TEC/Peltier effect CPU coolers - and why they're not very useful.
I put this link in another old thread, but thought I would add it here for (hopefully) easier finding via search in future. We've had a few threads I can recall where people ask about peltier effect coolers aka. Thermal Effect Coolers (TEC). This review is a useful summary of the pitfalls, and shows the poor performance that is achieved: http://www.anandtech.com/show/10695/...-cooler-review
Quote:
Originally Posted by http://www.anandtech.com/show/10695/the-phononic-hex-2-0-tec-cpu-cooler-review
There are however a few problems when working with TECs:
1. Condensation. A typical TEC can produce a temperature difference of up to 70 °C between its cold and hot side. Assuming that a heatsink is mounted to the hot side and that it is capable to maintain a near-room temperature, the cold side of an uncontrolled TEC can be significantly colder than its ambient surroundings. That will cause condensation, which will be disastrous inside a PC.
2. Efficiency. TECs are generally inefficient, with an efficiency usually lower than 15%, which means that they consume disproportionally high amounts of electrical energy for the work they actually offer.
3. The electrical energy losses that the TEC inserts are converted directly to thermal energy and transferred to its hot side. Therefore, the heatsink has to deal with the thermal load of the system plus the energy losses of the TEC, increasing the size and performance requirements.
Quote:
Originally Posted by http://www.anandtech.com/show/10695/the-phononic-hex-2-0-tec-cpu-cooler-review/6
The thermal performance of the HEX 2.0 is a complicated matter. The presence of the TEC allows the HEX 2.0 to provide exceptional performance when having to deal with low thermal loads, but the performance degrades as the thermal load increases. For very high thermal loads, the HEX 2.0 is equivalent or worse than an advanced air cooler, hardly justifying its price and energy consumption. That being said, while the HEX 2.0 is good for low temperatures in low power environments, it does add another 20W to the power consumption, perhaps negating the point of a low power system. It should work wonders if installed on a modern energy-efficient CPU, even with some moderate overclocking, but it cannot compare against top-tier air coolers and liquid coolers for advanced overclocking. Bluntly put, if thermal performance is your primary concern and space is not an issue, this is not a product for you.
In summary, the Phononic HEX 2.0 is a very interesting product but it is clearly aimed at a very specific segment of the market, which is users that have serious space limitations and or cannot use liquid cooling solutions for any given reason. If large air coolers or liquid coolers can fit, it is highly unlikely that the HEX 2.0 will have any distinct advantage, let alone justify its $150 retail price.
Peterb also dug out this technical paper which summarises some of the efficiency issues faced by Peltier coolers: http://www.asee.org/public/conferenc.../3976/download For the range of tests undertaken peak efficiency was only 2.2% for a deltaT of 68 degrees. That is about the limit of CPU temp above ambient - and ideally you'd want it lower than that.
To run a TEC in a PC you use more electricity than you need to, have to deal with condensation around the CPU and mobo electrics (!), still have to cool the hot side of the peltier - which is even hotter than the CPU would have been itself, generally don't get the magic solution you might hope you would at first.
the only interesting case I could see for it would be for low-power CPUs with ULV in laptops where potentially the hot-side could be made the entire base of the chassis. But the power hit probably still renders this impractical as it would dint battery life.
Re: TEC/Peltier effect CPU coolers - and why they're not very useful.
Oh, yes, I rember that discussion! Fundamentally the problem still remains on how to remove the heat from the hot side of the peltier junction and take it to the outside of the case.
As you say, realistically a laptop I'd probably the most practical application, but as you say, the power hit may not be worth it.
Perhaps an indication of the practicality might be gauged by looking at the number of commercial designs that use peltier cooling in laptops or desktops. :)
Re: TEC/Peltier effect CPU coolers - and why they're not very useful.
I had to check the review to see what on earth you were talking about. Peltiers have long been used in extreme overclocking circles because it's absolute temperature that matters there, not efficiency.
That all in one unit is an interesting piece of technology but, like closed loop coolers, seems to be principally aimed at those who choose based on the coolness-factor rather than anything practical.
Re: TEC/Peltier effect CPU coolers - and why they're not very useful.
Yikes! Are people still using TEC's??? I use to use them back in the day when I was overclocking to the max and managed to kill the odd motherboard & gpu due to condensation & voltage overload. :) Assumed, like me, these had gone the way of all flesh and that cooling tech & hardware had moved on.
Re: TEC/Peltier effect CPU coolers - and why they're not very useful.
Before I left to go back to university earlier this year, I was working on a cloud chamber cooled by a pair of TECs and a Fractal CPU watercooler I won in a Hexus competition.
I can't remember the figures but the pelts were pulling nearly 180W, the radiator was toasty. This was as far as I got;
http://img.photobucket.com/albums/v6...8.jpg~original
http://img.photobucket.com/albums/v6...5.jpg~original
http://img.photobucket.com/albums/v6...4.jpg~original