Can a PSU damage the motherboard capacitors?

[emy12]

Re: Can a PSU damage the motherboard capacitors?

I have the following caps on sli motherboard:
3x KZJ 1000uF 16v
5x Rubycon 3300uF 6.3v MBZ
29x KZG 1000uF 6.3v
16x E.CAP 100UF 10V SS
4x KZJ 6.3V 1800UF
Do think that those are not reliable?, or which once do you think should I replace?

[PCBONEZ]

Re: Can a PSU damage the motherboard capacitors?

KZJ should be replaced with something like HN or MCZ
KZG should be replaced with something like HM, MBZ, WG
.

[PCBONEZ]

Re: Can a PSU damage the motherboard capacitors?

If the 3x KZJ 1000uF 16v are on the 12v side of the VRM you could use 3x 470uF 16v Polymer there.
.

[mariushm]

Re: Can a PSU damage the motherboard capacitors?

KZG are known to fail without showing any signs...

[PCBONEZ]

Re: Can a PSU damage the motherboard capacitors?

KZJ too. - Same electrolyte.

[emy12]

Re: Can a PSU damage the motherboard capacitors?

Just for the time being i installed the GA-K8NF-9, it has better caps than sli one. On sli one it is a lot of soldering to be done and i do not have the right equipment for now, and I am not able to get the level of soldering of the original once.
What is amazing is, since i changed the caps with sanyo once the temps of the cpu dropped from 40-50 to 25-30, just watching tv online.
I did not know that the caps could cause higher temps of cpu.

[emy12]

Re: Can a PSU damage the motherboard capacitors?

It is ok to replace the KZJ 6.3V 1800uF 8x20 with Nichicon HN 16v 1800uF 10x25 ?
Nichicon has higher voltage and thicker/taller a bit.

[c_hegge]

Re: Can a PSU damage the motherboard capacitors?

It will work fine. The voltage rating is just the maximum it can handle. The diameter won't be a problem as long as you can squeeze them in.

[PCBONEZ]

Re: Can a PSU damage the motherboard capacitors?

The 10mm will have a 5mm lead spacing and the 8mm with have a 3.5mm spacing.
Not a big deal but you will have to bend the leads a little to fit right.

You don't want any side-pressure from the leads on the rubber bung as installed because that can create leaks along the leads over time.

Also don't pull against the rubber while bending.
Use a needle-nose or similar to keep the stress/pulling from the bending off the rubber.
.

[emy12]

Re: Can a PSU damage the motherboard capacitors?

Thank you both for help.

I will replace the KZJ 6.3V 1800uF 8x20 ---> Nichicon HN 16v 1800uF 10x25 or Nichicon LG Polymer 6.3v 1800uF 10X13mm (i think are ok)

If the 3x KZJ 1000uF 16v are on the 12v side of the VRM you could use 3x 470uF 16v Polymer there.
.

You mean this:



What about replacing with Nichicon LG Polymer 16v 820uF 10X13mm.
Its is ok?



By the way, it is normal for a PSU, ex:750TP Blue I have, to have a soft electrical buzzing noise which I would need to put my ear close to the PSU to hear it? 100% is not the fan, it barely spins, just 164 rot/min all the time.
Some experts say it is normal due to some “conversion” inside of it.

Thanks

[larrymoencurly]

Re: Can a PSU damage the motherboard capacitors?

An Antec 350W SmartPower's +5Vstandby circuit failed and fed 13V to the motherboard. The USB ports and everything plugged into them all seemed OK, but a USB mouse blew and the built-in Realtek sound no longer worked but was still recognized by the system.

This +5Vstandby circuit had its Teapo or Koshin capacitors replaced years ago with Sanyos, and a 2000uF, 10V Sanyo bloated. The circuit board turned brown in the area of its transformer.

[Wester547]

Re: Can a PSU damage the motherboard capacitors?

How is it that it partially survived being fed 13V but that, say, those Bestec ATX-250-12Es failing and feeding 12V to any motherboard fries the board instantly? Unless it has to do with the length of time that the board is exposed to that voltage. Just curious.

[mariushm]

Re: Can a PSU damage the motherboard capacitors?

The ATX specifications say any device must accept input voltages within +/-5% of the configured value.
So since the motherboard has to accept 11.4v - 12.6v, the circuit that takes 12v at the input is designed with the idea that it may receive up to 12.5-13v and every component in reality is chosen in such a way that it will tolerate it.
For example, the processor is powered using a switching regulator that has 16v rated capacitors at the input and mosfets that are rated for at least 15v (most modern mosfets can handle 20-40v easily)

In contrast, the motherboard expects 5v standby to be 5v give or take a few 0.1v and some components on the motherboard may be powered directly from that 5v standby, for example the chipset. Less often, there are linear regulators used to convert that 5v to 3.3v or 2.8v or 2.5v, for example to power up the BIOS flash chip and these linear regulators may be designed to handle only up to 6v.

Also, generally for 5v, motherboards use 6.3v or 10v rated capacitors, so when a power supply dies in such a way that it sends 12-15v on 5v, a lot of things can break down.

It's all about through which wires it got 12-13v.

[larrymoencurly]

Re: Can a PSU damage the motherboard capacitors?

How is it that it partially survived being fed 13V but that, say, those Bestec ATX-250-12Es failing and feeding 12V to any motherboard fries the board instantly? Unless it has to do with the length of time that the board is exposed to that voltage. Just curious.

I don't know how long the overvoltage lasted, but the LEDs of the USB mouse glowed normally (decorative blue one remained steady, red LED for the motion detector dimmed when not used and brightened when the mouse was moved), but the mouse pointer didn't move. I think I turned off the power within 2-5 seconds. The motherboard was an ECS GeForce 6100SM-M.

[c_hegge]

Re: Can a PSU damage the motherboard capacitors?

And that would be why you always replace the small <100uF caps. If the killer cap on the 5vsb circuit had been replaced when it was originally recapped, it wouldn't have happened.

[momaka]

Re: Can a PSU damage the motherboard capacitors?

How is it that it partially survived being fed 13V but that, say, those Bestec ATX-250-12Es failing and feeding 12V to any motherboard fries the board instantly?

Probably the fact that many of those killer Bestecs were fitted in E-machines with cheap, cost-cutting, poorly-designed motherboards. I have one - some socket 462 Fic (forgot the model # at this moment). It had a bunch of TTL SMD ICs hooked directly to 5VSB without any protection of any kind. The absolute maximum ratings for many of these ICs were 5.5V. No surprise the motherboard failed at all. All of these ICs were shorted when I found it.

Modern motherboards have many of their ICs running behind a linear regulator of some sort, thus being well-protected from over-voltage on all of the PSU rails. Older boards (P4 and older), on the other hand, often ran the SB directly off of 3.3V. In the case of SDRAM motherboards, it was the norm to have the RAM running directly off of the PSU's 3.3V rail (or through a MOSFET at the most). That's why even the cheapest of the cheapest PSUs regularly have the filtering section on the 3.3V rail at least a little more "beefy" than the others.

[Wester547]

Re: Can a PSU damage the motherboard capacitors?

Probably the fact that many of those killer Bestecs were fitted in E-machines with cheap, cost-cutting, poorly-designed motherboards. I have one - some socket 462 Fic (forgot the model # at this moment). It had a bunch of TTL SMD ICs hooked directly to 5VSB without any protection of any kind. The absolute maximum ratings for many of these ICs were 5.5V. No surprise the motherboard failed at all. All of these ICs were shorted when I found it.

Well, I'm not sure how poor the eMachines Socket 939 motherboards were anyway. I believe they were OEMed by MSI or/and Gigabyte which isn't too shoddy. Okay, those boards obviously came with crappy capacitors (KZGs) and had northbridges and southbridges with inadequate heatsinks that ran hot enough to burn your fingers, along with nVidia chipsets bearing the substrate issue that beleaguered everything from the GeForce 6xxx to 9xxx series. But, with a simple recap and fitting a 60mm fan to the northbridge and southbridge, and maybe turning off the "SmartFan" feature of the board so the fans run at +12V (if you don't mind the noise, but that "feature" has them running a bit too slow for my taste - less than half of their full speed until the CPU gets pretty hot, around 60*C, in my experience, at least at the default settings), that should solve most if not all of those issues. I can definitely vouch that they are better than anything ASUS has put out of relatively recent times. I'd say they cut more corners than any other OEM (of common knowledge at least).

Modern motherboards have many of their ICs running behind a linear regulator of some sort, thus being well-protected from over-voltage on all of the PSU rails. Older boards (P4 and older), on the other hand, often ran the SB directly off of 3.3V. In the case of SDRAM motherboards, it was the norm to have the RAM running directly off of the PSU's 3.3V rail (or through a MOSFET at the most). That's why even the cheapest of the cheapest PSUs regularly have the filtering section on the 3.3V rail at least a little more "beefy" than the others.

Yes, older boards make much more usage of the +3.3V rail. I believe the AGP 1x-2x slots ran directly off of +3.3V, at least v1.0 did (except for the GPU fans which of course ran off +12V, full speed or not), along with some PCI slots. I also believe that the Socket 8/387 Pentium II "Overdrive" CPUs ran off the +3.3V rail since their Vcore was +3.1 to +3.3V. I wonder how well protected the Socket 478 boards (made by Foxconn and Intel) are from overvoltage issues. I haven't been impressed upon much by the treatment of the output filter for the +3.3V rail in cheapo power supplies. I have noticed that the +3.3V rail tends to have a much smaller toroid with many less turns than the +5V/+12V toroid, in many power supplies, but that is understandable. I do like that Delta 300W-500W units tend to have a magamp for the +3.3V that's almost as large as the other toroid. I imagine that makes quite a difference for how clean the output is (even if it wouldn't be quite as clean as a linear regulated +3.3V circuit). I also really like the fact that 300W-400W Hipros tend to have larger ferrite coils in the PI filter circuits than usual. But back to cheapo power supplies, yes, I think one reason that cheap units get away with good +3.3V filtering is because they occasionally defer to linear regulation which is another sign of cheapness but also is equally convenient.

[socketa]

Re: Can a PSU damage the motherboard capacitors?

Interesting thread
i read it because this happened to my best motherboard, which i never even got to use.

Got me thinking about a close enough related issue:

Just checked the PC power button, and note that it is momentary switch - not on/off.
Also checked the 5vsb rail with a light bulb and was surprised to see it come on without the psu spinning.
i don't like the idea of power being supplied to the motherboard when the computer is not on.
When i push the PC power button, i want all power to be disconnected from the motherboard - is there there a mod for this, or will cutting power to the the 5vsb mobo rail achieve this result?

This means that i can't forget to turn off the power at the wall, otherwise there is the potential for wear and tear, or even damage
Sometimes i don't bother to turn the power of at the wall, which means that voltage (or perhaps even current) could be unnecessarily across some component.

If the 5vsb wire is removed from the psu, would the computer still start?,
or does it depend upon how old the mobo is?
Does the 5vsb come into use when i select start>turn off computer>standby?
i think that that is perhaps something different - because the psu fan is still spinning when the o/s is on standby.

[Agent24]

Re: Can a PSU damage the motherboard capacitors?

Interesting thread
i read it because this happened to my best motherboard, which i never even got to use.

Got me thinking about a close enough related issue:

Just checked the PC power button, and note that it is momentary switch - not on/off.
Also checked the 5vsb rail with a light bulb and was surprised to see it come on without the psu spinning.
i don't like the idea of power being supplied to the motherboard when the computer is not on.
When i push the PC power button, i want all power to be disconnected from the motherboard - is there there a mod for this, or will cutting power to the the 5vsb mobo rail achieve this result?

This means that i can't forget to turn off the power at the wall, otherwise there is the potential for wear and tear, or even damage
Sometimes i don't bother to turn the power of at the wall, which means that voltage (or perhaps even current) could be unnecessarily across some component.

If the 5vsb wire is removed from the psu, would the computer still start?,
or does it depend upon how old the mobo is?

In a standard ATX PSU the 5vSB (StandBy) rail is exactly that. It runs all the time while the PSU is plugged into the mains and the switch (if it has one) is on.

It supplies standby power to the motherboard, and when you press that momentary power button on your case, the motherboard sends a signal to the PSU telling it to come up to full power.

If you cut the 5vSB rail then your computer will no longer turn on, and the PSU would still be running anyway.

If you want a computer that doesn't have any power to it when off, either switch it off at the wall, or get an old Pentium 1 machine (or earlier) which uses an AT PSU that is switched by a big clunky mains switch and doesn't have a 5v Standby Rail.

Does the 5vsb come into use when i select start>turn off computer>standby?
i think that that is perhaps something different - because the psu fan is still spinning when the o/s is on standby.

Yes, if you are using S3 (Suspend to RAM) state. It sounds like you have your computer currently set to use S1 (Power-on Suspend). You can change this in your BIOS settings.

[momaka]

Re: Can a PSU damage the motherboard capacitors?

Well, I'm not sure how poor the eMachines Socket 939 motherboards were anyway. I believe they were OEMed by MSI or/and Gigabyte which isn't too shoddy.

The ATX-250 12E was mostly used during the socket 462 (AMD Athlon XP) and early socket 478 (Pentium 4) era. By the time they switched to socket 939, the ATX-250 12Z and ATX-300 variants took over - none of which had that 2-transistor 5VSB.

I can definitely vouch that they are better than anything ASUS has put out of relatively recent times. I'd say they cut more corners than any other OEM (of common knowledge at least).

Yes, ASUS boards don't just cut corners - they make them fully round. Their socket 939 boards were the worst, though. They always had severe discoloration due to overheating MOSFETs near the NB and/or RAM. Just check eBay.

Yes, older boards make much more usage of the +3.3V rail. I believe the AGP 1x-2x slots ran directly off of +3.3V, at least v1.0 did (except for the GPU fans which of course ran off +12V, full speed or not), along with some PCI slots.

3.3V, 5V, and 12V are present on all versions of AGP. The signaling voltage between NB and the AGP card is what has changed between the versions, though. In V1.0, it was 3.3V like you said, and usually directly from the PSU's 3.3V rail. In V2.0, it was 1.5V IIRC. Then finally 0.8V in V3.0.

I wonder how well protected the Socket 478 boards (made by Foxconn and Intel) are from overvoltage issues.

Depends on the age of the SB. If the SB is old enough to work with 3.3V, it probably is powered directly from the PSU's 3.3V rail. Haven't looked at much else on them, though.

Now I realise that I never bothered to look at how the RTC and hardware monitor chip are powered on any board. My guess is, those are the next most susceptible items to overvoltage damage.