/ Home & Energy

Why are you phased by the light bulb phase out?

Smashed energy-saving light bulb

When we wrote about the phase out of incandescent light bulbs recently, the response was overwhelming. So what got people’s backs up? I trawled through hundreds of comments to pick out the main complaints…

Your complaints ranged from cost to being left with less choice. Having tested energy-saving bulbs for the latest issue of Which? magazine, this topic is fresh in my mind, so I thought I’d try and tackle each of these issues one by one.

Unfortunately, I couldn’t address all the issues raised in our previous Conversation. Others included claims that energy-savings bulbs aren’t as bright, can’t be used with dimmer switches and don’t always fit light fittings.

1. Energy-saving bulbs don’t last as long as claimed

This can definitely be true if you have a bulb that’s poor for longevity, or doesn’t like being switched on and off frequently. The regulations only require 50% of a bulb’s sample to achieve the claimed lifetime – cold comfort if your bulb is in the half that didn’t.

Which? energy-saving light bulb tests show big differences between bulbs in terms of how long they last and how they cope with being switched on and off. We test three samples of each bulb and, at worst, all three fail long before our 30,000 on/off switching test is over. The best of the bulbs have all three samples working at the end of the test.

It’s the same with longevity – the top performers’ samples survive 5,000 hours continuously turned on – the equivalent of five years’ use. But poor performers will see some, or even all, samples fail before that.

2. Health side-effects including migraines, UV radiation and mercury

The most common type of energy-saving bulb, CFLs, do contain mercury, but only a small amount. An average CFL has no more than five milligrams – and would fit on the tip of a ballpoint pen. Breakages are easily dealt with if a bulb breaks, as detailed in our advice guide.

UV radiation is only potentially a problem if you sit very close (within a few centimetres) of a CFL-type bulb for prolonged periods. For migraine sufferers, using a different type of energy-saving bulb altogether (i.e. a halogen or an LED) or using a CFL with an outer surface as well as an inner (i.e. not a stick-shaped one), will also help.

3. Loss of heating from traditional incandescent bulbs

Incandescent bulbs did waste 85% of their energy as heat, but this is still a very small amount – the heating effect on any room would be negligible. Even using a higher-powered bulb of 100W would only give you around 85W of heat.

Compare that to a typical fan heater rating of 2kW and factor in that many of the bulbs will be in exactly the wrong place to heat a room – on the ceiling – and incandescents seem to be an inefficient method of heating!

4. No overall saving of energy

Some of you were concerned about the energy used to manufacture energy-saving light bulbs, but a 2009 Defra study found that incandescent bulbs had the highest environmental impact of all types of bulbs. That’s because “energy in use” is the key factor, and incandescents are much less efficient than energy-saving bulbs.

Our recent light bulb tests showed the same trait – that the most efficient bulbs were also the most environmentally friendly over the whole lifecycle of production and disposal, because energy used while the bulb is on is the most important factor.

5. Don’t tell us what to buy, give us a choice

It’s obviously annoying for a lot of people not to have a choice of which bulb to buy. But the phase-out is a “done deal” and although it forces us to buy only certain types of light bulb it does force us towards bulbs that will save us money and use less energy.

6. Energy-saving light bulbs are more expensive

The savings from energy-saving bulbs are nothing like as obvious as the extra cost of buying them in the first place, but they will save you money in the end.

We’re looking at ways to calculate payback times for all the bulbs we show test results for. In the meantime, any energy-saving bulb should more than pay for itself, but the amount will depend on various factors, especially how long it lasts.

Based on the bulb being on for 2.7 hours per day and electricity costing 14.5p per KWh, we calculate a 100W incandescent bulb costs around £14 to run for a year. An equivalent CFL energy-saver (such as the Osram Dulux Superstar 30W) would cost £3 or £4 in the same conditions, depending on how much light it emitted (watts being only a measure of how much power you pay to put in, not how much light you’re getting out).

7. CFLs are hard to dispose of

It’s true that disposing of CFLs isn’t as easy as incandescents – you can’t just chuck them in the bin because of the mercury. But we’re used to recycling much of our waste now, so isn’t it just a mind shift that’s needed to get into the habit of recycling bulbs too?

Recycling CFLs can be done at any Ikea, Homebase, Robert Dyas, 250 larger Sainsbury’s or at council facilities. Or you can find your nearest recycling points by typing in your postcode on Recolight on Recycle Now.

Do my responses make any difference to your view on the light bulb phase out? Do you think the phase out is all bad news or can you see some positives to the change?


The LED bulbs I bought were supposed to be dimmable. They were slightly more expensive than the non-dimmable version. I suspect, though, that they might need special dimmer-switches, although there was no mention of this in the instructions.

However, one has failed after only 6 months of use and I have returned it. I am waiting to hear from the suppliers as to what the problem might be. I have noticed, though, that when the LEDs are dimmed the electrics buzz and I am wondering whether this is indictative of a fault. I am now keeping them on full power in case the dimming is the problem.


The buzzing is likely to be due to the sudden increase in current through a small inductor, fifty times a second. It will probably disappear at full brightness, when the AC supply can pass uninterrupted. Dimmers can buzz even with incandescent lamps and hopefully the sound does not warn of imminent failure.

I still don’t know much about mains voltage LED lamps but have read a lot about 12 volt lamps where there are big differences between brands in radio frequency interference and resistance to damage by voltage spikes. The general advice seems to be to avoid cheaper products.

I am interested in why some brands of LED lamps are suitable for dimming and others are not. This may reflect a difference in circuit design or just that some manufacturers are being cautious.

I will be interested to hear what the supplier has to say, Richard, but I suspect that they will just send a new lamp or issue a refund.


Just more of the same old here, but I thought I’d let everyone know anyway.

Last night, at 8:44, (just before the end of Any Questions? if anyone wonders how I know the time so precisely) there was a glassy “TINK” noise from upstairs and I glanced round and out of the living room door to realise that the landing light had gone out..

5 mins later, when I finished listening to AQ, went to look and gently tilting of the glass globe revealed that it had loose broken glass in the bottom. Screwdriver out, released the globe, and the Pro-Lite 30W 1200K energy saving spiral, fitted Aug 12th 2012 (according to date I pen onto them each time now because I’ve had so many failures) was reduced to two “stumps” of glass protruding form a blackened plastic base, and 3 pieces of glass spiral tube in the bottom of the globe. ….. and a horrible stench of burned electronic components.

I had in stock a direct replacement, because I bought two at once last year given that it’s so hard to get 30W CFL’s anywhere round here, but annoyingly when I opened the box of the new one, I tried to fit it and it would not stay in the lamp holder. Quick look at the cap on the bulb and I realised that it had only 1 of the two bayonet pins fitted, and the other one wasn’t in the box so I guess it left the factory like that.

So I’ve had to shove a Sylvania Lynx “stick” type in, only 20W so rather dim, but I’ve had Sylvania Lynx’s that have lasted up to 12 years in other places so I reckon it’ll just have to do as I’m fed up of wasting money on useless cfl’s now.

I’m not up for reopening the whole debate about what’s wrong with cfl’s – I’m sure we all know what’s wrong with them: essentially they are made shoddily in order to make them cheap enough for the public to buy them, they are made too compact so they’ll retrofit older fittings (so the public will buy them), the electronics should be separate from the lamp (like in “2D” fittings) and they don’t give out as much light as we were all persuaded to believe (although the ratings have now been made more realistic) so we have to try to buy higher wattages than the government woudl like us to to get the right light. I don’t dispute any of that, and I absolutely agree that we need to save energy, but I am sick of being told that we will save money using these things: we don’t. Those two Pro Lite spirals cost me £7.99 each and I have not saved anything like £15 in electricity by using them (well, the one I could use) for less than 6 months instead of a 150w Incandescent that would have cost me, at most, about £1.50 to buy.

Anyway, there you go – more a “blog” for regular contributors and readers than anything else.

Have a great day everyone!!!


Can only say my CFLs generally last at least 5 years – Some I’ve had working since 1980. But they are NOT 30w but either 11w or 15w – The 15w bright enough for all my purposes.



From what I have seen in the past year or two, it is increasingly obvious that overheating of electronics is a major reason from premature failure of CFLs. Enclosed or semi-enclosed (including traditional glass lampshades) fixtures will result in the electronics running at a much higher temperature than in a large lampshade with ventilation at the top. I would expect CFLs to perform best in table lamps and standard lamps, where they are used cap-down. Here, the electronics will not be affected by heat rising from the lamp. I would have like to put CFLs in a fixture, but the lamps would be in glass shades and operated cap-up. A quick test with one CFL showed that it got far hotter than one used with good ventilation, so I went back to 60 watt bulbs. I am now looking for affordable LED lamps.

Poor quality electronic components will not help but it’s asking a lot for even good quality electronics to survive at the temperatures they have to operate at in a CFL. Compared with having the electronics separate from the lamp, the CFL is a flawed design. We have been let down by the manufacturers of light fixtures.


Absolutely agree with Wavechange, although I will be very interested to see how my Sylvania “stick” CFL fares, since my mum has three of the same bulb (cap up) in her fully enclosed, glass, 1950’s dining room light fitting, and (touch wood) they have all been in since 1998, just after my father died, and have not yet failed. Mum has that light on from well before dusk until gone midnight every day, so they have done many many times more than the stated 8,000 hours on the carton. If they can manage that then I’m hoping the one in my landing globe will also do well, and if that turns out to be the case I have to say I’ll be more inclined to argue that the issue is in trying to make the electronics too cheaply. Only time will tell us on that one.

Out of interest I had to replace the 5w CFL in one of my outside lights just after Christmas. THe other one had done 11 months and I did note that it failed on the first evening we had really low (as in sub-zero) temperatures. I am minded to wonder if sudden temp changes are as much to blame as consistently high temps, because the glass on that had cracked (but not broken) too. I’m thinking thermal shock when it starts up?


Dave D, I don’t know the origin of Pro-Lite lamps, but missing a BC pin suggests quality might be iffy. There are lots of CFLs around that may not be up to scratch.

A 30W CFL in an enclosed globe may well have overheated the electronics, and overheating the spiral will also reduce light output. Even so a decent quality lamp should not have come apart.

I’d stick with the household names if you want decent quality – Osram, Philips, Sylvania, GE for example. These would be expected to deliver the life and light output they claim. CFLs from reputable sources are effective and do save significant energy.

I have said elsewhere just what you say about integral electronics – they are technically worse than the separate electronic controllers that you will get in commercial CFL products, and a waste. Why the EU does not promote them beats me – but I suspect domestic consumers would not fork out the money for a decent new light fitting when they can buy a cheapish retrofit lamp. Its curious how we spend £000s on kitchens, bathrooms and holidays and yet skimp on other things.


Thermal shock could be greater if a fluorescent lamp is struggling to start up in cold conditions, due to increased power dissipated by the heaters at the end of the tube.

In the 90s I had Philips 20 and 23 watt CFLs survive in enclosed fixtures but the plastic became brown, showing how hot they were getting. Electronics can survive high temperatures, and I have seen some amazing examples (unrelated to lighting) but there is no doubt that keeping electronics cool can make a big difference to longevity. Electrolytic capacitors are probably the weakest link.

Having said that, it is really interesting to hear of CFLs that have survived adverse conditions. It would be interesting to know how much the plastic caps have discoloured as a result of overheating.


@wavechange: I have a number of CFL’s in various fittings, including 5w CFL’s in bedside lamps which have sizeable “drum” shades – completely open at the bottom and top and rated to take 100w tungsten lamps. All of the CFL’s have brown, in some cases nearly charred, plastic caps, regardless of whether the caps are up or down, including those which are still functioning fine.

Quite apart from the scientific aspects of this, which I think you and I and many others agree on, I still think that there is a huge variability in quality of components and in their tolerance levels.

My kitchen has a ceiling light which I must say I don’t often use but it’s a completely enclosed white glass globe rated for a 60w tungsten lamp. It has in it a Wickes 23w CFL (cap up), which has been in since I bought the fitting (also from Wickes) well over a decade ago. It’s not failed yet and that’s in adverse conditions – in fact it sounds like it’s almost identical to Richard’s bathroom lights. I also have under and over cupboard traditional fluorescent strips (8 foot above the wall cupboards and a 6 foot and a 2 foot below them). They last ages and I have them on almost every moment that I am in the house. But as you and I know traditional strips with external control gear are infinitely superior to CFL’s. Then there is a wall light, It’s a plaster uplighter with a completely open top and has an 11w cfl in it mounted horizontally. That has been replaced twice in the last 12 years and that’s on the same circuit as the cupboard lights and so is on for huge lengths of time. The first CFL it had was a wickes one, I can’t remember what the second was and the one in now is Crompton. The caps on the 1st and 2nd were badly burned and brittle when they failed.

I don’t knwo who makes Wickes’ cfl’s for them, but they seem to last a very long time and are comparable with Sylvania and Crompton which I find do really quite well. Conversely Pro-Lite, Mazda and B&Q’s own CFL’s have all been unmitigated disasters for me, regardless of where used, but Pro-Lite and Mazda are the only brands I can find in shops or on line where you can get higher wattages (above 23 watts).

Now, does that mean that Wickes’ and Crompton are better made, or does it mean that higher wattage CFL’s are simply unstable and so the bigger name companies won’t make them because they are a poor deal?

Which? don’t help because (as has often been pointed out to them forcefully before) they don’t test brands which people outside the M25 can buy easily, nor do they test higher wattages, so it’s really hard to know what the real story is.

Perhaps we conversationalists should start our own research project between us.


I think the lighting conversationalists would have to jump ship to another forum if we wanted to do some serious research. I would be very interested in an analysis of failure modes of modern CFLs sold in the UK, circuit design and choice of components. It would be a help to examine the workings of CFLs, but it’s not easy to dismantle them. On the rare occasions that I have a failure I do try, but generally give up in disgust.

If you check the Which? website there is a photo of CFLs under test. They are not in any sort of shade, which I believe is not representative of how the lamps are likely to be used in a domestic environment.

To test the hypothesis that cheap CFLs might not be durable I have been using a couple of 950 lumen (15 watt) supermarket spirals in lights that are on all evening (cap up and ventilated), and apart from slight browning of the plastic, both are still working fine. Hardly a scientific test, I know. It would be very interesting to know which companies supply the ‘own brand’ CFLs, but bear in mind that this could vary.

I would be happy to give Richard some 11 watt Philips CFLs that I bought for 10 pence each. They are not really bright enough for me, though I have never needed anything as large as a 30 watt lamp.


I have two “stick” cfls in the fully enclosed luminaires in my bathroom. They are mounted cap up and completely surrounded by a globe – probably the worst conditions they could experience.

They have both been installed for about 5 years and are still working perfectly – although, of course, they are not used continuously.


I have seen plenty of other examples of enclosed fixtures in bathrooms, Richard, and the fact that they are not in continuous use probably helps – as you say.

There are sensible rules about lighting in bathrooms and when I decided to replace my enclosed bathroom light with its overheating CFL I decided that it probably was not worth using energy saving lighting in the bathroom. I chose a fixture with four 25 watt halogen bulbs, which is fine but I’m fed-up having to replace bulbs that have failed prematurely. At least I am never left in darkness.


The quality of CFLs will vary enormously, so I’ll repeat that if you want reliability and performance, use a reputable manufacturer (e.g. Philips, GE, Osram-Sylvania). Where they might get them manufactured (e.g. China) is not an issue, it is the design, component quality and quality control that matters.

Wattages are available up to 42W, but these should not be necessary at home. The light output from reputable makes up to 23W should be fine for most applications.

Operating position is universal (cap up, down, side) unless the packaging says different.

Useful information on lamps can be found at http://www.thelia.org.uk/lighting-guides/lamp-guide/


@Malcolm R – You make an excellent point about using reputable brands, however Which? fails to recognise (when testing and reporting) the lack of availability of some brands, including some of the best brands – which in turn includes some that you give as examples, in many shops up and down the country.

As for wattages available and what should (or should not) be “necessary” in the home, well that rather depends on people’s homes and we should not be forced into using multi-lamp fittings, multiple fittings and other “work arounds” to compensate for inadequate light out put or non-availability of suitable lamps. When you have a single lamp lighting a 13 foot square dining room, or a single lamp lighting a 2 story stair-well (both of which I have) which used to use 250w or 150w tungsten lamps respectively, it should not be necessary to buy and install extra fittings into which to install extra lamps just because [readily available] CFL’s can’t offer the same light output as the old lamps did. That said, I think your first point about quality and brand reliability does make a difference to the light output too. The Sylvania Lynx “stick” cfl’s rated 23w seem to give really very nearly the same light output as the Pro-Lite 30w spirals and the Crompton 5w spirals most certainly give a much greater light output than the B&Q own label 11w spirals. This (apart from coming back to the old argument about the LUMENS OUTPUT being what matters rather than the watts input) suggests to me that the better brands actually give a better light and, if I’m right, is another reason why Which? should be testing more of the readily available brands so as to expose any anomalies in this regard.


In fact, apart from the early failure of one exmaple, I am very pleased with my LEDs and will probably not nuy any more CFLs.

The LEDs light immediately, use only about half the current of CFLs, and are supposed to last many times longer.


Richard, glad you are pleased with LEDs. Like CFLs, buy a good make, which won’t necessarily be cheap. That way you will get good light output and reliable electronics.
Be aware that if you want a colour similar to CFL or GLS lamps (e.g. Warm White) they will look pleasanter but not be as efficient. Some of the blue-white ones can look pretty nasty in the home, but should be OK outside.
Efficiency is around the same as CFLs but they are very sensitive to temperature – their light output can drop significantly if they run too hot by, for example, being high power in enclosed fittings without a good heat sink.
Their life will be similarly affected but if well-constructed should last 50 000 hours (that’s 12 years if left on all night!).


According to their literature they are about twice as efficient as CFLs and last over twice as long.


Efficiency is often quoted under perfect conditions, not in the situation where you will use them, so must be taken with a big pinch of salt. And it depends which colour you compare – as I said, the good quality warmer colours are less efficient. LEDs vary from 50 – 100 lumens per Watt (subject to white colour and operating temperature) with CFLs starting at 45 Lm/W and upwards. To get the equivalent light output to a CFL will require multiple LEDs which, if grouped together, will get hot and reduce their lumens.
Twice the life is OK.


My LEDs are single “bulbs” containing 6 separate LED “units”. They do get quite warm but I assume that the efficiciency quoted is for the units, not a hypothetical efficiency for each LED on its own.

Of course, when buying a product right at the beginning of its life-cycle once always takes a chance; I have no doubt that the equivalent of my LED units will, in a couple of years time, be half the price and twice the efficiency.


Half the price or a tenth of the price perhaps, but the maximum efficiency of the lamps is governed by physics. I’m not sure where we stand at present.


The colour of an incandescent lamp (2700 K) has been traditional for home use, with the exception of kitchens and utility rooms – where a colour temperature closer to daylight is acceptable or preferred. One of the reasons why halogen lights have proved popular is the whiter light and I would not be surprised if there is a gradual shift in preference towards higher colour temperatures in the future. Even though LED lighting is relatively new, there does seem to be a greater range of colour temperatures than we have seen with CFLs. With LED lamps having many light emitting elements it may be possible to have ones that offer some adjustment in colour temperature.


GLS lamps have a colour temperature of between 2500-2700K. Tungsten Halogen 2700-3200K, which is a whiter, but still warm, colour.
White LEDs can go from 2500 – 7000K as they use similar phosphors to fluorescent lamps. Colour temperatures above 4500K can seem too cool for domestic use.
White LEDs are in fact a blue LED with a phosphor coating. In principle by using different colour-temperature LEDs on dimmers close together you could vary the overall appearance. Another way is to use say red, green and blue LEDs and alter their proportions by varying their outputs. Commercial colour changers use this method. Not too sure why you would want to do this at home – experience suggests that once the novelty wears off, variable things stay set. Dimming is probably the best way to vary the appearance of a lit space. Suitable CFLs as well as LEDs can achieve this.


One reason for changing the colour temperature is to match one lamp with others in a set, since differences between lamps are very obvious.

Most people will be familiar with the production of different coloured light – including white – from the red, green and blue phosphor dots on a cathode ray tube, which could easily be seen on close inspection.


The colour temperature of the LEDs I have is much higher than that of the CFLs. I have both ceiling lights and wall lights and the difference is very noticeable. In fact I prefer the whiter light from the ceiling mounted LEDs – but, as the CFL wall lights are primarily decorative, it’s not a major issue.


wavechange – the normal reason “white” LEDs differ in colour temperature is because they are poor quality. Reputable manufacturers select LEDs to place them in groups of carefully-controlled colour temperature – called binning – so you don’t get this problem. It should not be necessary to adjust yourself if you buy wisely!


Thanks. I am familiar with binning in the context of selection of semiconductor devices with similar characteristics. For example, some transistors with a particular code number were colour coded according to their gain range.

I have not bought any white LEDs, but it is not uncommon to see a row of downlighters with an odd one. Even LED lamps will not operate indefinitely, so the ability to colour match may be desirable.

Safety is a good enough reason for avoiding cheap LED lamps.


David D above asks about the availability of reputable brands. I haven’t bought lately – a few years ago I stocked up from Robert Dyas when they were 10p each! Philips and GE brands and I’m still living off that hoard.
However, looking on line:
Robert Dyas sell Osram
John Lewis (and presumably Waitrose), Amazon, Sainsburys and Wickes all list Philips.
Not 10p each now, I’m sorry to say.


I’d never even heard of Robert Dyas until Which? started to list them, but the nearest Robert Dyas store to me is Solihull, which is about 120 miles away.

Sainsbury’s stores near to me (3 branches I use semi regularly) appear only to have cfl’s up to 13watts.

Wickes are good – own brand seem to be very good indeed (maybe they are made by Phillips?) and both stores up here have a decent range, but nothing over 20watts.

Waitrose (where I do almost all my ‘normal’ shopping have the poorest range (nothing lower than 11w and little or nothing higher than 13 watts but, worst of all, almost all screw fitting not BC. I don’t go into John Lewis in Sheffield because their staff are so abominably rude, but I don’t image what they stock will differ much from Waitrose since it’s all the same company.

I absolutely agree that on-line stores (including the online version of all above) makes it much easier to get lamps (and indeed anything else) but you then have to add the cost of P&P and the risk of damage in delivery, plus the (usually very short) wait for delivery and, we should remember, not everyone is willing or able to do online shopping.

Whilst I personally have no trouble shopping on line and also have a superb local independent specialist electrical shop which stocks and will order on demand anything from BELL, Osram and Sylvania, this comes back to the point that Which? don’t appear to test what is readily available in the high street, unless you are in the affluent south-east.

What would really help would be if Which? tested (and where appropriate ‘slated’) all the stores’ own brand lamps, and brands like Pro-Lite which seem to be available in every corner shop, market stall, co-op, lighting store and so on and so on. Possibly then these manufacturers would up their game, or other storse would stock a better range of the higher quality competing brand?


Most of the CFLs that I have used have been made Philips, probably because I started off with using the Philips ‘jam jars’ (choke ballast switchstart) that Dave and I have discussed at length. They have proved very durable and I also had success with the Osram and GE CFLs that I put in my parents’ house. As an experiment, I bought a couple of cheap spirals and a small reflector CFL from Tesco and they are doing well so far. I bought some 11 watt Philips Genie sticks for 10p each but have not managed to find a use for such small lamps. If Malcolm has found a use for his cheap lamps, he has achieved more than me.

I have a friend who had numerous failures of Megaman reflector CFLs in his large kitchen. When I stayed with him at Christmas, all the lamps had died and been replaced with incandescent reflector bulbs. I suspect that the lack of ventilation provided by the downlighter fixtures contributed to their demise, and there are plenty of negative reviews of these Megaman CFLs online.

Over the years I have seen numerous CFLs that have been killed by use with dimmers and in motion-activated outside lamps, thanks to users failing to follow instructions.

David Parsons says:
1 October 2016

I puchased a 30W CFL (called a energy saving stick) from Morrisons today. It claims to give the same light as a 130W conventional bulb, 1897 Lumen equivalent. Cost: £3.91