Gigabyte K8VM800M - Attack of the bad KZGs continues

[Th3_uN1Qu3]

Got this in tonight from my first floor neighbor. It stopped booting into Windows, would display random lines in text mode instead of the XP bootscreen. He thought the onboard video could be going out, but a look at the board revealed something else.

I know this is a low-end board but that's some serious cost cutting. The VRM has all of THREE caps. All three 3300uF 6.3v KZG, all three had gone bad. 1.3 ohm, 0.74 ohm and 100 mOhm. If the third cap had gone any worse it's likely that the thing wouldn't have even POSTed anymore. It was only running a Sempron 2800+ CPU, not some power hungry dual core. And this is a well-cared for machine, i mean just look at it, there's almost no traces of dust, that's a low power CPU and the caps still failed. The whole board is KZG btw, with a couple tiny Teapos thrown in the mix, but knock on wood, the rest of them seem to be fine.

I was out of 3300uF caps of any kind but i found something else in my cap box, take a look at the pics. Booted up first time, passed IntelBurnTest, so we can call this another successful poly mod. The OSCONs have 17mOhm ESR each, so with 4 of them i get 4.25 mOhm. The three KZGs (assuming they ever met their spec of 12mOhm) woulda been 4 mOhm. The difference is insignificant. It doesn't seem to complain about the change in capacitance either.

You should see the insides of that joke of a power supply, but that's for another thread... Anyway, it'll probably stay in spec given the consumption of a Sempron with onboard video, so it doesn't worry me that much. It also has a partially seized fan, but i ain't fixing that for free. I'll suggest him to get another power supply, whether he takes my advice or not isn't my business.

[Uranium-235]

Re: Gigabyte K8VM800M - Attack of the bad KZGs continues

nice

good thing is this kind of shit is slowly but surely stopping. I have a hard finding any new board that dosen't have AT LEAST the the (output) vrm poly'd

[goodpsusearch]

Re: Gigabyte K8VM800M - Attack of the bad KZGs continues

KZG are failing like crazy here too! It is like the electrolyte breaks down whether the caps are kept cool or not, working hard or not.. Just a matter of time.

[c_hegge]

Re: Gigabyte K8VM800M - Attack of the bad KZGs continues

Yeap. I have a Gigabyte GA-M56S-S3 board to fix. It had mostly KZGs, with a couple of small Nichicon HMs. A few of the ones on the VRM out were bulging. I polymodded the VRM with 470uF Nich LF on the VRM in and 1500uF OSCON SPEC on the VRM out. I'll replace the rest with Nich HM when they come. I'll also put a 40mm fan on the chipset heat sink, since it has an nVidia chipset.

[Uranium-235]

Re: Gigabyte K8VM800M - Attack of the bad KZGs continues

nvidia chipset?

put it out of its mysery right now

[RJARRRPCGP]

Re: Gigabyte K8VM800M - Attack of the bad KZGs continues

I just noticed a font I never noticed before on one of the KZGs.

[Th3_uN1Qu3]

Re: Gigabyte K8VM800M - Attack of the bad KZGs continues

Good news. Got paid for the cap swap, AND got a free IBM Netvista Pentium 4 machine on top of that. He told me smoke came out of it. It's most likely the weird form factor Hipro PSU as the mobo is Gigabyte and looks okay.

[momaka]

Re: Gigabyte K8VM800M - Attack of the bad KZGs continues

I know this is a low-end board but that's some serious cost cutting. The VRM has all of THREE caps. All three 3300uF 6.3v KZG, all three had gone bad. 1.3 ohm, 0.74 ohm and 100 mOhm. If the third cap had gone any worse it's likely that the thing wouldn't have even POSTed anymore. It was only running a Sempron 2800+ CPU, not some power hungry dual core. And this is a well-cared for machine, i mean just look at it, there's almost no traces of dust, that's a low power CPU and the caps still failed.

According to cpuworld, AMD Sempron 64 2800+ for socket 754 is rated for 59W TDP. It's not power hungry indeed, but it's not a tame dual-core Intel either. That aside, I think the problem has more to do with the cost-cutting like you said. I've seen better filtering on many old Pentium II motherboards. Heck, 3 caps is not even cost-cutting... it must be a joke really .

I see you didn't change the caps on VRM high side... hope they last somewhat otherwise your neighbor might come back. Bad PSU + heat from MOSFETs + heat from CPU can dry them out fairly quickly. Given that the board still posted with 3 bad caps on the VRM low, at least it doesn't seem like a picky board.

[Th3_uN1Qu3]

Re: Gigabyte K8VM800M - Attack of the bad KZGs continues

I see you didn't change the caps on VRM high side... hope they last somewhat otherwise your neighbor might come back.

I don't think i've ever had to replace any of those yet.

[momaka]

Re: Gigabyte K8VM800M - Attack of the bad KZGs continues

They do fail, but rarely.

I had to replace them (3x 16V 1800uF Nichicon HN with bad date codes) in a GX270 motherboard. Also seen many bad Teapo in that section of the Asus P4SD motherboards found in HP DC5000 and D530's SFF computers. But those things are mini-ovens that cook Rubycons and Nichicons too, so I guess not much surprise there.

The 3x 6.3V 3300uF Ost RLX on VRM high in my ECS P4VXASD2+ run scorching hot when the CPU goes to 100% usage. In contrast, the 4x 6.3V 3300uF Ost RLX on VRM low run very cool. Similar thing can be observed on a Jetway N2PAP-LITE I have (it still has the original GSC ME caps in it, too).