- Oct 20, 1999
- 7,132
- 0
- 0
Hey Everybody(this time, including Thermaltake PR):
I just got some test results comparing some heatsinks via internal diode/external thermistor(directly behind cpu core). All clips were untouched(so the clip pressure is higher than necessary for intel chips). All tests done with Arctic silver, and done 3 times.
Tested on A C2@ 927/1.9V(approximately 34.5W)
Ambient Case temp of 25C
The orbs(both of them) exhibit the same type of behavior I have suspected. Something with either airflow or design causes thermistor readings to be drastically lower in comparison with internal core temps(diode readings) than do other heatsinks. For example, the FOP32-1 has a difference of 4.5C going from internal diode to thermistor, whereas the Super-orb has a difference of 9C.
From more testing, it seems as though the thermaltake ratings on thermaltake's website are not too far off actual readings. Transplanting these readings into a t-bird 1.1ghz/1.85V(~65W), you would get core temps of
with a 30C ambient case temp)
Fop: 57.3C
Cek: 63.15C
S-orb: 66.4C
C-orb: 75.27C
The S-orb would appear to be a suitable, but not great heatsink. The C-orb has its uses, but below 900mhz. past that and you're running into cpu core temp "red-zone" with AMD chips.
The Importance of this test, is, to again show why socket-thermistor tests are inaccurate, compressed, and un-relative.
Charts and Such of These Tests
Mike
I just got some test results comparing some heatsinks via internal diode/external thermistor(directly behind cpu core). All clips were untouched(so the clip pressure is higher than necessary for intel chips). All tests done with Arctic silver, and done 3 times.
Tested on A C2@ 927/1.9V(approximately 34.5W)
Ambient Case temp of 25C
- Globalwin FOP- Internal Diode:39.5C; Thermistor:35C
- Taisol CEK733092- Internal Diode:42.5; Thermistor:37
- Thermaltake Super-Orb- Internal Diode:44; Thermistor:35C
- Thermaltake Chrome-Orb- Internal Diode:48.5; Thermistor: 39C
The orbs(both of them) exhibit the same type of behavior I have suspected. Something with either airflow or design causes thermistor readings to be drastically lower in comparison with internal core temps(diode readings) than do other heatsinks. For example, the FOP32-1 has a difference of 4.5C going from internal diode to thermistor, whereas the Super-orb has a difference of 9C.
From more testing, it seems as though the thermaltake ratings on thermaltake's website are not too far off actual readings. Transplanting these readings into a t-bird 1.1ghz/1.85V(~65W), you would get core temps of
Fop: 57.3C
Cek: 63.15C
S-orb: 66.4C
C-orb: 75.27C
The S-orb would appear to be a suitable, but not great heatsink. The C-orb has its uses, but below 900mhz. past that and you're running into cpu core temp "red-zone" with AMD chips.
The Importance of this test, is, to again show why socket-thermistor tests are inaccurate, compressed, and un-relative.
Charts and Such of These Tests
Mike