Measurement method for the IT power consumption page

The oscilloscope was a Tektronix TDS3052. The current in the neutral conductor was measured. The neutral was used in order to have a common, low-potential point for the current and the voltage measurement signals going into the oscilloscope channels (the oscilloscope was plugged into an un-earthed extension lead, to allow its inputs to have the supply neutral on their sheaths). If the measured device has an earth connection and a significant admittance to earth (e.g. from the filter circuit common in IT equipment) there could be a small amount more current, mainly reactive, in the phase conductor than in the neutral. This is not important for power consumption, but might be significant for reactive power in the off state.

In early measurements the current passed through a 1 Ω shunt consisting of ten parallel 10 Ω 0.6 W resistors. This was destroyed at the end of measurements of the first item, after which it was removed and a clamp-on current probe (Fluke 80i-110s) was used, set to 100 mV/A.

Oscilloscope Channel 2 (1 MΩ input, DC-coupled, 20 MHz low-pass filter) was used for current measurement. DC components in the current measured with the clamp-on probe should not be taken very seriously, as they may be from poor zeroing of the ammeter -- the zero-adjust was used until the oscilloscope showed about zero, but more precise ranges were not used. This is not important for active power calculations, as the voltage is quite close to sinusoidal.

Voltage was measured as the voltage across a 2.7 kΩ resistor in series with a 100 kΩ and 47 kΩ resistor, all in series across the supply on the supply side of the current measurement. Oscilloscope Channel 1 was used to measure the divided voltage, also set to 1 MΩ input, DC-coupled, 20 MHz low-pass filter.

The time-base was set to give 40 ms (2 cycles) on the screen, with the trigger level at zero. The voltage input (channel 1) was used for triggering.

When steady, representative waveforms were obtained on both channels, the oscilloscope's capture was frozen, to allow files of the 10000 values for voltage and current channels to be transferred to a computer. All calculations of power were done using the simultaneously measured voltage and current pair: i.e. instantaneous power.

Filtering of both channels' data was done before plotting, to remove HF noise by use of 8-point averaging.

It would have been preferable to have re-made each measurement once or twice, to confirm that the measured powers in the various states of use are consistent over time. This was not done, but the traces on the oscilloscope, before freezing the capture for data transfer, were seen to be either steady or varying only by a few percent.

Page started: 2009-02-13
Last change: 2009-02-13