G.Skill 6400 HZ

Rroff · 29 Oct 2006 at 01:11

Got a 2x1gig set of this stuff... overclocking it...

first had it stable at 1000MHz 4-4-4-12 @ 2.3v and seemed fine, dropped timings down to 4-4-4-8 but ran into some issues, however I just did a super long stress test on the ram and had one error, so now I've dropped it back to 4-4-4-15 @ 2.35v still 1000MHz and its gone twice as long as the first intensive stress test without any problems showing... my query is even with active cooling the heat spreaders are getting very very warm, can touch them but they feel hot... should I be concerned? I have an 80mm fan on them has anyone had problems with them running 24x7 at 2.3-2.4v?

easyrider · 29 Oct 2006 at 01:24

Rroff said:
Got a 2x1gig set of this stuff... overclocking it...

first had it stable at 1000MHz 4-4-4-12 @ 2.3v and seemed fine, dropped timings down to 4-4-4-8 but ran into some issues, however I just did a super long stress test on the ram and had one error, so now I've dropped it back to 4-4-4-15 @ 2.35v still 1000MHz and its gone twice as long as the first intensive stress test without any problems showing... my query is even with active cooling the heat spreaders are getting very very warm, can touch them but they feel hot... should I be concerned? I have an 80mm fan on them has anyone had problems with them running 24x7 at 2.3-2.4v?

That ram does not need that voltage at that speed.

Rroff · 29 Oct 2006 at 02:55

Right now thats a moot point

After changing the voltage and rebooting the computer locked hard just after showing the second BIOS screen (PCI bus scan, etc.) and after reseting I got nothing, fans spin up, hdds spin up but no POST beep, no video display. Did a CMOS reset and still same problem, left CMOS battery out, tried changing the RAM, video card, monitor, everything still same problem... at which point whilest moving my shiny new 7950GX2 around I managed to bang it on the PC case and knock off one of the tiny components, I think a resistor, gulp, really really not my night... fortunatly the component still seems to be in one piece *finger crossed* and I have access to equipment for working on micro-electronics so gonna have a go at soldering it back on, if not guess I'll have to hope it is a resistor and I can pencil it enough to work :s not to mention it looks like I need a new motherboard - no idea what on earth I've managed to do - but most of the parts work seperatly that I've taken out.

Last time I spend upwards of 2 grand on a PC this one has been nothing but trouble so far.

EDIT: it is a resistor - hopefully I can get its value with the equipment I have.

dark4orz · 29 Oct 2006 at 05:44

You've been Frying Chips and now you have Burn't Chips

No. seriously check out DFI Street

ted34 · 29 Oct 2006 at 11:08

Id put a 80mm silent fan blowing over the spreaders at those volts. Will help keep them a little cooler which will help there stability and maybe get you stable on the single error on long stress run.

As for 7950 just rma it - doubt they will notice a resistor missing but if you soldert it back on and then it doesnt work its a bit obvious

Rroff · 29 Oct 2006 at 13:51

I had an 80mm fan on the memory...

I've taken everything out and tested all but the motherboard and CPU seperatly - unfortunatly I can't test either of these 2 in another rig... so I'm not sure which one I have to replace :s

Left the motherboard overnight with the battery and power disconnected and still nothing (using an old PCI gpu to test)

My bro is building a similiar rig next week but different motherboard so hopefully I can test the CPU.

The problem with RMAing the GPU is that judging by the component that has been damaged, the card will still work under some conditions, but fail under specific conditions, i.e. when/if they tested it it would probably display the OS, etc. find but randomly lockup/crash whilest gaming.

Rroff · 29 Oct 2006 at 14:17

Forget it gonna scrap the lot

Justintime · 29 Oct 2006 at 17:42

My HX 2GB Kit is problematic on one of my DS3s too, fails memtest86+ after 2 or 3 passes with the case cover on, Cpu is a 6300@3200MHz, with case cover off it seems to be running fine so the fact that it makes it past a few passes makes me believe the memory is fine its just after it really warms up it starts to flake out, this at 2v too, nothing really high, gonna try a fan on it with the case closed to see for sure.

Rroff · 31 Oct 2006 at 01:25

looks like its the CPU thats dead - can't tell 100% til thursday but motherboard is passing all diagnostics, measuring voltages, test cycles, etc.

a tad alarming that it died with less than 2 weeks running mildly overclocked with hardly any vcore increase +0.05 and plenty of cooling.

OC_A64 · 31 Oct 2006 at 09:38

2.3v is nothing for D9 memory. It's about the max safe to run 24/7 with moderate additional cooling.
Don't be afraid to run more through them either, run ~ 2.5 through mine on a regular basis when pushing the dimm's.
2.3 maybe a little on the high side for 400 FSB, but every kit is different - wouldnt worry about running 2.3 through them as long as you can keep the dimm's cool.

Rroff · 1 Nov 2006 at 02:17

was running them on a 500FSB (1000MHz) but even with a couple of 80mm fans on them they were getting quite warm...

trojan698 · 1 Nov 2006 at 02:36

I'd send easyrider the bill tbh, sounds like his fault

Seriously though unlucky on the rig, managed to get the board, cpu and ram back up and running? seconding the RMA suggestion, it could have fallen off whilst in the box

Rroff · 1 Nov 2006 at 20:00

loool

nah its my fault for overclocking, I'm well aware of the dangers, but first time in almost 10 years that I've actually had to do more than clear the CMOS when its failed...

Gonna get the motherboard and cpu sorted this weekend - the RAM seems to be fine.

As for the GPU, gainward don't wanna know and by the time I'd RMAd it to overclockers, they got it back to gainward and replacement eventually got back to me... well I'm too impatient for that and would have bought the g80 in the mean time

which is prolly what I'm gonna end up doing anyway.

Rroff · 2 Nov 2006 at 21:08

well just tried the CPU in my brothers ASUS board... looks like the CPU is dead... while the board POSTs with it in (which mine doesn't) it won't go into the BIOS (ignores all key presses) and then gives a media error - no CPU error but doesn't have the CPU listed in the startup stuff either.

Rroff · 4 Nov 2006 at 22:51

ack

CPU AND motherboard seem to be dead and I've got a bad feeling I've killed another E6600 establishing that...

While I've done the best job I can of repairing the GPU I'm not confident its going to work... this looks like being expensive... gonna need a new motherboard and new GPU as a minimum and quite likely another E6600.

Devious · 6 Nov 2006 at 19:35

wow, that sucks bad...

just so i dont do it, in your second post you said you changed the voltage, what voltage was that and was it up or down ?

trojan698 · 7 Nov 2006 at 13:10

RMA it all, or sell the dead stuff to me!? I'll give you £20!

Rroff · 7 Nov 2006 at 16:32

I don't believe in RMAing stuff when I killed it myself :s

Well what I thought I was doing was backing the RAM voltage down to 2.1v - however I was very tired at the time and the Northbridge voltage box is right by the RAM box and unfortunatly allows voltages above 2v (default is 1.6v) judging from the damage done I've got the bad feeling I accidentally left the RAM at 2.35 and set the NB to 2.1...

Rroff · 8 Nov 2006 at 00:41

hmm new board does not have a 2.1 setting in the Northbridge - tho its running an older BIOS atm...

Rroff · 8 Nov 2006 at 19:43

wooohooo!

new motherboard and new CPU - everything is working - ran the GPU through 3D Marks 06 without any problems and about to try some other tests...

I'm absolutely stunned that the soldering job seems to have worked (with replacement parts scrounged off an old FX5200 - see they are good for something)

Fingers crossed it all works happily ever after without repeat of the first problem...