GA-7DPXDW-P Rev 1.0 fails to post

[shimjocky]

Have had this server humming away faithfully for the better part of 5 years now. Had to bring her home from the data center recently after a sudden failure that could not be easily fixed.

It will power up but will not post. No vid signal according to the monitor. Power light comes on and hard drives show activity. Also for some reason it will not power off anymore using the power button, have to kill the psu. PSU tests good. Removed all extras (hard drives, vid card, memory) but still no post or beeps.

I looked it over and quickly noticed some obviously bulging caps. There are two rows of six caps, one row right next to each cpu. Before becoming a member of this fine site I would never have noticed them.

So, could these caps, assuming they are bad, be the culprit here?

[370forlife]

Re: GA-7DPXDW-P Rev 1.0 fails to post

Most likely the problem.

Make note that these are United Chemi-con Capacitors, which are usually good, but a certain series had a defect that made them absolutely hate heat (I can't remember the series, a more knowledgeable member will fill in.)

Also, is that a piece of a broken off agp card in the agp slot?

[shimjocky]

Re: GA-7DPXDW-P Rev 1.0 fails to post

Most likely the problem.

Make note that these are United Chemi-con Capacitors, which are usually good, but a certain series had a defect that made them absolutely hate heat (I can't remember the series, a more knowledgeable member will fill in.)

Also, is that a piece of a broken off agp card in the agp slot?

Thanks for the reply. Would bum me out to know that they failed from heat as I totally overcooled this thing. It has, in addition to others, two vantec tornado's moving air through the 2u chassis. I was told I had the loudest server in the datacenter,lol. Will look at replacing those caps and will probably use some panny's.

As for the AGP slot, thats some sort of spacer you are seeing there. I never used the slot so I never paid it much mind. Since this is a server it just had an old MX200 pci card in it.

[Toasty]

Re: GA-7DPXDW-P Rev 1.0 fails to post

Check the PSU also. Although you don't mention what type/brand it is, it probably has the same problem.

Check the badcaps thread/sticky on the main page to identify what series' were bad.

The spacer is if you use an AGP Pro card. You remove it to use it.

[severach]

Re: GA-7DPXDW-P Rev 1.0 fails to post

Would bum me out to know that they failed from heat as I totally overcooled this thing.

The fault is with the improperly formulated or weak caps, not you. The caps are supposed to last even if you don't over cool. Your new caps will outlast the other OST and G-Luxon (WAG) caps.

You'd have noticed them but may not have considered them a likely suspect because "They're supposed to be like that, right?"

[hardwareguy]

Re: GA-7DPXDW-P Rev 1.0 fails to post

The United Chemicon caps that hate heat are the KZG and KZJ series....the others are OK.

With that many failed caps, Im not surprised it won't POST.

As others have stated, check your power supply....its likely in a similar state of disrepair. Check the chipset fan on that board too...Gigabyte was notorious for nasty fans back then. Any 40mm fan with at least a few CFM will work. It doesn't need much air to stay cool.

The spacer is there to prevent a regular AGP card from being inserted improperly into the AGP Pro slot if the unit was put together without a case for testing.

AGP Pro is normal AGP but with some extra power pins for high power cards. Gaming cards didn't use it, it was for workstation cards only such as the 3D Labs Wildcat series.

[gg1978]

Re: GA-7DPXDW-P Rev 1.0 fails to post

I've got one of these boards, but it had another fault, the VRM's wouldn't turn on.. Had to "hotwire" the VID bits to fix it. My board has Rubycon MBZ in the VRM though, instead of UCC..

[hardwareguy]

Re: GA-7DPXDW-P Rev 1.0 fails to post

I've got one of these boards, but it had another fault, the VRM's wouldn't turn on.. Had to "hotwire" the VID bits to fix it. My board has Rubycon MBZ in the VRM though, instead of UCC..

Indeed...later Gigabyte boards had Rubycon MBZ all over the place except for the 12V input side of the VRM which was Nichicon HM/HN. (no worries...they weren't the sized that popped, I've never seen one dead)

So you had to manually give the VRM controller the right bits....what do you think failed? VID pin in the CPU socket not connected any more? VRM controller gone nuts?

[KeriJane]

Re: GA-7DPXDW-P Rev 1.0 fails to post

Hello.

I've been using ECS boards for budget builds for a few years with no troubles... mostly the 945GCT-M boards with Celeron 430 (Fry's combo)
I generally use Seasonic-built Antecs for PSU, which are full of OST, not the dreaded Fuhjyyu.

I've noticed a wide variety of caps on these boards... generally the few polymer caps are a "CS" brand which appears to be affiliated with Elite. (yuk)
Usually, the boards have a lot of OST except around the CPU where either NCC or more commonly "TK" brands are found.

Recently, I've begun to preemptively recap the "TK" 16v and 4v caps near the CPU with Rubycon MBZ. Comparing the caps off the board with an ESR meter shows the TK brand to be higher in ESR than Rubycon, measuring .01-.03Ω instead of the 0-.01Ω on a new MBZ (1800uf/16v used for comparison).

In particular, the few tiny 680uf/4v caps on vCore do not inspire confidence. Shouldn't there be some polymers here? Fortunately the Intel-chipset boards often have tiny NCC caps that fit under the stock Intel heatsink, but I just had an ATI-chipset/AMD CPU build with 6 of those "TK" 680/4 in vCore. Fortunately, there was enough room for Rubycon MBZ 820/6.3 which seem to work fine.

Does anyone have a suggestion for vCore polymer caps to replace these tiny 680uf/4v electrolytics?
Should one fill the "empty" cap slots at vCore as well?
Should "CS" brand Polymers be replaced?

I have a feeling that much of ECS's poor reputation was earned right here at vCore. Of course a lot of it is probably due to cheap PSUs being used with these cheap boards... you should probably have decent caps in one or the other if not both. Even PC Chips boards work OK with a decent PSU.

Though I have never had a problem with ECS, their reputation is awful. I particularly like that the boards have no fancy gimmicks like dual BIOS, Overclocking or other such humbug that just gives more opportunity to go wrong or confuse the user.
Their one great feature is on the Driver CD... auto-install of all the boards drivers in one step.

Surprised at the relative goodness of ECS,
Keri

[hardwareguy]

Re: GA-7DPXDW-P Rev 1.0 fails to post

Such is the difference between an OEM board and a board aimed at the enthusiast market.

In defense of the enthusiast boards I must say I have found them useful in day to day computing.

Dual BIOS can save you a LOT of trouble.... it was originally made to guard against bricking due to malicious writes to the firmware. (think CIH virus back in the day)

If a customer updates the BIOS and fouls it up, they are REALLY happy when I help them recover over the phone instead of having days of downtime.

The overclocking stuff, well that's useful too. Even my systems I sell to non enthusiasts have some degree of control, maybe not as much as the gamer boards, but some things can be tweaked. Its REALLY nice being able to change the memory voltage +/- 0.3V..... This ability gives you much more flexibility when it comes to memory, particularly with customers that just bought some quirky RAM.

The classic example is someone that buys EPP (enhanced performance profile) RAM and their board doesnt know what to do with the weird SPD and it applies 1.8V to a 2.1V stick and the system is unstable as a result....one phone call can rectify that.

Oh and while you have had good luck with ECS, I have not.... they don't last too long in the Texas heat.

[Phenos]

Re: GA-7DPXDW-P Rev 1.0 fails to post

I have this board at my workplace running as an fileserver.
Replace the Choyo 330uF 25V Caps as well because those fail alot without visible signs of malfunction. (I used 470uF 16V)

TK is an japanese partner of Ost and is good quality. (Listed green on capacitorlab.com)

[370forlife]

Re: GA-7DPXDW-P Rev 1.0 fails to post

I have this board at my workplace running as an fileserver.
Replace the Choyo 330uF 25V Caps as well because those fail alot without visible signs of malfunction. (I used 470uF 16V)

TK is an japanese partner of Ost and is good quality. (Listed green on capacitorlab.com)

Uhhh, I think williwake must have been drunk when he did that then, because ost is known crap.

=whoops=

Ok, I checked it out on his site, it just said they are distributed by OST, not manufactured.

[PCTechNY23]

Re: GA-7DPXDW-P Rev 1.0 fails to post

Anyone know who runs capacitorlab.com? They list a couple shady places to buy caps like computekinc.us who sells beat old caps. I bought some from them and they looked like they had gone through the wash a couple times. Didn't even bother to install them, direct to trash. http://www.waifong.com.hk/ never has anything in stock and I wouldn't trust them to be real. Not many places to buy Rubycons in the US.