OzSnoal

Hey Roger Sounds like the problem I had with my OCZ VX ram back when they first came out. I found that if I allowed both sticks to cool down to room temperature, I could boot with both in. It was also only one of them that would 'die'. Eventually the 'dieing' stick stopped working completely. I recommend you get your RAM replaced with normal voltage RAM, not this 3V+ stuff. Chris

Follow the link in my signature. Edit here it is, should I change my sig and the link disappears. http://www.dfi-street.com/forum/showpost.p...480&postcount=3

I recently fixed a persons machine that was having similar behaviour. PC was gine until you started using it. It turned out to be the keyboard, a wireless Logitech something or other. Somehow the sleep/shutdown button was being activated.

Thanks Jon. Your fix worked for me. My NVidia LAN also stopped working ages ago, thanks for sharing.

I have seen where a PC would not boot because of dust build up on the memory stick closest to the CPU heatsink. I mean there was so much dust on it, I could not see the PCB anymore.

You do the PSU volt mod to reduce the power dissapation (read heat) of the regulator. If you want to work out the approximate power dissapation of a linear regulator you can use the formula: (P)ower = (V)oltage * (I)Current Example 1 The voltage is taken as the Input Voltage - Output Voltage Input voltage is either 3.3V or 5V, depending on your jumper setting on the motherboard. Output voltage as that you set in the BIOS. Let's take Vout = 2.9V Choosing an arbitary value of 4A for the current: Power dissapation of the regulator at Vin = 3.3V would be P = (Vin - Vout) * (I)Current = (3.3 - 2.9) * 4 = 1.6 Watts (Coolest) Power dissapation of the regulator at Vin = 5V would be P = (Vin - Vout) * (I)Current = (5 - 2.9) * 4 = 11.6 Watts (Hottest) If you do the volt mod to your PSU, say turning up your 3.3V rail to 3.5V. Vin = 3.5V P = (Vin - Vout) * (I)Current = (3.5 - 2.9) * 4 = 2.4 Watts (Middle) You always want your regulator (or any part for that matter) to run cool. The best way to do this is to keep the (Vin - Vout) to a minimum, this is where the PSU volt mod came in. So, in the above example, the RAM was being run at 2.9V, there is no reason to do a volt mod or run on the 5V jumper. Now I am repeating the excercise with 3.1V ram or as I call it Vout. In this case you can no longer use the 3.3V rail, now you either need to do the volt mod or run on the 5V rail. Increasing the Vout to 3.1V should also increase the current to around 4.28A. Power dissapation of the regulator at Vin = 5V would be P = (Vin - Vout) * (I)Current = (5 - 3.1) * 4.28 = 8.13 Watts If you do the volt mod to your PSU, say turning up your 3.3V rail to 3.5V Vin = 3.5V P = (Vin - Vout) * (I)Current = (3.5 - 3.1) * 4.28 = 1.71 Watts EDIT: Aah, read Angry's post too

I can only speak for the VX, and I recommend not to get it. You may be lucky, you may not. I was not, even with the second set i received and eventually settled for the ram in my signature. I don't trust Anandtech's reviews anymore, not because they don't know what they are doing, but more because they get samples to review. Search on cold boot... maybe one of the OCZ officials would care to comment.

No

Well, if that is what he wants, I suggest he gets rid of the VX and replace it with some lower/normal voltage ram. As for the answer to the fix, there is no definitive answer. I have not seen one from DFI, nor from OCZ, or for that matter other memory manufacturers that have also had cold boot problems. If he does not want to fiddle, get rid of the ram or the board. I agree it is frustrating, there is no easy answer... if there was he wouldn't be asking! I know from experience, I fiddled for ages trying to get my VX to cold boot, and eventually found some of the causes of which a bios update helped for some. For me, gradually the cold booting got worse and worse, eventually my VX would not boot at all (one of the sticks anyway) since then I replaced by OCZ with the ram in my sig. I have not had a cold boot since, and this has been many months. He must try what is in the link I gave above, otherwise, contact OCZ directly. End of story.

http://www.bleedinedge.com/forum/showthread.php?t=11998 Their you can try everything that OCZ says to try to fix your cold boot. Good Luck. P.S. jeffbar1, unless you elaborate on what you have tried, you cannot expect people to 'know' what you want.

6964

Well, bridge and 2nd Leadtek card arrived today, and SLI is up and running on my modded nf4 ultra: 3DMarks2003: 15361 (DNA 81.85 balanced driver) I have the extreme versions which come at default 550/1120 GPU1 @50C GPU2 @39C I'm happy. I suppose the wizer choice would have been to sell my one 6600GT and get a 7800 GT instead. But then where's the hardware fiddling and mod fiddles then? Chris

I don't believe you save things to the BIOS chip (except of course a new BIOS), that would be done on a chip seperate from the BIOS. Also, the battery normally only handles the clock and possibly the data you save to CMOS (or saved to a EEPROM/Flash chip seperate from the BIOS), so being able to boot without a battery does not really tell you anything. http://en.wikipedia.org/wiki/Nonvolatile_BIOS_memory

Resolved. I _think_ it may have been the bios setting that appears in single VGA mode to support dual 6600GT. I set it to Dual support, moved the jumpers to SLI mode and now a single card works in either slot with the jumper set to SLI mode. Thanks for all the feedback guys. I'm now thinking of getting a second 6600GT + bridge. According to the new nvidia drivers the clocks speeds no longer need to be the same on the cards.

Soldering iron. Can anyone confirm what wangerin says?