Test Platform And Overview
Our (Temporary) Test System
Since one of the test’s objectives is the assessment of Thermalright's new base plate, expected to improve cooling performance on Haswell-E-class processors, we're using an X99- and LGA 2011-v3-based system.
Our subject is a six-core Intel Core i7-5820K equipped with Hyper-Threading technology. Selling for around $390, it constitutes an affordable entry point into the company's highest-end platform. We expect many readers to choose this processor for price and performance reasons.
Sporting a base clock of 3.3GHz and a maximum Turbo Boost frequency of 3.6GHz, this CPU is actually slightly slower than Intel’s four-core LGA 1150 flagship, the Core i7-4790K. But for many applications outside gaming, the two extra cores improve performance markedly.
We picked Gigabyte's X99-Gaming 5 for our supporting board; it's able to handle most of the overclocking scenarios we'll throw at the processor, and it offers one distinct advantage over larger motherboards: the cooler’s heat pipe doesn’t extend over the rear RAM slot. That's a good thing, since a large heat sink's pipes dissipating thermal energy over your memory can be counterproductive.
The operating system and the test programs reside on a fast, inexpensive Transcend SSD370. Its 128GB capacity is almost overkill for this test.
Because this machine is partly the author’s personal PC, picking RAM brand was easy. We went with 32GB of Crucial DDR4-2133. While these four 8GB DIMMs are budget-priced, they're a solid choice since they can typically be overclocked to DDR4-2400, even at 1.2V. At $100 per DIMM, the Crucial RAM is an excellent choice from a price/performance perspective.
The real surprise may be our PSU choice. We went with a 400W be quiet! Straight Power E8. Since this is not a graphics-heavy story, a 400W PSU is more than sufficient, and this specific model's fan is nice and quiet.
The graphics card we used, an XFX Radeon HD 5450, won’t inspire envy, either. But at least its passive cooler is completely silent, allowing us to more accurately judge the CPU fan.
We conducted all measurements with the system mounted on a Cooler Master Test Bench.
Last but not least, we used an Aqua Computer Aquaero 6 controller to set and measure the fan speed. While most motherboards offer some kind of fan control built-in, they're not as fine-grained as the Aquaero 6. Nor are their read-outs as precise.
Overview of the Test Candidates
In order to make this Macho family comparison more exciting, we invited lots of derivatives to the table, in addition to Thermalright's brand new HR-02 Macho Rev. B. One of them is the HR-02 Macho Rev. A (BW). It offers less eye candy, a different fan and an older version of the base plate.
We could have omitted the Macho Zero, since it's basically the same as the HR-02 Macho Rev. B with black nickel plating and without the latter’s fan. However, we're including it for the sake of completeness.
The biggest and baddest participant is the huge HR-22. A few days before we started testing, we received a TY-121 fan, which now ships with the small Macho 120 Rev. A. Hence, we put the latest fan on our older Macho 120 and included it in our test.
Like the Macho Zero, the HR-22 doesn’t ship with a fan. It's intended for (semi-)passive operation, after all. But in order to level the playing field and because fan mounting components are included with all of these coolers, we put a TY-147A fan on both of them.
As mentioned, we also included the Macho 120 (our Mini Macho). We had an older version in our lab and upgraded it with the fan of the current Macho 120, for fairness' sake.
Thermalright Macho 90
It takes 70 minutes to measure the CPU temperature. After a soak time of 60 minutes, the remaining 10 minutes are logged. We plot the average of the six core temperatures. Note that we report the difference to the ambient temperature, subtracting the room temperature from that average.
If you want to know what CPU temperature to expect in summer conditions, add your room temperature to our reported temperature value. Of course, this is only an approximation, since cooling performance degrades as ambient temperature increases.
We put the ambient temperature sensor 8“ in front of the CPU fan’s intake, within the top third of the fan. We took care to position the PSU in such a way that its exhaust air is not fed back into the processor fan. In fact, we reversed the PSU and put it at the bottom of the test bench. Thus, the ambient temperature sensor measures actual room temperature.