View Full Version : My motherboard hates me and my brand new OCZ PC4000

10-26-04, 06:50 PM
Damn it. I got my OCZ PC4000 Gold Rev. 2 in today to discover two things. One - that the PC4000 Gold Rev.2 is the exact same RAM as the PC3700 Gold Rev.3, which I just owned and RMA'd. Two - It still doesn't work.

Basically, I'm trying to achieve an overclock of 2.5 GHz. To do this I was going for a 250x10 HTT/Multiplier and 1:1 DRAM ratio. With the OCZ PC3700, this was simply not possible, and the RAM wasn't even stable at the tiny overclock of 240x10 1:1. This pissed me off a good deal, of course, considering every review I've read of this ram ranted about how amazing an overclocker it was, posting stable overclocks of 270+FSB at 1:1 ratios. But mine can barely POST before crashing or corrupting my BIOS at slightly higher than rated speeds.

So, I RMA'd the PC3700 thinking I just got a bad pair of sticks and opted for the PC4000. I got it today, put it in the system, and set the timings to 2.5-3-3-8 (which is what it is rated for), the FSB at 250, and the multiplier at 9, just to be safe. DRAM voltage at 2.7 (which OCZ recommends for this RAM), 2T command rate, and I figured I was ready to go. So I get past POST and the computer restarts when Windows begins to load. I decided to back off the clock speed a bit and went to 240, and booted again. My much maligned "Bad BIOS Checksum" appeared again and I had to put in the motherboard CD to let it re-flash the BIOS. So, I updated my BIOS to the most recent and tried again. This time I did 3-3-3-8 timings, and booted to Windows. Oh boy! I decided to run Prime95 to test for stability. Much to my dismay, Prime95 couldn't run for even a minute without failing. Good times.

I called OCZ to have their tech support guy inform me that the RAM I had was not only the exact same RAM I just RMA'd, but that the chips on the RAM weren't very good at handling the clock speeds I was running. Which was kinda funny, considering there are stickers on the headspreaders that VERY CLEARLY SAY DDR500 PC4000 2.5-3-3-8.

Anyway, I guess I'll have to RMA this stuff too. OCZ will apparently pre-test a replacement module with my motherboard (they said they had one there) and give me sticks with better chips in them (Samsung something-or-others) if I send it to them. I guess I will, but first I wanted to tap your guys' limitless experience and see if maybe you knew any way to get this stuff to work. Googles have turned up squat.

10-26-04, 08:01 PM
Just checking but is your HTT divider set to 4x?

Also, you might give 2.8v a whirl just to see what happens.

BTW dude, you might consider getting yourself some PC3700EL instead if it isn't to late to exchange. That stuff is rated @ 2-3-3-8 233mhz. Fastest DDR433 I know of and is pretty flexible memory. This PC3200EL I'm using is stable @ 2.5-3-3-8 250MHz 1:1.

10-26-04, 08:27 PM
Ninjaman, I'm wondering if you don't have something else going on.....My PC3700 Gold rev2 runs perfectly on the Asus A8V-2.0 @ 2.5 3-3-7 T1 2.8v
1:1 250(fsb) HTT x4

10-26-04, 08:29 PM
Are you sure your mobo is an A8V rev 2.0?

10-26-04, 08:42 PM
Actually, I believe my HTT is set to 5x, now that you mention it. I'll try 4x and see what happens. BTW, what the heck is the HTT divider anyway?

I'm fairly certain it is rev 2.0, but I'll double check. Off to try your suggestions, hopefully I'll be back shortly...we'll see.

Oh, and yes, I tried 2.8v as well. The OCZ tech support guy actually told me not to use it, and that the chips actually functioned better at 2.7v. He assured me that was enough power to get them to DDR500, at least.

10-26-04, 08:52 PM
He's probably right since my 3200EL runs fine @ 250 w/2.7v. Still, it's just an experiment and probably won't hurt to try out.

Here's a discussion regarding what the HTT is:

10-26-04, 09:08 PM
OK, well I'm back after yet another BIOS corruption. My Corsair is back in there, reliable as ever. Anyway, I have confirmed that I have the Asus A8V Revision 2, it says so on the board. Under HTT settings in the BIOS I had the following options:
200 MHz
400 MHz
600 MHz
800 MHz
1000 MHz

I assumed by 4x you all meant 800 MHz so I had that selected. I ran the RAM at 250 fsb with both 2.7v and 2.8v at 2.5-3-3-8 2T command rate timings. The 2.7v almost made it past the welcome screen before crashing, and the 2.8v corrupted my BIOS. So here I am again. I'm wondering if I just have really bad luck because this is two different pairs of the same RAM doing the same thing, but it seems that other people with a very similar setup as me can run what I'm trying to run with no trouble. I'm more confused now :P

10-26-04, 09:50 PM
Well effin 'a that sounds like a real fun experience. :rolleyes:
You have my sympathy. :)

And yes, 800MHz = x4 as you said which is OK for 250. Honestly if the RAM is rated for that I have no idea what the problem is. You're only trying to run it at the rated speed.

I'd almost have to wonder if maybe your particular board doesn't like running @ 250.

10-26-04, 10:16 PM
I'm beginning to arrive at that conclusion myself. I am currently running at 250x10, 1:1 4x HTT with the RAM at 2.7v and 3-4-4-8-2T timings (which suck). The weirdest thing is that all my errors are memory related and I have been able to run at 260 FSB before with this board, only on a lower ratio. About to run memtest86 AGAIN (and probably fail). If this doesn't work I'm just gonna RMA the RAM - the OCZ guy told me they'd hand-pick new sticks for me, so maybe that's the thing to do. It's just that this has happened twice with the same RAM - maybe I just can't run at 250 FSB after all :P

10-26-04, 10:18 PM
OCZ gold chips still use laser modified samsung tcb3 chips, no?

10-26-04, 10:25 PM
Try a 3XHTT(600) just to see if it will help.

10-26-04, 10:39 PM
Not sure what's on the gold. I know platinum 3200/3700 use TCCD.

10-26-04, 10:42 PM
Actually ragejg, apparently these sticks do not. One of the things the OCZ tech guy mentioned was that they would send me sticks with samsung tcb3, and that my sticks don't have it.

Ahead of you, Waste. It didn't seem to have any effect. Actually, changing it from 5x to 4x didn't appear to have any effect either.

Anyway, the setup in my previous post appears to be somewhat stable, but I'm not sure how long it would last. I put the Corsair sticks back in and I think they'll be there for a while. Back to stock speeds and better temps for now, I guess. I appreciate your help, everyone.

10-26-04, 10:47 PM
OCZ and Mushkin both take real good care of their customers from my experience. :)

10-26-04, 11:43 PM
OCZ and Mushkin both take real good care of their customers from my experience. :)

Don't know about OCZ, but Mushkin has been nothing but trouble for me. I've given up on them after waiting more than 6 months for them to replace a stick of PC3200 the tech support guy said they would after I'd sent it back to them. I've already replaced the memory so I don't care any more, but I'm never going to buy that brand of memory again.

10-27-04, 07:37 AM
Wow that just blows me away. They've never been anything but very accomodating every time I've had to get memory replaced with them and it's been twice now. They even let me return my 3200 dual pack when they dropped prices two weeks later and exchange for 3500.

10-27-04, 08:33 AM
I'm going to see if OCZ will let me trade for some 3200EL Platinum. Gotta wait till they open to find out, heh. I did manage to get a stable system with 3-4-4-8-2T clocks, but frankly, the performance increase (or lack thereof, rather) wasn't anything to get all giddy about. Besides, I didn't pay for 3-4-4-8 @ 250Mhz - I paid for 2.5-3-3-7 @ 250MHz. I think I'm just gonna give up on overclocking my CPU/memory, it's been a very expensive, very unsuccessful venture, and the performance gains absolutely do not justify the costs. By now I could have bought an Athlon FX-55 for the amount of money I've spent on O/Cing stuff, and be seeing much better performance. Blah, oh well, lesson learned.

10-27-04, 08:36 AM
Well if they let you exchange for that 3200EL give it another shot. I know mine and also circuitbreaker's runs OK @ 250x10 2.5-3-3.

BTW, running 2T is killing your bandwidth if you didn't already know. :)

10-27-04, 09:04 AM
I know, my Corsair has no trouble running 1T. Weirdly enough, my entire system feels "smoother" with the Corsair in it. I guess A64s really do love better timings.

10-27-04, 09:18 AM
Mainly just a command rate of 1T, everything else can be countered with higher clock speeds. Or so I have found with my brief testing and from feedback on forums.

2T is evil, but on Athlon 64 systems, it is the devil himself.

10-27-04, 03:35 PM
I just got this, Buffalo PC3700 and it has Micron -5b C chips on them, the same mem chips at the OCZ 3700EB. IT only cost $87 here.
I running my 1x512 dimm with 3.3V at a speed of 245HTT 2.5,2,2,5,T1.
Not bad for $87US :clap:

10-28-04, 10:24 AM
I am getting my computer warmed up for 275+Mhz FSB my memory is Geil PC4000 and it loves juice and i would say all memory does after 400MHz, 2.85v is pretty much minium for 500+Mhz memory mine is garanteed upto 3.1v I am hoping at 3.1v i can get it to do 550MHz or maybe higher :cool:

10-28-04, 10:31 AM
That is typically true but there is some newer stuff that just came out, "GSkill" memory, that seems to want no more than 2.8v even at high clock speeds. Or so I've heard...

10-28-04, 01:23 PM
That is typically true but there is some newer stuff that just came out, "GSkill" memory, that seems to want no more than 2.8v even at high clock speeds. Or so I've heard...
Its using Samsung TCCD chips.