PCPower carefully chooses, and quality tests every component they place in their chassis, and only handicraftsman can adequately accomplish this goal. In so far as the PCB, PCPower has them custom made, as seen below;
One of the simplest ways to determine a mediocre PSU, from a quality PSU is weight. Pure copper wiring, large transformers, capacitors, and heatsinks, equate to weight. Just as the best heatsinks are 600g copper monsters, so should your PSU possess mass. There is a direct correlation between mass and heat. More mass distributes more heat. PCPower&Cooling occupies every square millimeter of their steel chassis with highest quality components, and fan placement is right where it should be. Earlier I'd spoke of a "purest" or "minimalist" design. While other's sell units based on a large number of fans, chassis material, windows, and LED fans, PCPower eschews such ergonomics investing every penny into components. They "minimalize" so that manufacturing costs go towards a better capacitor, or inductor.
This is also what is meant by a "purest" approach, employing tried and true electrical theory and circuitry, featuring ripple and noise suppression, and extensive line conditioning. They back it all up with a 5-year warranty, and that truly says it all. There are many whom might think paying over a $100 for a single fan unit, is crazy. What is in fact crazy is the effectiveness of marketing which has us focusing on prima facie accoutrements rather then basic electronics. When the designer spec's the unit with electronics exceeding the requirements of their intended function, they're not running so hot as to require localized cooling. This is where PCPower&Cooling sets the bar. They use only one fan, because that is all that is required, and place it where it's needed most as seen below;
The photo above indicates four adjustable pots, and the yellow arrows point out the 3.3V, and 12V rails which are accessible with the chassis in place. The 5V rail is also accessible with the chassis in place, located on the side of the unit. The photo below is a close-up of the 3.3v, and 12V pots with chassis in place;
The cabling supplied with the TurboCool 510 Deluxe is the finest I've seen with any PSU. Each cable is individually braided, and just the right lengths are used. Often PSU makers in their attempt to please everyone will supply overly lengthy cables.
While this is helpful for those of us with very large towers, most ATX towers regardless of size have the motherboard and PSU within close proximity. PCPower has made the primary ATX, and Intel connectors optimal length, and others long enough for the tallest of cases. None-the-less they're a company renowned for custom orders, and adding a few cm's to specific cables can be resolved with a phone call;
The true test of any PSU is its ability to hold the rails, and hold them as things heat up. For the TurboCool 510 there may be a challenge for which it was born. I'm speaking of the infamous, literal gateway to a silicon Black Hole, otherwise known as the Prescott pipeline. The TurboCool 510 shall be our guiding light as we enter a vacuous realm from which few motherboard power circuits have egressed unscathed. A point of singularity even Intel hasn't been able to fully explore. It has been whispered, when Prescott reaches speeds of 5GHz, (error free of course) its bandwidth shall cast a shadow over the entire Central Valley of Southern California. Will Prescott, leave the TurboCool 510 a quivering mass, of smouldering steel, ball of fused plastic and glass inside? Or will TurboCool 510 prevail, leading us back into the light?
Our test system will feature the Abit AI7 (Springdale chipset) featuring the µGuru
utility. µGuru derives its capabilities from a custom made Winbond microchip. I searched Winbond's site and emailed both Winbond and Abit concerning this chip, but haven't received a response yet. The chip, (Winbond W83L950D) is located in close proximity to the PLL crystal clock-generator, seen below;
The chip is basically a W83L785R(pdf)
on steroids, monitoring 8-voltages, 3 fan RPM's and 3-temps, one being the Intel on-die thermal diode. In addition, and its claim to fame is the ability to adjust the FSB frequency, AGP-voltage, DDR-voltage and Vcore. As seen below;
The utility does in fact work, and for the most part changes are recognized, however; certain programs are not able to read these changes. WCPUID and Sisoft Sandra benchmark are not able to recognize these changes. While this may not be a problem for most, as an Enthusiast or Reviewer, measuring performance after making such changes is essential. While the anomaly may be attributable to the hardware, the fact CPU-Z does acknowledge the changes seem to refute this theory.
ClockGen suffers from the same problem
in that WCPUID, nor Sandra’s memory benchmark cannot recognize its changes unless the system is re-booted. One can see the dilemma. If you have to re-boot for all changes to take place, you could just have easily raised the FSB, and any voltages in the BIOS as well. In fact entering the BIOS is the preferred method. It's just not the simplest method. In some respects this kicks Guru back to the Fuzzy Logic, or Easy Tune-III/IV from Asus and Gigabyte respectively.
On to the testing ->