Stress Test: Power Supplies Under Full Load

PC Power & Cooling Turbo-Cool 850 SSI: Excellent Price And Performance

Top performance at a great price: PC Power & Cooling's Turbo Cool 850 SSI delivers 850 watts for $500.

The graphic above shows the power supply connections and their length in centimeters. A cornucopia: like the second power supply from PC Power & Cooling, the Turbo-Cool offers 6 SATA connections. Drawback: only one floppy connector - but it does support dual core.
Feature And Test Results

The black monster of this comparison demonstrates it once again: Enough is never enough, at least not when it has to do with sheer output in supplying a powerful system. What may be sufficient for a generously equipped PC is by no means optimum for a top-flight maxed-out mega-PC.

Its temperature, measured at 142.7°F (65.1°C), demonstrates that you can easily burn your finger on a lead to a power supply unit. The measurements of the individual voltages are not as convincing as those of the Epower 470 watt power supply unit, nonetheless they do lie within the confines of the specifications. In any case, its extremely high efficiency (80.4 percent) is of seemingly record proportions.

Input: 1058 watts (213 volts, 5.02 ampere, cosphi 0.98)

Output: 850 watts (21.2 ampere/3.3 volts, 26.0 ampere/5 volts and twice 26.3 ampere on 12 Volt, 20 watts on -5/-12 volts, 5 volts standby)

Degree of efficiency at maximum load: 80.4 percent (208 watts dissipation increases the air temperature by just under 21.6°F (12°C)

Noise reading at maximum load: 68.3 dB(A) in 4.7" (12 cm) gaps

Storage oscillo-scope: Excellent values are demonstrated here, the 12 volt CPU fluctuates by 25 mV - it's 40 mV for the 3.3 volt cable.

Fantastic values: Oscillogram of the 850 watt monster from PC Power & Cooling.

The power supply unit is examined after the successful test run.

Upscale technology: The four voltage paths of the individual cables can be clearly seen.

Gets extremely loud despite high-grade ventilators: The fan impeller has only three blades.

Standard at the$500 level: A network filter removes the voltage spikes from the supply network.
PC Power & Cooling Turbo-Cool 850 SSI (850 watts)
Test Phase 1
Test Description Lower Limit Upper Limit Result
Increase load from 20% to 100% of nominal load limit 170 watts 850 watts Passed
Test Phase 2
Test Description Load Result
Steady load at 100% nominal load 850 watts Passed
Test Phase 3
Voltage Stability
Power Rail Measurement Min/Max according to ATX Spec. Result
+ 3.3 V + 3.32 V +/- 5% (+3.14 to +3.46 V) Passed
+ 5 V + 4.96 V +/- 5% (+4.76 to +5.25 V) Passed
+ 12 V + 12.04 V +/- 5% (+11.4 to +12.6 V) Passed
+ 12 V CPU + 12.18 V +/- 5% (+11.4 to +12.6 V) Passed
+ 5 V Standby + 4.96 V +/- 5% (+4.76 to +5.25 V) Passed
- 12 V - 13.1 V +/- 10% (-10.8 to -13.2 V) Passed
Power Rail Measurement Min/Max according to ATX Spec. Result
+3.3 V 40 mV 50 mV (Peak-Peak) Passed
+12 V CPU 25 mV 120 mV (Peak-Peak) Passed
Further Data
Value Measurement
Noise Level (max.) 68.3 dB(A)
Voltage 213 V
Current 5.02 A
Cos Phi (Distorsion of Phase) 0.98
Efficiency 80.4%
Statement From PC Power & Cooling

We would like to take the opportunity of including a statement from CEO of PC Power & Cooling that we just received.

An honest power supply rating is for CONTINUOUS operation at FULL LOAD. 24 hours is a very short test. There should be no complaints! A PSU should actually run at full load for weeks and months with no problems.

The voltages and output ripple should be within ATX specs at full load. The tests are NOT too tough. In fact, they're easy, because they are run at nominal input line voltage and room temperature. I suggest that tests be conducted at low line voltage and high line voltage using a variac. These tests would demonstrate the PSU's input operating range and whether or not the PSU could withstand the sags and surges in today's world at full load.

Another legitimate test condition is to run the PSUs at full load with the ambient temperature elevated about 15C to represent the real-life condition in the computer, rather than the condition on the test bench (many power supplies have exaggerated wattage claims based on low ambient temperature assumptions). One more important test would be a dynamic load test (instantaneous pulse of the +12V output). This would demonstrate a PSU's ability to handle multiple hard drive access (servers) or quick changes in CPU and video power loads (gaming PCs, especially SLI).

In conclusion, the stress tests being conducted by Tom's Hardware are more than fair. Products that do not pass should be redesigned or de-rated to realistic power levels.

Doug Dodson, CEO, PC Power & Cooling