The GAW Miners Fury is a low end, inexpensive, scrypt coin miner. As such, it lends itself well to modifications. Much like the power supply, the cooling system can be improved without much effort or technical ability. Does modding the cooling system do anything for the Fury's performance? Let's take a look.
The Fury's ASIC chips are designed to transfer heat through the bottom. Proper application of quality thermal paste can improve heat transfer into the heat sink. I removed the main PCB and applied Phobya HeGrease Thermal Compound.
I'm a little skeptical about adding heat sinks to the top of the ASICs, but I'm almost positive that they're not going to hurt. It seems that pulling a little of the heat out of the top can only help. I used Cosmos Mini Aluminum Heatsinks.
|An attempt at measuring chip temp with a thermocouple and Kapton tape.|
I measured chip temperatures with and without heat sinks and the body of the Fury itself in an attempt to see if the thermal paste was transferring heat more efficiently. Almost all of my temperature data is flawed. I put a bunch of time and effort into collecting it, but if it's crap... it's crap and it can't be used. My instruments and subsequent methodology we not up to the task.
Though I don't have any reliable data on how the cooling mods effected temperatures, we can look at the overall performance of the cooled Fury in the form of hash rates, rejects, and hardware errors.
|Stock Fury Cooling|
Using an ATX power supply to power the Fury, the stock cooling averaged right around 6% hardware errors across the tested clock rates. That's representative of the Furys I've seen. They all seem to hover around 5-7%.
|Fury w/heat sinks and new thermal paste|
After upgrading the cooling, the hash rates, rejects, and hardware errors show no significant change. I was able to try a slightly higher clock rate of 354 and found error rates were up 3-4%. My assumption is that the ASICs are starting to run out of voltage past a clock rate of around 350.
The test results indicate that the stock Fury cooling is adequate at clock rates under 350. If your Furys live in an environment under around 75-80°F, there doesn't seem to be anything gained by improved cooling. Harsher environments might be able to get away with additional fans. Heat sinks and better thermal paste are probably not required at less than extreme ambient temperatures or extreme clock rates with Furys that have been modded to provide higher voltage to the ASICs.
Next up: Volt mod!
See also: Fury Power Supply Mod
|Fury Test Bench|
Criteria: Watts, amps, volts, hash rates, rejects, and hardware errors are an average of several samples. Data was recorded after running for several hours to allow for the numbers to stabilize. Measurements were made with the following tools: Kill A Watt Electricity Usage Monitor, Fluke 62 MAX Plus Infrared Thermometer, Fluke 87-V Digital Multimeter, Extech EX330 Autoranging Multimeter.