X-bit labs

ATX Power Supply Units Roundup. Part V (page 9)

Category: Other

by Oleg Artamonov

[ 11/21/2006 | 03:51 PM ]

When the PSU was working in pair with an APC SmartUPS SC 620, the UPS overload indicator would turn on at a load of over 330W irrespective of the power source (electric mains or batteries). There were no problems with the UPS at lower loads.

The high-frequency pulsation of the output voltages is negligible (less than 20 millivolts on all the rails), but there is a low-frequency (100Hz) ripple on the +12V rail that is as high as 64 millivolts at full load (the permissible maximum is 120 millivolts). The ripple subsides at lower loads, amounting to 28 millivolts at a combined load of 240W (half the maximum output power of this model).

The cross-load diagram of the Phantom 500 doesn’t have as much green as the previous model’s, yet it looks good anyway. The voltages deflect from their nominal values by no more than 4%. The range of loads typical of modern computers (a high load on the +12V and a moderate load on the +5V and +3.3V rails) is all green.

I had to measure the speed of the fan three times, for each of the three positions of the switch. The diagram suggests that only the first and second positions differ much. The third position of the switch doesn’t bring anything dramatically new into the behavior of the fan. In every case, the fan only begins to work when there’s a load of 200-300W on the PSU (I performed my measurements at a room temperature; in a real computer system the fan is going to start up sooner). The speed of the fan is a little over 2400rpm at the maximum.

As a kind of drawback, the fan speed controller doesn’t have some kind of hysteresis between the turn-on and turn-off thresholds. As a result, at a load of about 250-300W the fan periodically turns on and works at about 1400rpm. The PSU receives a portion of fresh air and gets cooler by a few degrees – and the fan turns off again for a couple of minutes. Ideally, the fan turn-off temperature threshold should be lower than its turn-on threshold. There would be no such cycles then. Well, there’s not a high chance that the power consumption of your computer will fit exactly into this range of cyclic turning on of the fan.

The efficiency of the Phantom 500 reaches 86% at a load of 200W and doesn’t change after that. That’s an excellent result, but quite expectable from a fan-less PSU. It would just overheat otherwise.

The Phantom 500 doesn’t have power factor correction (at least its American version has not – a PSU must have at least passive PFC to be sold in Europe), so its power factor is only 0.65. This also means that you have to manually select the input voltage with a switch (110V or 220V). Be careful when you first turn the PSU on – a wrongly set switch has been the death of a great many power supplies.

Generally speaking, the Phantom 500 is a rather expensive, but interesting option for people who care about silence. This PSU is absolutely silent under low loads. At high loads, its fan is very quiet, too. The other parameters are up to the product class. I have no complaints about the quality of assembly. The stability of the output voltages is excellent. The voltage ripple is within normal. I only wish the manufacturer implemented hysteresis between the thresholds of turning the fan on and off so that it didn’t enter the cycle of turning on/off with a period of several minutes in a certain range of loads.