Energy Consumption of energy efficient bulbs....

[ant]

Hi guys, I'm sure I know the answer is "they're no 100% efficient", but just wanted to check whether I have got some really wrong readings...or whether I'm missing some vital text on the packet which says these bulbs don't consume what they imply they consume...

NOw I may have been given a trojan horse of a bulb of course, but npower sent me a load of energy efficient light bulbs a while back (Philips branded).....I then received a free energy consumption monitor more recently...so I decided to do the @nul thing and test every individual appliance, bulb, cat and dog in the house as well as 'combinations' thereof to try to avoid the monitor giving crap readings where the consumption was below a level it could read...or where the increment was too small for it to notice.

Anyways...I came around to the 3 energy efficient bulbs from npower that I'd put in a 3 way spot light in the kitchen. Turn the lights on and the monitor shows 90W...I was somewhat surprised given I'd put in an 11W + 11W + 8W bulb which according to the bumf would actually give me about 60W + 35W + 35W = 130W equivalent in light..which as we all know isn't quite true but either way...lumens aside...if the box says

11WATT IN (little green arrow saying "Energy" pointing INTO the bulb...and big 60 WATT inside the bulb and the words "Light") surely it's saying this is only supposed to consume 11 watts regardless of what output it claims..so should be a total consumption of 31W..not 3 times that. Similarly, I turn on 2 more 11W spots, and a further 2 x 12W softone spots..so expecting a total consumption of the original 31 + 22 + 24 = 77W...the monitor reads 230W !

I know the energy consumption monitor thing 'could' be a bit rubbish but it does appear to be linear based on all the other things I've tested...so inclined to believe it's giving a fairly reasonable estimate...but maybe a numberof you will fall of your chairs laughing at me ;-) I know they're only supposed to be an indicator as opposed to a measuring device...but I'm sure it can't be 'that' far out....

Have npower just sent me a load of innefficient bulbs hoping I'd leave them on 3-5 times longer ? ;-)

Ant.

[Andy Burns]

Were all the other things you tested purely resistive loads (e.g. kettle, incandescent lamps, heaters, oven etc)?

CFL lamps (and PCs, most phone chargers and wall warts) draw current in a non-linear way, so are difficult to measure with the simple power meters.

[ant]

I measured everything...even the kids plugin night lights..for the PC/ chargers/Sky+/Dreambox/PS3/Amp I was aware the consumption would fluctuate..but the meter was also showing that, so I was able to compare 'normal' use (e.g difference between the plasma screen showing light scenes versus dark scenes of a movie)...PC consumption would be all over the place but I figured my desktop and network which are 24x7 are generally in an awake/active but low utilisation state so took readings individually for the PC/NAS/Router etc and then combined as a 'working environment' kind of measure to get a decent figure.

Now...for the energy efficient bulbs, are you saying that they may fluctuate in consumption so rapidly, that the meter would show a bad reading or just show their peak consumption all the time ? or that they need to be left to warm up for quite a long time before they drop back to the 11W/14W etc consumption ? (or both !)

I'm thinking of going round the house and totting up all the Wattages for the energy efficient bulbs I have installed...and turning them all on simultaneously to see if they average out to a more reliable reading...by the time the meter is getting up to 1Kw I'd expect it to be in it's comfort zone for accuracy and If I can do that with just 20 or so bulbs..would sugest the readings I'm getting are reasonable (I could also take a meter reading I guess......oh my kids and wife are going to love me tonight !....sorry, 2 hours, lights only..)

Ant.

[Bob Minchin]

These monitors are potentially inaccurate when measuring small loads and at all load levels are only going to measure power properly on resistive loads like kettles, electric fires etc. The meter on which your bill is based is sensitive to both current and voltage and the relative phases of both and so can work out true AC power consumed.

Bob

[Andy Burns]

With a "normal" bulb the current varies smoothly 100 times a second (with each half of the mains cycle) however with a CFL the current will typically vary tens of thousands of time per second, and the current will be very spiky rather than a sine wave.

Therefore the current and voltage samples that the power meter reads are not likely to be frequent enough to calculate power accurately.

If anything the true consumption is likely to go up when they've warmed up to full brightness.

[Tim Streater]

Then there's this:

gives you the power factor and so can display watts or VA.

[John Rumm]

Well indeed not, however you are not measuring the light output so efficiency hardly comes into it! ;-)

Bit of both really...

[snip tale of measurement woe]

The thing that is causing the confusion here is that the meter will give accurate results for some loads, while giving cloud cuckoo readings for others. The trick is spotting which ones are likely to be ok, and which will fool it.

All power meters will be fine on straight resistive loads - heaters, conventional filament lamps etc.

Some may also be ok on loads with classical non unity power factors[1] - things with induction motors like fridges, fans, CH pumps etc, or old linear strip lights, altho some of the poorer meters may be fooled by these loads and over read by a bit/ How much a "bit" is will vary, but multiplying the number they give by say 0.7 will give a closer answer in many cases.

Finally you get the really "difficult" loads to measure - the switched mode power supplies, especially the smaller ones like those in your CFL bulbs. These will fool most meters - even those that can cope with classic leading or lagging power factors - since these supplies draw power for only a short part of the whole mains cycle. So any attempt to computer power as a product of amps flowing and applied voltage is doomed to failure.

The fuller answer to all this is here:

[1] Have npower just sent me a load of innefficient bulbs hoping I'd leave

;-) hmm would be a nice scam for them I am sure, but in this case just your meter throwing a wobbly.

[The Natural Philosopher]

s/simple/badly designed.

[therustyone]

The important thing for your wallet is what the Electric Co's meter reads as the consumption. Does your house meter allow readings of

1/10 kwh so consumption could be checked by leaving the bulbs on for several hours.

rusty

[Mike Clarke]

No need to take so long as that, you can check power consumption in just a few minutes by watching the supply meter. The meter will have either a rotating disk below the dials (if the old analogue type) or a flashing led (if the newer digital type). Either way the front of the meter will be marked to indicate how many rotations or flashes correspond to 1 KWh. On our digital meter it's 1000 flashes per KWh, so if it flashed 1000 times per hour (3.6 seconds per flash) you have a 1KW load and if the OP had only his kitchen light with 2 x 11W and 1 x 8W CFLs (= 30W) connected it should be 3.6 x 1000/30 = 120 seconds per flash. That's with a meter like ours, other meters might have a different rate of flashes or turns but the principle would be the same.

[Ant]

THanks for all the responses here...very interesting info. I'll find a suitable time to power down the whole house ! and report back with my findings by reading off the meter...

[ant]

RIght...looks like this now makes sense ;-)

I powered the whole place down then turned on as many energy efficient lights as I could find accessible and totalled the Wattage for them up to 126W I then took meter readings every 10 minutes and after 1 hour had used very approximately 0.28 Kwh however my wife came home in the middle of my test and I inadvertaintly, being a good husband...put the kettle on...Doh!. Anyway, I reboiled after the readings to figure out it used about 0.042Kwh (2.6Kw for 1 Min) ...subtrating that from my reading gives 0.238 Kwh implying a draw of 240W which is double the ratings of the sum of those bulbs but not as bad as my original readings looked...I need to therefore run this for a longer period, and without boiling the kettle ;-)

Incidently, My meter has 800 Imp/Kwh next to the flashing light and I recorded about 29 seconds between flashes (eliminating the kettle error)...that works out as 1 Hour / 29 Seconds =3D 124 FLashes.......

124 / 800 =3D 0.155 Kwh...so 155 Watts which is much closer to what those bulbs added up to...

Ant.

[Ronald Raygun]

You have a problem here, because you've got two conflicting answers, one based on the meter reading and one on the flash rate. Since they're both from the same meter, the results you get should be the same!

Your flashes tell you 155W, your readings tell you 238W. The ratio is about 0.65, and coincidentally that's also the ratio of 19 to 29. Are you sure the flash intervals weren't 19 seconds instead of 29?

[ant]

the flashes were definately 29 seconds but I think the discrepancy is more likely to be the meter readings I took (since it's very difficult to read the 1/10th markings on the 1/10th scale (i.e 1/100ths)..to be honest I didn't even realise the 1/10ths were subdivided until I looked closely, and since the reader only clocked up 3/10ths during the hour mixed in with a silly kettle boil I figured the meter reading was likely to be erroneous...I'm going to retest over the weekend if I get chance as I don't like conflicting answers either ! Ant.