Antec logo  

Power Supply Specifications, Testing, and Antec TruePower.
~ What Those Numbers On The Back Of The Box Mean, And Why You Really Do Care ~

By Han Liu
Edited by David Taue
and David Forster


If you stop and think about it, which most people don’t, the power supply is actually a pretty critical component of your computer. It’s connected to everything. If you overload it or it misbehaves, your computer has a reduced chance of acting properly, and an increased chance of never acting again at all. And in high-performance computers with less tolerance for error, this is even more true.

So now that you have yet another thing to worry about, what can you do about it?

You can start by reading this article, for one thing. And you can keep an eye out for planned future articles.

One very important feature of any power supply is that it provide steady power. Other features are important too, but we will begin with this one.

DC Voltage Regulation

An ATX power supply should supply steady output (within the voltage regulation limits) under any load condition within its power range. When Antec tests a unit, this is the first and main thing it looks for.

An ATX form factor power supply consists of 6 outputs: +5VDC, +3.3VDC, +12VDC, -5VDC, -12VDC and +5V Standby (SB). Intel has defined the voltage regulation limits in the ATX12V Power Supply Design Guide. They are:

Output
Range
Min.
Max.
+5VDC
±5%
+4.75V
+5.25V
+12VDC
±5%
+11.40V
+12.60V
  -5VDC
±10%
-4.5V
-5.5V
-12VDC
±10%
-10.8V
-13.2V
+3.3VDC
±5%
+3.14V
+3.47V
+5V SB
±5%
+4.75V
+5.25V

By comparison, Antec TruePower supplies are designed and built for workstations or servers that require more stringent voltage regulation. They usually require allow only ±3% variation vs. Intel’s allowed ±5% limits for +5V, +3.3V, +12V and +5VSB output; ± 5% vs. Intel’s ±10% for 5V and 12V output.

Output Voltage Regulation Limits - TruePower vs. Intel’s Design Guide

DC
OUTPUT

RANGE

MINIMUM

MAXIMUM

TruePower

Intel

TruePower

Intel

TruePower

Intel

+5V

±3%

±5%

+4.85V

+4.75V

+5.15V

+5.25V

+12V

±3%

±5%

+11.64V

+11.40V

+12.36V

+12.60V

+3.3V

±3%

±5%

+3.20V

+3.14V

+3.40V

+3.47V

-5V

±5%

±10%

-4.75V

-4.50V

-5.25V

-5.50V

-12V

±5%

±10%

-11.40V

-10.80V

-12.60V

-13.20V

+5V SB

±3%

±5%

+4.85V

+4.75V

+5.15V

+5.25V


Okay, great. Now we know what the expectation is. We still have to test to see if an ATX power supply meets the spec.

The Electronic Load Test

An electronic load test machine usually consists of several modules. Each module is capable of drawing a programmed constant current simultaneously from each output of the power supply to check the stability of each output voltages.

 

Example of an electronic load test machine.

We will use one of the Antec’s TruePower supply supplies as an example to show you how we test it.

Model: True430 430W TruePower Supply

A/C Input: 115V/230V
Maximum Output: +5V/36A, +12V/20A, -5V/0.5A, -12V/1A, +3.3V/28A, +5V SB/2A
Maximum Power: 430 watt
+5V, +3.3V and +12V maximum output: 410W

One of the characteristics of a typical ATX power is the so-called " +5V and +3.3V combined maximum output." In this traditional design the +5V and +3.3V outputs share the same voltage output circuitry, and increasing the load on one decreases the remaining capacity available for the other. TruePower has dedicated output circuitry at +5V, +12V and +3.3V output, subject only to the maximum of the total power supply. This enables the power supply to output at its full rated maximum output at +5V, +12V and +3.3V on each of these outputs, in this case up to the designed total of 410W. (The total maximum output of True430 is 430 watts, in which about 20 W is generated on the 5V, -12V and +5VSB lines.)

Preparation

Before conducting the test, we need to program the load machine according to the specification of the power supply. We will start the test from 10% load and gradually increase the load to 100%. After that we will increase the load to 110% and beyond to see how far the power supply can handle the load. We have filled in the rated maximum output current according to the spec of the power supply, then calculated the test value of each cell and filled in that number. Results appear below.

 

+5V

+12V

-5V

-12V

+3.3V

+5V SB

Total
Theoretical
Output (W)

LOAD

A

V

A

V

A

V

A

V

A

V

A

V

Rated

36.0

20.0

0.5

1.0

28.0

2.0

10%

3.5

2.0

0.5

1.0

3.0

2.0

73.65

25%

9.0

5.0

0.5

 

1.0

7.0

2.0

150.35

50%

18.0

10.0

0.5

1.0

14.0

2.0

278.45

75%

27.0

15.0

0.5

1.0

21.0

2.0

406.55

100%

36.0

11.5

0.5

1.0

28.0

2.0

432.65

100%

15.5

20.0

0.5

1.0

28.0

2.0

432.15

100%

30.0

15.0

0.5

1.0

24.0

2.0

431.45

110%

17.0

22.0

0.5

1.0

31.0

2.0

473.55

%


Note:

  • Since the maximum output current of the 5V, -12V and +5V SB are relatively small (0.5A, 1A and 2A respectively), we will test these outputs at maximum current throughout the whole test.

  • We have calculated three combinations of load that add up to 100% of the rating for the power supply. This is because we want to test the power supply under various load conditions.

Testing
During each load test, we’ll let the power supply run for at least 5 minutes before recording the number. We’ll let the power supply run for up to 30 minutes during 100% load to stress the power supply.

What to watch for:
If you are doing this test on your handy-dandy home electronic load test machine, watch the voltage display on each module during the test. The test machine displays to the fourth digit to the right of the decimal point. That represents 1/10,000 of a volt. We can tell whether an output voltage is steady by observing the reported digits. You can consider the output to be acceptably steady if the second digit (1/100 Volt) fluctuates. If the 1/10 volt digit fluctuates, we know the power supply is not steady.

What to listen for:
Listen to the fan noise to see whether it is a quiet power supply. A power supply equipped with thermal controlled variable speed fan will be quieter than one without this feature.

Listen for a high-pitched humming noise. A good power supply should perform without significant hum. If the humming noise increases as the load increases, it usually suggests that the power supply is over rated for its design or components, or the quality of the power supply is otherwise poor.

What not to smell:
Any burning or overheated odor is a bad sign. This is most likely to occur, if it will, when the load is stepped up to 75% or higher. A poor quality power supply, especially overrated power supplies, will likely burn or blow out at higher loads.

What to feel for:
Use your hand to feel the air blowing from the exhaust fan. If the air is warm or getting hot, it indicates that the components are working very hard. It may also indicate that they are underrated for the work they are being asked to do. The cooler the air is, the cooler the power supply is running, which suggests a longer life and a longer Mean Time Between Failure (MTBF).

Test Results

Below you will find actual test result data for an Antec True430. For results of testing for other TruePower units, please click True330, True380, True 480, or True550.

Voltage readings in blue are the numbers recorded from the electronic load machine display. These are the voltage outputs under load conditions.

Total Actual Output in red is the calculated total output power in watts.

+5V +12V -5V -12V +3.3V +5V SB Calculated Actual Output (W) Total Theoretical Output (W)
LOAD A V A V A V A V A V A V
Rated 36.0 20.0

0.5

1.0

28.0

2.0

10%

3.5

5.01

2.0

12.26

0.5

5.03

1.0

11.97

3.0

3.35

2.0

5.08

76.75

73.65

25%

9.0

4.98

5.0

12.24

0.5

5.06

1.0

12.00

7.0

3.32

2.0

5.05

153.89

150.35

50%

18.0

4.94

10.0

12.19

0.5

5.11

1.0

12.05

14.0

3.28

2.0

5.01

281.37

278.45

75%

27.0

4.89

15.0

12.14

0.5

5.16

1.0

12.11

21.0

3.23

2.0

4.96

406.57

406.55

100%

36.0

4.85

11.5

12.14

0.5

5.19

1.0

12.14

28.0

3.19

2.0

4.93

428.13

432.65

100%

15.5

4.89

20.0

12.13

0.5

5.17

1.0

12.12

28.0

3.21

2.0

4.95

432.88

432.15

100%

30.0

4.87

15.0

12.12

0.5

5.18

1.0

12.13

24.0

3.20

2.0

4.93

429.28

431.45

110%

17.0

4.93

22.0

12.12

0.5

5.15

1.0

12.10

31.0

3.25

2.0

4.97

475.82

473.55

116%

22.0

4.96

22.0

12.06

0.5

5.16

1.0

12.13

31.0

3.23

2.0

4.94

499.16

500.80

Special Note: Antec without exception does not recommend that users put a load on a power supply larger than the unit is rated for. Antec without exception does not warrant that its power supplies will perform within specification at loads above the rated capacity of the power supply. Antec disclaims all liability for the consequences of using Antec equipment outside of its rated capabilities or not in accord with Antec's recommendations and requirements. Putting a load on an Antec power supply larger than the power supply's rating will void the manufacturer's warranty. You have been warned. The results given above are merely the results for the tested unit, shared purely for interest's sake. If you attempt to recreate this experiment, or otherwise use Antec equipment outside of its rated capabilities, you do so entirely at your own risk.

Voltage Regulation. Compare each actual output voltage in blue with Intel’s voltage regulation limits. Each must be within the specification limit. (± 5% limits for +5V, +3.3V, +12V and +5VSB; ±10% limits for 5V and 12V). A power supply exceeding the voltage regulation limit may cause the computer to crash or fail to boot, may damage some or all of the parts, and will almost certainly affect the longevity of those parts.

The True430 is designed and built for a server or a workstation with ±3% for +5V, +3.3V, +12V and +5VSB; ± 5% for 5V and 12V limits. We can see this power supply has met expectation.

The True430 design incorporates two features for increasing the stability of a power supply.

  1. Separate individual voltage output circuitry. With the separate individual output circuitry at +5V, +3.3V and +12V, a power supply can output at its maximum rated power without interference between the lines, and can improve the stability of its voltage regulation to +/-3% at positive end and +/-5% to negative end.

  2. Individual +5V, +3.3V and +12V voltage feedback circuitry. Each circuit is also designed to detect voltage change at the output end and compensate for the loss of voltage that occurs as electrical current passes along a wire. This enables the power supply outputs to be more stable than in a power supply lacking this design.

Total Actual Output Power. We can (and did) calculate the Total Actual Output Wattage to see whether the power supply has performed as rated. (If you like arithmetic and want to calculate it for yourself, the formula is Power (Watts) = Voltage x Current (Amps).)

For example, Actual Total Output Wattage =

(4.89V x 15.5A) + (12.13V x 20A) + (5.17V x 0.5A) + (12.12V x 1A) + (3.21V x 28A) + (4.95V x 2A)
= 432.88 Watts

Since the maximum power of True430 is rated at 430W, and since all the outputs are within the specified design limits, this power supply has met the manufacturer’s specification.

Maximum Power (Continuous Output). Don’t confuse Maximum Power ratings with Peak Power ratings. Peak power is the largest power a power supply can output for at least 15 milliseconds (15 thousandths of a second, or less than the blink of an eye). It really means nothing to a regular user with a real world PC. Always check and use the maximum power specification.

A good power supply should perform within regulation under maximum load conditions. We usually let the power supply run for at least 30 minutes in maximum (continuous) load conditions. We have found that overrated power supplies usually die in seconds at maximum load.

Liability and MTBF. When we design a power supply, we rate the power supply lower than the actual maximum power so that we won't overrun it. A power supply constantly running at its maximum power will have a shorter life. Like a car that can reach speeds of 180 miles per hour, the fact that it can reach that top speed does not mean you want to drive it from San Francisco to Chicago like that.

During our testing we increased the load from 100% to 110% and let it run for 30 more minutes. We do not recommend that you ever attempt to do this, and specifically disclaim all liability for the consequences of using Antec equipment outside of its rated capabilities or not in accord with Antec's recommendations and requirements. As a result of this test, we find that the True430 has passed the 110% load without any problem. After that, we continued gradually increasing the load until we found the point where the power supply finally fell out of regulation. In the case of the True430 power supply, we reached 116% of rated load before output fell out of the regulation. This is actually 499.16 Watts of continuous output.

In Conclusion.

There are other tests that a power supply must also be designed to pass. They require other equipment and different methodologies, and represent an opportunity for potential future articles.

For now, we hope this has been helpful. Even if you don’t have electronic load test equipment, or if yours is in for its 60,000 mile check, you can still look for signs of impending failure on your home or office computer.

Is the air exiting your power supply very warm? Hot? Is there a slight hot or burning odor? If the answer to any of these is yes, you may have a poor quality power supply or one that is about to fail. Even if the quality is good, you may be overloading it, and that means you are reducing its expected lifespan. And since you don’t know what will happen to your other components when the power supply fails, you might want to consider taking action and replacing the unit with a better quality or higher capacity one before it fails.