Quote from: Trillium on July 13, 2013, 04:52:42 AM

Quote from: ssateneth on July 12, 2013, 05:41:40 AM

Quote from: Trillium on July 11, 2013, 10:39:48 AM

Nothing rare, but my two little USB erupters arrived today. Set them up on a shelf above my desk with a D-Link DUB-H7 7 port hub I got for $23. I quickly realised that they were reaching 70+ deg C temps when I was setting them up and the 2x 15 watt server fans really help (but they are running at 3.3 volts so they are essentially silent). Miners now get better temps of about 45 deg C (25 deg C ambient) after running for a while. Getting <1% hardware errors in bfgminer I might have to look at that later.

Is there any components on the other side of the PCB? Or is the aluminium heatsink removing heat through the PCB itself? Some micro-sized heatsinks might be in order for the ASIC chip on the top side in any case...

You realize your fans are easily using 6 times (or more) as much power as the USB miners themselves? Those fans are also noisy as FUCK. You can be fine with just a gentle breeze. The USB miners actually don't need any cooling at all. The manufacturer has tested them with zero airflow and even removed the heatsinks (the chunk of metal with the BTC symbol on it) and they still ran just fine with no increase in HW errors.

If you really feel you need to run fans, get something slow and quiet. Something that uses very low power. A 240mm fan that only draws 0.8 amps @ 12v will cool 50 usb miners without a problem.

As I said they are 12 volt fans running at 3.3 volts. At 3.3 volts they use 0.145 amps each (measured with DMM) = 0.48 watts each. So actually the two miners use about 5x more power than both fans. And as I said they are silent at 3.3 volts.

Also the average longevity of the ASIC at ~45 deg C is going to be much better than some at 70+ deg C. Sure the manufacturer tested them but they didn't test them for months continuously nor would they have conducted a long term study of their failure rates vs temp. Given that these miners will probably never pay for themselves (even if they last for years) it makes sense to operate them at lower temps to try and maximize their working life. Also if the ASIC and other components are heating up the entire circuit board (especially with the assistance of the heat spreading 'heatsink' on the bottom of the PCB) while there is no airflow then the other components will also reach unfavorable temps...

http://i.imgur.com/QnaIvC7.jpg

AOZ1020 - Maximum ambient temperature 85 deg C
CP2102 Maximum operating temperature 85 deg C
ATtiny2313 - Absolute maximum operating temperature 125 deg C
74HC574 - Maximum operating temperature 125 deg C

It's not surprising that some of those components could reach the temps above if they are sitting on a PCB thats over 70 deg C, especially the voltage reg which would have to dissipate quite a bit of heat. Also as the spec sheets make clear, many of those devices become unstable at higher temps, often well before their rated limits.

Totally right.