![]() | ||
|
|
General Liquid/Water Cooling Discussion For discussion about Full Cooling System kits, or general cooling topics. Keep specific cooling items like pumps, radiators, etc... in their specific forums. |
![]() |
Thread Tools |
![]() |
#1 |
Cooling Savant
Join Date: Feb 2003
Location: USA - Boston area
Posts: 798
|
![]()
I'm looking for a reasonably inexpensive way to figure my temps. About all I can say about my CPU diode temps is that my idle/load temps go up and down in lockstep with the thermometer in my wall thermostat (and they go up/down between idle/load
![]() I suppose I should just "not care" as I'm getting the overclock I want and running stably - but the most recent "post your temps" post over on oc forums has just pushed me over the edge... Thanks! Bob |
![]() |
![]() |
![]() |
#2 |
Big PlayerMaking Big Money
Join Date: Aug 2001
Location: irc.lostgeek.com #procooling.com
Posts: 4,782
|
![]()
If your interest is only for "show and tell" then I'd just go with what you have now and qualify the temperatures reported by including motherboard and bios revision.
If you want to integrate temperature monitoring for water (best spot is right before wb inlet), air (maybe watch the intake air and exhaust air on radiator) and CPU, then I think personally that the Dallas One Wire probes are the way to go for the average person. They are transistor package digital thermometers that are accurate to 0.5C out of the box and wire length/solder points does not affect them. You can either build a reader or use the Matrix Orbital MX2/4 or Crystalfontz 633 to read them. The nice thing about these readers from MO and Crystalfontz is that you can link fan speeds and even PC power to the temperature probes. This lets you adjust fan power between idle and load to lower noise and gives you a hardware-based way to protect the system in event of a problem. Oops I should add that epoxying the temperature probe underneath an AMD CPU is only going to be close to the CPU die temperature if you are running normal water cooling or air. If you chill the CPU substantially below zero then the probe on the backside will be off by a good bit
__________________
Getting paid like a biker with the best crank... -MF DOOM |
![]() |
![]() |
![]() |
#3 |
Cooling Savant
Join Date: Feb 2003
Location: USA - Boston area
Posts: 798
|
![]()
Thanks for the info!
Hmmm... will have to think about whether I want to build a reader or buy a bay device. The Panaflo L1Aa that I use are noisier when undervolted with PCM than with a simple resistor/rheostat (just have 'em on the 5V bus right now). Looks like that project would extend to soldering on a SMD, which my soldering skills, equipment and even my freak'in eyesight (don't get old if you can avoid it) aren't up to. Bob Edit: Will those Dallas one wire probes fit under the processor (AMD XP-mobile)? What if I use "frag tape" to attach it (will probably want to move the sensor to another processor... eventually.) Last edited by bobkoure; 05-12-2004 at 12:14 PM. Reason: add question |
![]() |
![]() |
![]() |
#4 |
Big PlayerMaking Big Money
Join Date: Aug 2001
Location: irc.lostgeek.com #procooling.com
Posts: 4,782
|
![]()
They fit fine:
![]() Use 30 gauge (wrapping wire at Radio Shack) wire and I suspect you could get it between socket pieces. I cut a slot in my socket for clearance of much larger wires for diode reader. Best bet is to use thermal epoxy and just wire up a different CPU sensor when needed.
__________________
Getting paid like a biker with the best crank... -MF DOOM |
![]() |
![]() |
![]() |
#5 |
Big PlayerMaking Big Money
Join Date: Aug 2001
Location: irc.lostgeek.com #procooling.com
Posts: 4,782
|
![]()
Oh and I gave you this method because I feel like it's the simplest way to get decent accuracy. You can go with thermistors or thermocouples or diode readers, but all of them have to be calibrated (and esp the diode readers are a hassle to PROPERLY use and shield).
__________________
Getting paid like a biker with the best crank... -MF DOOM |
![]() |
![]() |
![]() |
#6 | |
Cooling Savant
Join Date: Mar 2004
Location: Lawrenceville, NJ
Posts: 254
|
![]() Quote:
__________________
Ghetto riggin'! |
|
![]() |
![]() |
![]() |
#7 |
Cooling Savant
Join Date: Feb 2003
Location: USA - Boston area
Posts: 798
|
![]()
At low voltage, particularly when the low voltage is supplied via PWM, they make a "rumbling" noise. Doesn't matter if they're hard-bolted to a high-mass plate, soft mounted with low-hysteresis material (rubber) or high-hysteresis (50 durometer silicone), held in my hand or sitting on my desk. It's not the same noise at all as that produced by an out of balance fan (which I regard as broken, BTW).
If I had an out-of-balance fan, and if, for the sake of argument, I wasn't going to just replace it with an unbroken one, I'd just balance the thing. If you've got a "dial back to zero" automotive strobe light you can use that to figure out where the heavy point is - or you can just look through another fan that you've got a speed controller on (harder, but do-able). If you've got a CRT, try looking at that with a fan to see what I'm talking about. I used those "inventory sticker dots" as balance weights by putting them off-center on the hub. It would probably be OK to put them on the underside of the fan blades (top surface does the lifting - coanda effect) but I didn't try this. Bob |
![]() |
![]() |
![]() |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
|
|