I Water Cool My Rig Too

Cooling/Water Cooling by SidneyWong @ 2004-04-02

Not all Water Cooling systems are created equal, the same applies to air cooling as we all know. A year or two ago, the bragging right was how loud your delta fan could run relative to how low the CPU temp was. Today, you?d better have water cooling just to keep a spot at any forum. W/C is in; Delta fan is out, at least for now.

  • prev
  • Go to mainpage

Observations & Conclusion

Observation:

I spent a day to get this “thing” setup, since it is still March in Chicago and snow storm is not uncommon, go golfing is out of the question. Although the thermal result is not excellent, I now could proudly say that I have built a W/C system rather than buying a turnkey unit. I did not get killed in this battle, perhaps slightly wounded. With the knowledge and experience under my belt, I shall begin to speak up a little instead of sitting on the side-line in front of the screen says “Water Cooling Forum”. The lower noise level can only be appreciated when I turn on non W/C unit. The nice feature of this unit are; 1) I can take it anywhere with no worry of leak or spill because there is no reservoir and the pump is external, 2) no second box to haul around all built in and 3) it’s custom made or at least 50% custom made.

Update:

Two days later I received the missing temp probes from a customer service oriented dot.com place. The missing probes came from Super Flower SF-609 Fan Master that has had received good reviews from several review sites. Knowing the MSI BIOS reports kind of inaccurate temp (besides most P4 so called diode readings are manipulated by BIOS computations/compensations; if board makers use different sensor vendors updating BIOS will be like chasing your own shadow), I proceed to using the temp probes. (I cancel the idea of using the single temp probe from the X-Dreamer case).

Madshrimps (c)
These were the missing ingredients: Two flat thin probes and two thermistors (bulb) type


Madshrimps (c)
Reveals the TOPSPEED filler with SF-609 Installed



I will provide detail of the Super Flower SF-609 in my next review; something that was not mentioned in other reviews. I installed two probes on two different spots under the water block to see if the temps are identical and if not; the difference. Since most AMD Athlon motherboards are using probe / sensor under the processor, it is only fair to have the same type of measurement used on P4 knowing diode reading MAY be more accurate at times and with higher temperature reading than using temp probe. Another reason is the maximum case temp recommended by Intel could be difficult to “capture”.

Madshrimps (c)
28 degrees C idle with max temp set at 65C from probe 1


Madshrimps (c)
27.5 degrees C idle with max temp set at 55C from probe 2


Both are within 0.5C difference, I’d say acceptable. Interestingly, these idle temps are very close to my three P4 800 MHz systems using external probes.

Madshrimps (c)
Prime95 load temp from Probe 1 – 37.3 degrees C


Madshrimps (c)
Load temp from Probe 2 – 37.1 degrees C


Madshrimps (c)
Pump disconnected for 30 seconds – 48.7 C




I’m a cynical person; unplugged the pump for 30 seconds just to see what would happen. Well, it works; temp shot up fast from 28 C idle to 48 degrees. Another stupid check to make sure, I put one of the probes into my mouth and it reads 38C and 101F (please don’t do this, I may have a bit of fever.); I’d say it is accurate enough for something not meant for lab work.

These temps (28 C Idle and 38C load) are from the same set up with identical V-core and FSB; P4 2A clocked to 3.015 GHz at about 81 Watts reported by Sandra. The fan settings are: 92 mm radiator fan at 2100 RPM, case intake fan at 2,200, case exhaust at 2,250 and power supply at 2,000; resulting to 43 dBA sound level 18” away from front of the case. I live in a very quiet neighborhood with less than a dozen cars passing by on an hourly basis with ambient noise level of 35 dBA. Dialing the 92mm to higher and noisier speed does not lower the temperature and further. Then, if I just use the system for office apps or browsing the net I could turn it down to minimum fan speed of 1,800 RPM at about 41 dBA and still within the acceptable temp range.

My final thought:

TOPSPEED may not have a product that suits the overclocking community due to not so clever design of the “housing” and “mounting position”. However, the radiator and water block are pretty decent, I only wish the 230 L/H pump could be a bit more powerful; doubling it would be nice. This simple project is not about “a” product, rather, the idea of W/C and the fun in modding. I’d say it turns out pretty good. In fact, this TOPSPEED has found a permanent home. Now, if TOPSPEED would just sell the pump/radiator and water blocks at the right price and change the pump to 12-volt, I would not mind paying for it. I found another good spot in mounting the unit. Until next time, take care!


Questions/Comments: Forum thread
  • prev
  • Go to mainpage