PDA

View Full Version : Coolermaster HHC-001 w/ 92mm fan update


Breadfan
02-04-02, 10:19 AM
Well, I did my cooling upgrade yesterday...got a bunch of pics too, but I'm at work so I don't have them here. I'm gonna write up a bit of what I did, etc etc.

Anyway, thought I'd atleast let you know how it went and what I did since I was talkign about it last week. I was curious to see if I could get a 92mm fan on my HHC-001 without losing much cooling capacity and also withough buying some cheesy $15 fandapter.

I ended up making my own fandapter out of scrap metal, which went very well. I used a thermaltake 92mm fan rated at 56cfm. Infact, the dead spot in the middle is only about 1mm more in diameter than the 60mm 36cfm fan on the unit. (Which looks likes a delta clone).

Anyway, because of how things progressed I didn't do a direct comparison of the coolermaster with the 60mm vs the 92mm. I still can though, and might in the near future.

Heres the end result though. I upgraded from a 900mhz tbird to a 1200 AXIA tbird I acquired a few months ago from someone for a really sweet price. I decided to hold onto it until I upgraded the cooling or had more parts for a second system, etc etc...

Anyway. I have a multimeter with a termistor wire on it. It works pretty well, but the thermistor is too big to put next to the die with the heatsink on...its taller than the die. I tried to crush it (it was a metal ball) and it seemed to shrink ok. However, when I was moving the wires around trying to mount it on the cpu, I ended up breaking the end off from bending it too much. The wires are really brittle.

Anyway, I was able to just twist the ends together and still get a temp reading, but now I don't even know if this is accurate or not. it seems like the metal ball that was on the end was just that...a metal ball. So, in theory, the metal wires connected should work the same....infact, it seeemed to reach its end temp alot quicker with just the wires twisted. Does this make much sense? I'm just curiosu if its still reliable, and also, how is the best way to get it attached to the cpu next to the die?

Anyway, becuase of this, I just used the thermistor temps, which while innacurate, will atleast show the changes. Neat thing is that I was using a 92mm antec fan on my WBK, so in a sense the air movement around the socket should be somewhat similar.

My old setup:
tbird 900@1000 w/ WBK = 45C

The new setup with the cooler master (same board...etc)
tbird 1200@1365 w/ hhc-001 = 41C

All these are at 1.88v and the same load (folding@home and sandra burnin)

I guess I'll have to get used to copper...for a while the coolermaster was at 37-38c but finally leveled off at 41c....took longer to level off than the aluminum wbk.

Anyway, thats pretty respectable, and its very quiet, not whining, but some wind noise obviously. I lost my side blowhole when I installed my homemade window, but I'm gonna put it back, with a 92mm sunon rather than 80mm as before. This will blow oustide air directly over the cpu hsf, which I think will get me below 40c....thermistorly speaking :)

Oh btw, according to my older version of Radiate, my axia tbird at 1365 should generate 81 watts and the 900@1000 should be generating 59.4. Thats a pretty big difference, and I still have cooler temps!

Anyway, I'll get more written up on it, and have some pics later. My fandapter is pretty awesome, and wasn't that hard to make if you have a dremel...

Mike
PS: I had run my old chip at 1066mhz for a while, but it was causing some instability so I went back down to 1000 and got rock stable. So far I'm rock stable with the AXIA at 1365, and I've maxed out my A7v's multiplier at 12.5x and 105fsb...oh interesting to note, I set the multiplier to waht should be 12.5x and the fsb was at 100mhz and it gave me 1300mhz when I booted and the asus utility shows that speed to. Whats up with that? Last time I checked 12.5x and 100fsb should be 1250mhz...

Monaco
02-04-02, 12:22 PM
that's pretty cool man! I'm impressed, that is a big-a55 fan on there you have. Some really good temps too.

The thermal probe will work fine even if you mess with it's wires- the metal ball on the end is a resistor. A resistor's value changes in linear fashion according to temperature changes, your PC interprets this to be a temperature. Screwing with the wires may slightly change the overall impedance of the probe, but not by much. And it will be a constant too.

Breadfan
02-04-02, 01:59 PM
Awesome, I'll probably give that shot soon. I'm curious what my actual temps are, and I'd like to find out if my ASUS is reporting hi or low...I've heard alot of people say their A7V's are actually reporting high which would rock, but with 81 watts being generated, I'm skeptical my thermistor temps are being reported too high...probably a few C low...but I wont know for sure unless I check.

With the impedence changed will the thermometer be innacurate now?

Oh, and one thing I forgot to mention which was cool, when I clocked up to 1365mhz I got 7168 in 3dmark2000. I know its 2000, but keep in mind I'm still using a GF2MX. Considering at 1ghz my best score was 5100. So I gained 2000 extra points in 3dmark2000 even with a gf2mx. Not too bad...I thought my video card was gonna hold me back from much more, but apparently not. I'm just waiting too see what GF4 prices are and then I'll decide on a GF4 or GF3ti200...

Mike

Monaco
02-04-02, 03:19 PM
wow-7168 with an MX!!

My GTS and Celeron@995 only gets 6500, for cryin out loud! And I thought I was cool.....:D

I use 3dMark2000 a lot too, 2001 isn't so hot unless you have GF3 level equipment IMHO. Plus I hate seeing myself score only 3200 in 2K1, so depressing.

The thermistor may be off a bit now due to the wire chopping, as the wire has it's own impedance, but it will be constant with itself- it will always be off by the same amount. So it is still perfectly usable as a temperature trend probe, just not 100% matched to what the real-world temps are. Those things are never too accurate anyway, they usually are within +/- 10% or so.

Since it is an in-socket thermistor the reading will not be at all accurate anyway. Have you tried to calibrate your diode? It's a pain but totally worth it-
check it out here! (http://www.arcticsilver.com/diode_calibration.htm)