ChillTec Thermo Electric CPU Cooler

Heatsink Test – Joe

SUMMARY: First look at a TEC heatpipe heatsink

Pic

The good guys at Ultra Products were nice enough to send a sample of their ChillTec Thermo Electric CPU Cooler to test out. A TEC is a solid-state heat transfer device which, in use, has one side that becomes very cool and one side that becomes quite hot. By dissipating the heat from the hot side, the cool side remains cool. A TEC when not powered is not effective to transfer heat.

Key Specifications

  • Thermoelectric Chip: 40 x 40 / 12V DC; TEC Temperature Response Time to Control Unit: < 1 Second
  • Total Power Dissipation: 0% Load = 5 W; 50% Load = 28 W; 100% Load = 50 W
  • Heatsink Material: 4 Heat Pipes, Aluminum Fins, Copper Base
  • 90mm fan specs: Low 2000 rpm / 20 dB(A); Medium 2400 rpm / 24 dB(A); High 2800 rpm / 28 dB(A)
  • Temperature Monitor Range: 0ºF~99ºC / 32ºF ~ 210.2ºF
  • Size: 128 x 104 x 147; Weight: 985g (With Fan)

One issue with a TEC is that in use, it’s either on or off – when on, the cool side (depending on heat load) can be under ambient temperature; this results in moisture condensing on the cold surface¹, an obvious problem in electronic cooling.

There are two approaches two controlling condensation:

  • Insulate all around the TEC: This requires an air-tight insulating blanket on any surface which is cooler than ambient air;

  • Temperature Controlled TEC: Use a temperature driven controller to ensure that surface temps surrounding the TEC do not go above ambient either by power cycling (on-off cycles @ 12 volts) or voltage control (0-12 volts).

The Ultra TEC Heatsink uses the latter method – a temperature controller:

Controller

As the back shows, there are ports for controlling the TEC and a power port.

Control Back

A closer look at the base clearly shows the TEC between two plates:

Pic

Very important – note that the four heatpipes are split, two dissipating heat with the TEC off, two kicking in when the TEC is powered. This means that when the TEC is off, only two heatpipes are dissipating heat – this may be enough under low power conditions; One question I have is are the two which dissipate the TECs heat when powered sufficient to deliver aggressive cooling performance.

The second question I have is fin density – The fins are relatively close together and long; pushing lots of air through this array when under load requires a high cfm fan to overcome system resistance.

Pic

Parts that ship with the unit for multiple CPU mounting schemes:

Parts

CONCLUSIONS

Only testing will tell how the questions I have will be answered – stay tuned!

Testing CONTINUED on page 2…

¹ Strictly speaking, moisture condenses when temps are below the dew point – however, during hot, humid weather, the dew point can be almost the same as ambient temps.

SUMMARY: Top-rank performer.

Use Test

The display on the controller unit shows the TEC temp:

Pic

The “CPU” refers to the image of the figure shovelling – it represents the CPU’s activity, although I did not see any real difference at idle or load. It also has a timer and total hours display.

The Test

The ChillTec was tested on an Asus P5WD2 motherboard P4 Motherboard Test Platform with a modified Pentium D 805 to read CPU case temps (both supplied by Directron). This setup is not inside a case.

Heatsink

Case Temp

Ambient Temp

C/W

On-Die Temp¹ / TEC Temp

ChillTec, 2445 rpm, 52 dBA²

34.7

20.5

0.15

44 / 27

¹MBM on-die temperatures.
²50 dBA measured 8″ from the fan intake corresponds to about 30 dBA measured 3 feet from the fan, a very quiet noise level.

Results place the ChillTec Thermo Electric CPU Cooler in the topmost rank of heatsinks tested to date (Heatsink Ranking) with the fan at its mid speed setting; there are no controls to allow users to vary fan speed.

At idle, at least for the setup I have, the fan continually cycled between its low and mid setting about every 10 seconds – could drive some people nuts. In all the testing I did, the fan never ran at its high speed setting. I also overclocked the CPU to 3.6 GHz (180 FSB) with the following results:

Heatsink

Case Temp

Ambient Temp

C/W

On-Die Temp¹ / TEC Temp

ChillTec, 2445 rpm, 52 dBA ²

35.6

20.9

0.11

45 / 27

¹MBM on-die temperatures.
²50 dBA measured 8″ from the fan intake corresponds to about 30 dBA measured 3 feet from the fan, a very quiet noise level.

It was not stable at 190 FSB and above, failing Prime 95 after five minutes with a hardware error; at speeds above 190 with the setup described above, I had not had overclocking success.

Conclusions

The ChillTec Thermo Electric CPU Cooler was a top performer, although it does place more of a power burden on systems compared to non-TEC heatsinks. Its performance is equal to the best aircooled heatsinks I have tested to date.

In addition, there is increased heat dissipation required due to the 50 watt TEC – case cooling may need a boost to compensate. MSRP $150.

Disclosure: Joe Citarella has a financial interest in a company developing thermosyphon products for electronic chip cooling.

Email Joe

Be the first to comment

Leave a Reply