Surprised me too. I would not have believed it if I did not do it myself many times over.
Surprised me too. I would not have believed it if I did not do it myself many times over.
I might give it a try then ;) I don't care much about the SB as it's really cool with the HR-05, but the heatsink on the NB really consumes some space. If I can find a way to route that nicely... well, we'll see :D
PWMs might come in handy. For now, I used a pretty strong airflow through the case for testing purposes. Something I'd want to get rid of for normal use and who knows how warm they might get then...
If you decide to use the MCW30 chipset block on the northbridge, be sure and get the "raising" kit sine there is a capacitor that would get in the way of the wing mounting plate.
I use MIPS blocks and am very pleased with them. A german retailer sells them at http://www.mips-computer.de/
I have the set of four blocks, 2 Mosfets & NB & SB.
They are very compact and fit perfectly, even with 2 8800GTX's and other stuff
I'm currently using BIOS 0602 with Vcore on auto too and Everest (newest version) also reports 1.47V. Do you mean that I could set it to 1.325 so Everest would read that too? I'm running a slight overclock (from 266FSB to 333FSB), but I think I could do with some less volts if that's the case. It would probably decrease the temps too.
@Radical_53,
How big a temperature decrease did you find between using the stock heat-pipe system and HR-05's? I'm trying to figure out whether or not I'll need them for some good overclocking which will come eventually. Right now, my Everest reports 34C for the NB on cool days and up to 38 on hot days, but that's idle (is there any difference for NB/SB between idle and load?).
I never used the stock heatpipe to be honest. I tried that on a P5W and it was so bad that I never wanted to use one again ;)
I first put a Zalman NBF47 on the NB first, but it got way too warm even with a 80mm fan nearby (55-60°C measured on the backside of the board). Now with the HR-05, I get readings of 30-33° ;) Much better and absolutely "enough".
Everest doesn't read any temps related to NB, that sensor always shows 28°...
Not only recent versions. But the temperature shown doesn't have anything to do with the actual NB temperature ;) Not even close...
I don't know a single board where there's a real sensor that could measure the NB temp accurately.
I did a review on NB waterblocks once and used a sensor on the backside of the board. I tried to check the reported temperatures of the onboard sensor as a reference, but they didn't have anything to do with the real temperatures of the chip.
The temperature shown is just a small sensor somewhere around the board. Put a fan blowing on your board and the temperature drops... even if you use a water block ;)
If you want to know your mainboard temps in detail, put sensors to the points of interest. The board doesn't monitor it.
Whizzer,
Yes if you choose 1.35V Vcore in the BIOS, then Everest would show that value+/- a little bit. I cannot believe Auto overvolted the Vcore so much. However, I am running 501 that came with my other machine and a Xeon 3070 and Auto shows 1.33V (1.325V) which is correct. I think it is from 601 and later that the auto voltages get weird.
Safan80,
The "raising" kit is only like $2.00. While it may work like you say, this kit sure makes things easier.
Radical53,
Agreed. I run Everest (latest version for this board in particular) and the "mobo" temperature has nothing to do with the NB. I water cooled it and this value did nit change from before. It is somewhere on the board, but not on the NB. Even back in the 478 days, people always say that the mobo temp is the NB.
Yes, I know ;) Funny that this myth still stays alive... mobo temp has never been NB temp.
If manufacturers weren't so vague about this sort of thing then myths as you call them would never get started. How hard would it be for them to list the exact location of the temp sensors on the motherboard layout diagram in the manual?
Why would the temp in question fluctuate depending on voltage to the northbridge. Do you have any idea where the actual diode is located on this board?
Alright then. I think I will go with the HR-05 later on then (sometime in January when I have the money again, I'm saving up for a 9800GTX in November and a new 22" monitor, Samsung 226CW). Thanks a lot for explaining that. Is there any reputable brand with sensors or could I just get them from about anywhere? I've never actually seen them at stores or online so that got me wondering.
And I'll be sure to lower my CPU voltage then. I'm using 0602 so it should be overvolting as I type this. Anything to lower temperatures I guess.
New 0805 bios on the net, ftp and http.
@whizzer: Well, I think the Striker has the option to plug in sensors. You could put one on the backside of the NB and then you'd have an accurate reading (at least as accurate as it gets, a sensor close to the side of the NB, on the front, doesn't read very good values).
@mrbusa: I'd have to look for it. But it can be found for sure ;) Was easier in the older days when they were larger and often painted in bright blue :D
PS: Do you have a link to the bios? I can't find it anywhere.
Thanks alot! I'll give it a try ;) Hopefully that stupid Razer bug isn't in this release also.
Edit: Just flashed it an hour ago. The Razer bug is gone, bios seems to run nicely so far. I'll check if I can find any improvements for my machine...
Finally guys! I got my new RAM!!! 4 gigabytes of OCZ FlexXLC ddr2-1150MHz (OCZ2FX11502GK) ! I couldn't get them to boot @ 1T, no matter what voltage or latencies. Maybe I'll try swapping the sticks, or upgrading to the 0805 bios. Now I'm running them @ 575MHz 5-5-5-18 2T, 2.25v.
I want 4-4-4-4 1T so badly...
Don't think 1T is possible over 900-950Mhz mate, for 4-4-4-8 1T you will need to drop ram speed to around 800mhz, 4-4-4-12 2T 1000mhz is surely doable...
Think you set ya goals too high, my Gskills do 1000mhz 4-4-4-5 2t but not 1T...
@leeghoofd: I haven't run any specific stress tests so far. But the system booted up fine and any "game" I threw at it ran as it did before, without a flaw.
@jultsu: If you get the system to boot with 1T, try Prime Large to test for stability. It seemed to be the only test run that gave me problems when I was running too high RAM clocks on 1T.
Thanks for the new BIOS. I'll be sure to give it a try this week.
@Radical, that sounds great. Perhaps the P5N32-E SLI Plus allows for this too. I would have to remove the motherboard, correct? However I'd have to do it, where can I get good sensors? And how do I read the temperatures (Everest perhaps)? I'll do some proper research right now, I promise.
I'm not sure if the board does have the connectors soldered to it. If not, you'd need something to read those sensors. Lian Li does make very simple ones, nicer ones come from Aqua Computer (aquaero 4.0) or mcubed (T-Balancer). Most watercooling stores should have sensors like these.
Btw., I did some Prime runs. The new bios seems to limit my overclock a little. Still testing, but I got errors much earlier than usual. Let's see if I can find some old benchmark values. Maybe they improved performance somewhere, tightened some timings, otherwise I coulnd't explain how this would have happened.