Anyone using the T-Balancer bigNG?

[tensim]

I was mostly joking about the superglue idea Figured it'd be a permanent mount and not the best for thermal conductivity!

Like the idea of a dab of thermal paste and a teeny bit of glue, might be a good solution.

Another quick question; the unit comes with 2 digital sensors (in addition to 4 analogue ones), I'm assuming these willl be the most accurate and would be best mounted near the CPU and GPU?

I guess the best spots will be directly on the heatsink base for each? Or just on the fins as close as possible to the heatpipes and base? Or somwhere else?

cheers

[peterb]

I'm not sure what they mean by a 'digital sensor'. The actual temperature sensing element will be a thermistor, thermocouple or a special diode that changes resistance/generates a voltage/changes its characterisrics in a predicatble way in response to temperature change - and is inherently analogue. It might be that it is bonded to an A to D converter so the output is digital - so teh calibration (if any) would be unaffected by teh cable length, but it while it may be more precise, it may not be any more accurate (apart from the calibration issue - and in this application, accuracy to a degree would be good enough - and about as accurate as I would expect - even though the precision may be higher.

As for the siting - it is part of a feedback loop. The CPU is generating heat - which is flowing through the heatsink (which has themal resistance) and is removed by the fan. The faster the fan, the greater the thermal difference, and the greater the heat flow. So if you put the sensor near the fan, it will measure the temperature there, and the CPU will be a few degrees higher, and so the sensor will be a little slower responding. Place it nearer the source and the sensor will respond quicker, but the fan may hunt more (speed up/slow down, which might be annoyinmg - it depends on the damping factor of the whole system.

In practice it probably doesn't make much difference - anywhere convenient on a flat surface of the block - the important thing is to get good thermal contact. If you want to experiment, you could tape (use sticky tape or something) to hold the sensor in different places and see what difference it makes - personally I would n't be bothered!