Rediscovering the Power Factor
…or, why clip-on power meters can give wildly misleading readings
Before I installed my solar PV system I wanted to get an accurate idea of how much background power I was using, to decide how big a battery I needed to cover overnight usage. I also wanted to know where in the house it was going; and whether there were any quick wins I could get by turning off power-hungry devices.
This turned out to be a longer voyage of discovery than I was expecting. Eventually some long-forgotten electronics knowledge saved the day!
TL;DR: power meters that clip onto your incoming electricity supply can give wildly misleading readings.
The starting point
I knew from my utility meter readings that I use around 3,600kWh per year, or 10kWh per day.
Dividing by 24 gives 0.417 kilowatt-hours per hour — or more simply, 417 watts. That’s my average power consumption. That seems quite high, given that devices on standby ought to use only 1 or 2 watts each.
We use electricity for cooking, so I know that will account for a big chunk of it (although hot water and central heating are gas). The kettle is used many times per day, for drinking tea. Then there are the washing machine and tumble drier and so on. All of these things consume a lot of power when running, but nothing when idle.
I wanted to find out how much of the usage is continuous background power drain, also known as “vampire power”, and whether there was anything I could do to reduce it.
Take 1: The clip-on power meter
Many years ago, I bought an “ElectriSave” wireless power meter, a model which is no longer on sale. This has a sensor that clips onto the incoming mains feed at the consumer unit, and plugs into a transmitter. There is a separate display which shows your usage. Both parts are battery powered.
I decided to dig this out and give it another go.
Firstly, I switched off the fridge and freezer. These are things which turn off and on intermittently, so would make accurate measurements difficult. I reasoned that I can make separate measurements for these later. In any case, we can’t live without them, so they’re not really “vampire” power.
Under these conditions, the power meter told me that the house was using: 310 watts!
Hmm, that’s more than I was expecting. 300 watts is a significant amount of power ultimately ending up as heat, and indeed not far off my total usage. That’s excluding big consumers like the oven. Could it really be that the majority of my usage is down to vampire power? I wanted to find out where this was power was going.
I switched off all the breakers in the consumer unit, and switched them on one by one. There was no obvious single circuit which could account for all this usage.
- Many circuits used nothing (e.g. oven, hob, garage, doorbell)
- One downstairs lighting circuit used 120W — with no lights on! However this circuit also has the TV amplifier in the loft, and the bathroom extractor fan
- Switching on a couple of rings used 120W — despite only having a few things on standby.
Something odd was going on. But if nothing else, it was clear that the resolution of the ElectriSave was poor: although it displays increments of 0.01kW (10W), I only ever saw it jump up in increments of 30W or 40W. And it only updates a few times per minute, making it very hard to pin down the source.
Time to get a better device.
Take 2: The current clamp meter
My ancient Maplin multimeter has a dodgy switch, so I decided to treat myself to a new meter, one with a current clamp.
This fits around the incoming mains wire, just like the ElectriSave.
Under the same conditions, here’s the measurement I got:
That’s 1.344 amps. The supply is 240 volts, so we get
Power = volts x amps = 240 x 1.344 = 322 watts
So the ElectriSave was right all along! Or so I thought.
Take 3: Measuring individual devices
One benefit of the clamp meter, though, is that it has extremely fine resolution, down to 0.001A (1 milliamp), as well as refreshing several times per second. This meant I could turn individual devices off and on, and see how that affected the power in real time¹.
Taking one example: with just the kitchen socket ring active, I measured 0.146A. By turning things off one at a time, and seeing how much the current dropped, I was able to break this down into:
- Fridge: 0.053A (it was plugged in at this point)
- Microwave: 0.012A (it has a built-in clock)
- Sonos Play 1: 0.062A
- Everything else: 0.019A
The Sonos, which was sitting there connected to the wireless network but otherwise not doing anything, appeared to account for 0.062 x 240 = 15 watts. This is an old model, but that’s pretty bad for a device which is more or less on “standby”.
The same story was repeated on other rings and with other kinds of devices I tried. 10 watts here, 15 watts there, and it soon adds up to the 310 watts of background.
I wasn’t happy. I was missing something.
Take 4: The Plug-in Power Meter
The next step was to directly measure the power consumption of individual devices, rather than the total power on a ring. That meant buying a plug-in power meter:
Then I went round trying some of the same devices as before.
Strangely, each individual device seemed to be using much less power. I couldn’t get them to add up to what I had found before. Was the plug-in meter inaccurate?
Let’s take the Sonos Play 1 again. Previously I thought it was using 15 watts. But the power meter says only 3.7 watts:
That’s odd, although it’s a more reasonable value for background power usage. Anything else it can tell me? Pressing the “Function” button, another screen shows me the voltage and frequency:
That’s what I was expecting. The next screen shows me the current:
0.070A is close to what I had measured indirectly before (0.062A), simply by turning the Sonos off and seeing how much the current measured by the clamp meter fell.
But then the penny dropped. The answer was staring me in the face — right underneath the current reading.
The Power Factor
All along, I had been using this well-known formula:
Power = Voltage x Current
And you can hardly blame me, because the ElectriSave was doing this as well.
This formula works for Direct Current (DC). But as it turns out, it doesn’t necessarily work for Alternating Current (AC). For AC, the real formula is:
Power = Voltage x Current x Power Factor
I had learned about the power factor many years ago, but partly I had forgotten about it, and partly I thought it wouldn’t be relevant in a small domestic environment. I was wrong.
I’ll try to explain it properly later. But in short:
- Multiplying the voltage and current gives you the “apparent power”. This is measured in volt-amps (VA) rather than watts (W)
- Multiplying this by the power factor gives you the “true power”
- The true power represents the real work that the electricity is doing
Why the clip-on meter is wrong
The clamp sensor on the clip-on meter measures only current, using the magnetic field emitted by the mains wire as current flows through it — just like my clamp multimeter.
In order to display a power reading, it has to assume that the voltage is 240 volts, and the power factor is 1.
For many large consumers of electricity in the house, particularly those devices which mainly convert electricity to heat or light, the power factor is indeed 1 or very close to it, and so this assumption is fair.
But for small devices this isn’t always true. What the ElectriSave shows is apparent power, not true power.
Take 5: The utility meter
What’s really important here is what I’m charged for. Does the utility company charge me for apparent power, or true power?
I read online that it’s true power, but I wanted to convince myself of this.
If you have a smart meter, this is easy (and you wouldn’t have bothered doing any of the things I’ve done in this article). But even without one, you can watch the meter itself to get a near real-time reading of power.
Old meters have a disc which rotates. Modern meters have a light which pulses periodically, and a label will say how many pulses it generates per kWh of energy usage — mine is 1,000 pulses per kWh.
What you do is watch the pulses (or disc rotations), and measure the average time between them.
There are smartphone apps available which use the camera to watch the LED pulses and perform the above calculation for you — if you can be bothered to hold your phone in the right place for long enough. I just used a stopwatch instead.
Here’s what I found, when trying it under the same conditions (fridge and freezer turned off):
- There are 26 seconds between pulses
- This means there are 3600/26 = 138 pulses per hour
- Each pulse is 1/1000 of a kWh
- Therefore I am using 0.138kWh of energy per hour. Or equivalently, my power consumption is 138W.
That’s way less than the ElectriSave was telling me. This implies that all of these devices together have an average power factor of about 0.43 (138W true power, divided by 322W apparent power)
Or put it another way: the ElectriSave was showing a value which was nearly 250% of the true power value. That’s grossly misleading.
Take 6: The true-reading CT meter
Unfortunately, the power factor is not constant. It varies from device to device, and how active that device is.
Always wanting to get to the bottom of a problem, I needed to know if there was a way to measure the true power to the house, using a clamp meter (so I didn’t have to touch the wiring) but which was also able to compensate for the power factor.
It turns out there is. What you need is a meter which plugs into a mains socket, so it can sense the voltage and phase, and has a clamp connection, also known as a CT (Current Transformer).
Some of these are expensive, and I didn’t want to spend a lot of money just to confirm a theory, but I found a really cheap module which can do it.
BEWARE! This is not a piece of consumer electronics. It has four screw terminals at the back. Two of them are where the CT connects, and two of them need to be connected to Live and Neutral of a mains supply. I connected them to a cable with a mains plug, and then I stuck some insulating tape over the back, and took care not to let my fingers go near that end.
Don’t attempt this unless you’re sure you know what you’re doing. This is designed to be installed by an electrician, and really it’s a module that should be installed inside a permanent panel.
Anyway, when plugged into a nearby socket, and with the clamp applied to the main feed into my consumer unit, this is what I saw:
There you go. A true reading of 134 watts, matching the utility meter, and showing the power factor of 0.42. Success at last!
If it stays like this all day, then over 24 hours this background usage accounts for 0.134 x 24 = 3.2kWh of electricity — about a third of my total usage. And as I said before, the rest comes from things like the fridge, freezer, cooking and so on.
The module, by the way, is surplus to requirements now. It can’t tell the difference between power imported and exported, so it’s not any use with a solar PV system.
Measuring the fridge
The plug-in power meter that I bought (see “Take 4” above) also allowed me to find out how much the fridge uses.
The fridge is something that switches off and on intermittently, so watching its instantaneous power draw is not useful. But the plug-in power meter also counts units of energy used, just like a utility meter.
So I simply left the fridge plugged into it for about a week. From this, I was able to determine that the fridge used 7.840kWh over 7 days and 18 hours. That works out as
7.840 / (7x24 + 18) = 0.042 kWh per hour
… or more simply, 42 watts on average. (Exactly what any Douglas Adams fan would expect).
A similar exercises for our standalone freezer came up with 25 watts — presumably because its door is opened less frequently.
Conclusion
Clip-on power meters are still useful, and will give accurate readings for your use of electricity for lighting, heating, cooking, and laundry: these sorts of devices typically have a power factor of 1, as well as high power consumption.
But for small devices, a clip-on meter may give you an exaggerated idea of how much you could save by switching them off. It’s much better to get a plug-in power meter to find out what these devices are really using, and therefore, what they are really costing to run.
The rest of this article is to explain where the Power Factor comes from. If you’re not interested in that, please feel free to end here.
Postscript: Understanding the Power Factor
So what exactly is this Power Factor, and from where does it arise? I’ll try to explain it in simplistic terms.
The electricity supply in your house is Alternating Current (AC), which switches directions repeatedly — in a UK household it’s 50 cycles per second (50Hz). AC power naturally arises from the rotation of generators in power stations. It’s also convenient because it allows transformers to be used to step up and step down the voltage.
Over a cycle the voltage rises and falls, and then switches direction, following a sine wave.
With a normal “resistive” load, as the voltage rises and falls it causes a corresponding rise and fall in current — in accordance with Ohms Law. Hence the current is in phase with the voltage.
When the voltage reverses, the current reverses too. At all times, power is going into the load, and is consumed by that load.
However there are other types of load, called “reactive” loads, which temporarily store and return energy. There are two kinds of reactive load: capacitors store energy in an electric field, and inductors store energy in a magnetic field.
Real devices have a combination of resistive and reactive behaviour. Power is consumed by the resistive part, but stored and returned by the reactive part.
In practice, what you see is that the current is out of phase with the voltage: it either leads or lags.
During certain parts of the cycle (highlighted in blue), the voltage is positive but the current is negative, or vice versa. During those parts of the cycle, energy is being returned from the load.
This explains why the true power consumed is less than the apparent power. When the voltage and current are out of phase, sometimes energy is being transferred to the load, and sometimes energy is being returned from the load. The true power is the net power which the load consumes. Your electricity meter doesn’t penalise you for “lending” some energy briefly, if you get it back in the same cycle.
The role of transformers
Many devices in the home need low voltages to operate, and have transformers inside them.
A transformer consists of two coils of wire wound together so as to be magnetically coupled. The primary coil creates an alternating magnetic field from the incoming supply, and the secondary coil turns that magnetic field back into a voltage (which may be higher or lower than the original, depending on the number of turns in each coil).
A transformer winding is inductive, so if the device isn’t drawing much power from the secondary coil, it appears as a reactive load.
What effect does the Power Factor have?
Resistive loads have a Power Factor of 1, so the true power and apparent power are the same. You can calculate the power using “voltage x current” as normal².
Devices with a reactive part have a power factor less than 1. As described above, power flows into them during part of the AC cycle but flows back out again during another part of the cycle. So overall, the power consumed is less than you’d expect just by looking at the current alone.
Or looking at it the other way: the current is higher than you would expect, when transferring the same amount of power into the load.
If you’re charged only for true power, does Power Factor matter?
The answer is, not much.
Most of the power you use is consumed by the devices themselves, but a small amount is wasted in the wires that deliver power to them. The wires have some resistance, and the current flowing through them makes them heat up a little.
With a power factor less than 1, the current flowing through the wires to your device is higher than it would otherwise be. So whilst the device may be using the same amount of power, the amount of power lost to heating the wires is increased.
In practice, this is usually tiny, compared to the power consumption of the device itself. If you notice the cables to your devices feeling warm, then you have a problem, and you should speak to an electrician immediately.
Power Factor Correction?
It’s possible to correct for the power factor, by adding extra components to bring the current and voltage into phase.
There are some devices with high power draw where power factor is important, primarily those with electric motors. Any devices you have like this will most likely have power factor correction built in. Without power factor correction, the increased current could cause your circuit breakers to trip.
Some people will try to sell you home power factor correction systems. Don’t buy them. At best, they will reduce the current flowing through your wires, but they will not reduce the true power consumed by your devices, and hence they won’t reduce your energy bills. The reduction in resistive heating from your home wiring will be negligible.
Such devices might be able to make the power shown by an ElectriSave-style clamp meter drop dramatically — but now you know why that’s not a true reflection of the power you are consuming.
¹You can’t just put the clamp meter around an appliance cable, because the live and neutral currents are in opposite directions and the magnetic fields cancel each other out. You have to put it around the live wire or the neutral wire only. In my case, I was using one of the thick cables coming into the consumer unit.
²The voltage and current are also varying throughout the cycle. But by using a special kind of average, called the Root Mean Square (RMS), you can multiply the RMS voltage by the RMS current to get the average power.
