How to test for bad caps?

[Nagoshi]

Just noticed my DFI has a bad cap. Id like to know if there are others on the way out.

Also, anything special I should take in consideration before changing a cap?

[davidhammock200]

Originally Posted by Nagoshi

Just noticed my DFI has a bad cap. Id like to know if there are others on the way out.

Also, anything special I should take in consideration before changing a cap?

Look at them closely under magnification with high intensity light, if you see bulges, cracks, splits or discoloration, they are bad.

When one cap has failed in any section, it is reasonable to suspect all of the major caps in that section, the usual practice is to just replace them all.

[onewecallgod]

You can't really test a cap short of visual inspection, but you if you're going to change one, might as well change them all.

Make sure the new cap(s) are similar in radial size (I'm assuming you're changing electrolytics), capacitance, lead pitch and come from a reputable company/country of origin. Going with a higher voltage isn't a bad idea as long as you don't forget your space constraints.

[Nagoshi]

Originally Posted by davidhammock200

Look at them closely under magnification with high intensity light, if you see bulges, cracks, splits or discoloration, they are bad.

When one cap has failed in any section, it is reasonable to suspect all of the major caps in that section, the usual practice is to just replace them all.

Yeah Im actually considering recapping the entire board... there might be other caps going bad, I will have to pull everything from the case and check it all out.

No wonder why the board acted so strangely recently...

[davidhammock200]

Originally Posted by onewecallgod

You can't really test a cap short of visual inspection, but you if you're going to change one, might as well change them all.

Make sure the new cap(s) are similar in radial size (I'm assuming you're changing electrolytics), capacitance, lead pitch and come from a reputable company/country of origin. Going with a higher voltage isn't a bad idea as long as you don't forget your space constraints.

Always try to go with both a higher voltage & temp rating, making sure they will fit.

[Nagoshi]

Originally Posted by onewecallgod

You can't really test a cap short of visual inspection, but you if you're going to change one, might as well change them all.

Make sure the new cap(s) are similar in radial size (I'm assuming you're changing electrolytics), capacitance, lead pitch and come from a reputable company/country of origin. Going with a higher voltage isn't a bad idea as long as you don't forget your space constraints.

Yeah Im changing the "taller" caps, not the smaller "solid" ones. Its an older board (DFI LP UT nF3 250Gb, socket 754).

Id go with similary spec'ed hi-quality caps, but do you mean I can go with higher voltage caps and it might actually help the board?

Additional Comment:

Originally Posted by davidhammock200

Always try to go with both a higher voltage & temp rating, making sure they will fit.

Thanks, will look for same specs with higher volts and temps ratings for my next caps

[onewecallgod]

Originally Posted by Nagoshi

Id go with similary spec'ed hi-quality caps, but do you mean I can go with higher voltage caps and it might actually help the board?

It shouldn't help, assuming the board was designed correctly, but it doesn't hurt to increase the margin a bit more. Also make sure the cap you buy specifies low ESR (equivalent series resistance).

[davidhammock200]

All caps will eventually fail, they must as that is how they are designed, but the higher the volt/temp rating the longer it takes to fail, assuming equal quality.

If you want a primer on cap quality read some of OKWolf's PSU reviews at http://jonnyguru.com

Best Wishes,
Dave

[onewecallgod]

Originally Posted by davidhammock200

the higher the volt/temp rating the longer it takes to fail

In case anyone is wondering, voltage rating in itself has no bearing on longevity of a cap since it just means the parallel plates are a bit thicker, but high voltage caps last longer because they're meant to deal with higher temperatures. A high temperature rated cap has the improvements to make a cap more durable, such as a more stable electrolyte solution to prevent and a stronger, environmentally resistant shell.

So comparing a cap that's rated for 16V and one that's 20V given all else equal probably won't have measurably different MTBFs.

[davidhammock200]

Originally Posted by onewecallgod

In case anyone is wondering, voltage rating in itself has no bearing on longevity of a cap since it just means the parallel plates are a bit thicker, but high voltage caps last longer because they're meant to deal with higher temperatures. A high temperature rated cap has the improvements to make a cap more durable, such as a more stable electrolyte solution to prevent and a stronger, environmentally resistant shell.

So comparing a cap that's rated for 16V and one that's 20V given all else equal probably won't have measurably different MTBFs.

Of course it will! In real life, if both caps are of the same design & manufacture & both are used side by side in a 12V circuit,
on average the 20V will have a longer MTBF.

[Nagoshi]

The cap that went bad and leaked was a 1000uF 6.3v from Michicon? Its black sleeved. 105c rated, A0506, HM(M) whatever that means.

Looking at the whole board shows some more bad caps, inc. one by the RAM slots, which might explain why my RAM was not working right.

So as long as the uF rating is the same, the voltage and temperature rating can go higher, as long as it physically fits the board, right? Is the color of the sleeving any important or it's only about the brand?

[davidhammock200]

Originally Posted by Nagoshi

The cap that went bad and leaked was a 1000uF 6.3v from Michicon? Its black sleeved. 105c rated, A0506, HM(M) whatever that means.

Looking at the whole board shows some more bad caps, inc. one by the RAM slots, which might explain why my RAM was not working right.

So as long as the uF rating is the same, the voltage and temperature rating can go higher, as long as it physically fits the board, right? Is the color of the sleeving any important or it's only about the brand?

Yes keep the same uF, but you can go "reasonably" higher on voltage, but 105C is pretty much top o line & is more than enough in a well cooled environment.

After reinstalling that mobo, I recommend getting a small fan to cool the area around where the caps died.

[Nagoshi]

Originally Posted by davidhammock200

Yes keep the same uF, but you can go "reasonably" higher on voltage, but 105C is pretty much top o line & is more than enough in a well cooled environment.

After reinstalling that mobo, I recommend getting a small fan to cool the area around where the caps died.

Good suggestion on the fan, altho it's hard to do since it's around the passively cooled chipset area and there are connectors in the way. Hmm, I will have to make something up!

Is there anything else I should be aware of? Known good and bad cap brands? I dont feel like getting ripped off and getting lower end caps in the end, as I would like the board to regain its original overclocking abilities

[onewecallgod]

Originally Posted by davidhammock200

Of course it will! In real life, if both caps are of the same design & manufacture & both are used side by side in a 12V circuit,
on average the 20V will have a longer MTBF.

Do you have proof? Or better yet, do you have an explanation as to why the higher voltage spec will last longer? The plates don't fail, it's the venting and boiling of electrolyte that causes problems.

Originally Posted by Nagoshi

The cap that went bad and leaked was a 1000uF 6.3v from Michicon? Its black sleeved. 105c rated, A0506, HM(M) whatever that means.

So as long as the uF rating is the same, the voltage and temperature rating can go higher, as long as it physically fits the board, right? Is the color of the sleeving any important or it's only about the brand?

Are you sure it was Michicon and not Nichicon? Small matter, really, but if it says Nichicon, then that's pretty much a top of the line (in industrial terms) Japanese cap. The color is just a manufacturer's designation; it's not like a resistor where the color of the body actually means something.

[davidhammock200]

Originally Posted by Nagoshi

Good suggestion on the fan, altho it's hard to do since it's around the passively cooled chipset area and there are connectors in the way. Hmm, I will have to make something up!

Is there anything else I should be aware of? Known good and bad cap brands? I dont feel like getting ripped off and getting lower end caps in the end, as I would like the board to regain its original overclocking abilities

Yes, as I suggested above go to http://jonnyguru.com & read some of the PSU reviews, OKWolf always disassembles the PSU's after testing & discusses the design & construction, that is your easiest way to get "up to speed" on who makes what these days. It has been a number of years sinice I personally recapped a PSU & I am sure things have changed.

Best Wishes,
Dave

[Nagoshi]

Thanks for the link David - I just finished reading a DIY on replacing caps and it has me a bit worried actually LOL

http://www.jonnyguru.com/modules.php...ory4&ndar_id=8

They got more ripple in the PSUs after recapping them? That actually kinds of tingles me, lol

Yeah after looking it's Nichicon, the cap is small and could barely read it.

Should I try and grab new caps from the same brands, if possible with a higher voltage rating?

The other caps are brown and only have the specs on them, with either "KZG" or "KMG" on them. Ill figure they are also top of the line caps?

[davidhammock200]

Originally Posted by Nagoshi

Thanks for the link David - I just finished reading a DIY on replacing caps and it has me a bit worried actually LOL

http://www.jonnyguru.com/modules.php...ory4&ndar_id=8

They got more ripple in the PSUs after recapping them? That actually kinds of tingles me, lol

Yeah after looking it's Nichicon, the cap is small and could barely read it.

Should I try and grab new caps from the same brands, if possible with a higher voltage rating?

The other caps are brown and only have the specs on them, with either "KZG" or "KMG" on them. Ill figure they are also top of the line caps?

Good questions to ask at a DFI forum where someone will have the same mobo & can at least look for you.

Best Wishes,
Dave

[JEskandari]

I've bee told from someone who was an engineer in Iran TV Broadcasting that on a board if you find a capacitor which have failed it's wise to search around it for a failed Resistor.

[larrymoencurly]

Originally Posted by Nagoshi

The cap that went bad and leaked was a 1000uF 6.3v from Michicon? Its black sleeved. 105c rated, A0506, HM(M) whatever that means.

HM is the series (probably Nichicon's third-lowest ESR type for wet caps -- HN is lower, HZ the lowest), (M) is the package type, I don't know what "A" means, but I think 0506 is the year 2005, week 6. Nichicon HMs are considered really good, unless they were made around 2003-2005 (lots of Dells and Macs from those years affected).

Everybody mentions capacitor dealers like Mouser, Digi-Key, and Newark, but there are also:

BadCaps.net
CapsMod.net
TheCapKing.
PC Motherboard Capacitor Store (eBay, probably the cheapest on shipping for small quantities)
B&D Enterprises (lots of Sanyo, some Nichicon, pretty cheap)

Be sure to check the actual ESR numbers from the data sheets of the original capacitors and the substitutes because verbal descriptions can be really imprecise, and one company's "low ESR" cap could have such high ESR that it will run hot.

Over at BadCaps.net, they don't seem to like Chemicon KZG caps and say they often fail without bulging, and because KZJ is chemically similar (but has half the ESR), they don't like them, either. OTOH KZE is OK but has twice the ESR of KZG.

I've also had to replace caps around the memory slots, even when the memory voltage regulator was the linear type and those caps were located far from the regulator's heat.

[Nagoshi]

Originally Posted by larrymoencurly

HM is the series (probably Nichicon's third-lowest ESR type for wet caps -- HN is lower, HZ the lowest), (M) is the package type, I don't know what "A" means, but I think 0506 is the year 2005, week 6. Nichicon HMs are considered really good, unless they were made around 2003-2005 (lots of Dells and Macs from those years affected).

Everybody mentions capacitor dealers like Mouser, Digi-Key, and Newark, but there are also:

BadCaps.net
CapsMod.net
TheCapKing.
PC Motherboard Capacitor Store (eBay, probably the cheapest on shipping for small quantities)
B&D Enterprises (lots of Sanyo, some Nichicon, pretty cheap)

Be sure to check the actual ESR numbers from the data sheets of the original capacitors and the substitutes because verbal descriptions can be really imprecise, and one company's "low ESR" cap could have such high ESR that it will run hot.

Over at BadCaps.net, they don't seem to like Chemicon KZG caps and say they often fail without bulging, and because KZJ is chemically similar (but has half the ESR), they don't like them, either. OTOH KZE is OK but has twice the ESR of KZG.

Where could I find the data sheets of the original caps? I thought we could find most of the data on the caps themselves

I was thinking about kind of upgrading the caps but this is ending up more complicated than I thought

I would have to check but I believe half the caps on the boards are those black Nichicons and they are all the same sizes. Should I upgrade them to HZ caps, or something like KZG caps? The other half of the caps are brown colored with no brand marking on them, and there are about 5 different size of caps.