If there’s one thing that attracts me to the Alpha line of Heatsinks, it is their uncompromising performance and standards of engineering. Back about 3 months ago I was unimpressed with what passed for socket 7 coolers: Under-performance seemed to be the rule.
Along came the Alpha – I was so impressed with the Alpha PFH 6035’s performance as a socket 5/7/370 cooler that, on impulse, I bought a bunch and started to sell them to fellow overclockers. My objective was to get some high-performance cooling gear into the market at reasonable prices (driven partly by my reaction to one company selling a $3 80 mm fan for $15 and mis-representing its performance to boot!). We sold a lot of Alphas for very little margin and I’m pleased that Alphas are now widely available at reasonable prices.
No doubt that Alpha raised the performance bar for other heatsink vendors. Global Win has stepped up to the plate with the Global Win FDP 32, a socket 7 monster. It is large, measuring 2 7/8 x 2 ½ x 2 ¾ inches high, compared to 2 ¼ square x 2 5/8 high for the Alpha. This prevents the Global Win from fitting into certain motherboards such as the ABIT BP6 (a smaller version, the FEP 32, remedies this but most likely at the cost of performance). Some basics on heatsink design tells us why the Global is so massive:
There are 3 basic parameters that determine a heatsink’s performance: mass, fin area and air flow. Designers can play these factors off each other and as such each design represent a compromise, with cost playing a large role. The cheapest heatsink to produce is an extrusion – basically forcing aluminum stock at high pressure through a die. What comes out is one very long heatsink that is then cut to size and machined. This is how the Global is made.
The Alpha is more costly to make – they are made by a forging press.
The heat sink base and fins are one piece of aluminum, formed by a large press. I imagine a block of aluminum goes in and the heatsink pops out after a die squeezes into shape using very high pressure. Pin fin heatsinks are inherently more efficient, with the benefit of reduced size at high performance levels. In addition, some air flow work convinced Alpha that a hood covering half the heatsink’s pins was more efficient in directing air to where it’s needed the most. More efficient pin fins, compact and highly directed air flow – a tough design to beat.
So Global’s challenge was to meet or exceed Alpha’s performance within the limits of the technology it uses – extrusions. Not an easy task. The heatsinks weigh about the same, so I believe Global increased its physical size to increase fin area to the point where it could compete with the Alpha. The constraint, however, is to avoid making the sink so large that in meeting performance it would be too large to use – and in fact Global exceeded the size constraint making the smaller FEP 32 necessary.
Let’s take it to the test pad and see if the Global Win is an “Alpha Killer.”
Now I test heatsinks in a very simple but fiendishly clever way: I use a 40 watt peltier to simulate a CPU which generates 40 watts of heat. I have a little test rig which places a thermal probe on the peltier’s cold side, giving me its temperature. I measure the ambient temperature, then subtract the cold side temperature from the ambient to give me a reading of how much heat is being dissipated by the heatsink. Notice: No system variables, no CPU variables, no case variables – just the heatsink and heat. I use 2 power supplies – one just for the peltier and one for the fans.
I took my time on testing one against the other – these test results represent multiple tests over a 3 day period. The fan on each is the YS Tech 26 cfm units. Results:
Alpha PFH6035: 67.4 F
Global FDP 32: 64.4 F
Three degrees difference – that’s pretty damn close to matching the Alpha, but not an “Alpha Killer.” Now I’m intrigued – what would it take to make the Global meet Alpha’s performance? There’s really only one thing I can vary – air flow. Break out the fans and let’s see what we can do.
The first thing I try is to mount an 80 mm fan on the Global. Result: Performance degradation – why? If you look at the size of the fan’s hub, it directs less flow to the center of the sink that with the 60 mm fan. OK – how about 2 fans? I place 2 60 mm YS Techs on the top of the Global and see a performance increase:
Global Win FDP 32 – 2 fans on top of heatsink: 65.2 F
Getting close, but problem is the fans overlap the sink and a lot of air is wasted. How to close the gap? How about 2 fans on the sides – one blowing into the sink, one sucking air out of the sink. But part of the fan does not butt up against the heatsink – the fan is taller than the heatsink’s fins. OK – let’s put a shroud over the top of the heatsink so the fan’s airflow is contained. A little tape, some index cards and I have 2 fans on each end of the Global with a hood on top.
Global Win FDP 32 – fans on side of heatsink: 66.6 F
Close but no kewpie doll. I try a number of variations – 3 fans (one on top, 2 on sides). Worse – probably too much air flow interference among the fans. Anything I tried was not better than the blow in/suck out side fan idea (this arrangement is similar to an Alpha variant of the P125C). Now whether this is practical is another matter – I’m sure in some instances this arrangement could work where the one fan Global does not.
The Global Win FDP 32 is the first serious challenger to the Alpha PFH 6035. It is ugly, large and inelegant compared to the Alpha, but it is a very close-performing second. With the potential to increase performance with some tinkering, I would rate the Global an interesting alternative to the Alpha.
P.S. I’m pleased to see one “main line” heatsink manufacturer step up to providing overclockers with a better product, and expect to see more Alpha-challenging designs on the market from others. Ain’t competition grand?