM caps, Panny FM will do. Nichicon HM (lower ESR) would be even better. […]
Show full quote

^ For that stuff's VRM caps, Panny FM will do. Nichicon HM (lower ESR) would be even better.

. . .

Nichicon HN, even better. Nichicon HZ, even better (lower ESR).

Once fully recapped with good jap lythic caps, and provided good cooling and decent PSU, you can expect a decade or more of longevity.

I wouldn't mind seeing some overclocking benchmark comparisons with UESR caps on pre-P4 VRMs.

Correct me if I'm wrong, but Slot1 or Socket370 CPUs are being fed directly by either the +5V or 3.3V rails. There is no Pi filter like 12V VRMs for P4s and later.

If that's true, then I'd have to challenge the wisdom of using HN or HZ on such a board. Again, you are getting a cap with much better endurance when you go HD/HV/KZE/FM/FR class. And their ESR is already pretty low, so again, the only time it would be justified is if the overall lower ESR might help with overclocking, which I doubt.

What is the reason for Panasonic caps not being recommended for VRM on P4 / Athlon 64? They don't make low ESR capacitors?

I have two AOpen P4 boards with Japanse capacitors. Interestingly they have a mix of brands. One has Panasonic and Rubycon, the other one Panasonic and KZE or something similar sounding.

Farnell has been advertising some new Panasonic Polymer caps with very low ESR. And there's always the the magnificent Nichicon L8 Polymers as well. Low ESR on these Nichicons as low as 0.008 Ohms.

PhilsComputerLab wrote:

What is the reason for Panasonic caps not being recommended for VRM on P4 / Athlon 64? They don't make low ESR capacitors?

I have two AOpen P4 boards with Japanse capacitors. Interestingly they have a mix of brands. One has Panasonic and Rubycon, the other one Panasonic and KZE or something similar sounding.

Panasonic caps were often used on P4/Athlon 64 boards and later... And for that matter, Asus even used UESR caps on Athlon XP boards (Unfortunately they used Nichicon HM from a time when they were a defective series, so you always see them bloated. The A7N and A7V series are a good example of this). But the series used were FJ or FL... The problem with FJ and FL is that they were never sold directly to the public.

These days, you're limited to the diminishing supply of UESR caps from Nichicon, who stopped production last year or so. Technically, Sanyo/Suncon will still manufacture WG series, but I can't confirm that. It's either that, or you use Polymer caps like carlostex just mentioned.

You will have to drop the capacitance though, especially on the VRM high where you always need a 16V cap for the 12V input. So typically, for VRM high caps where there's space for only an 8mm part, you use something like 16V 270uF (whereas you'd use 16V 1000uF with an electrolytic). The way polymer cap technology works, you don't gain capacitance with height, so even taller 16V 8mm caps will not be much higher than 270uF. I think 470uF is the maximum at 16V in 8mm. For the VRM low, you can get 2.5V 1500uF caps in 8mm. 2.5V caps in the VRM low are always acceptable replacements for 6.3V parts because the CPU voltage will never go above 2 volts.

Having said that, the theory behind using polymer caps in these types of VRMs is that you lose nothing by halving the capacitance because lower ESR is more important.

Ok this starting to make sense now.

To summarise, if a polymer replacement has significantly lower ESR, then I get away with using a lower uF. Voltage has to match though.

Voltage should be at least 16V for the VRM high caps which are taking the 12V from the PSU (caps come in degrees of voltage. 2.5v, 4v, 6.3v, 10v, 16v... That's why I say to use 16V for a 12V power rail) and as low as 2.5V for the VRM low caps which will never see more than 1.5-2.0 volts from the CPU. So where you normally find 6.3V caps for the VRM low, you can use 2.5V or 4V polymers.