That assumes people are using the CFL they got cheap. I'm not so sure they are.

For the LEDs I've just purchased...

5W LED vs 50W halogen.

Difference = 45W or 0.045kW
Cost of electricity = 0.14p/kWH
Cost of LED = £6.60

Hours = 6.60/(0.045*0.14) = 1047 hours

How many hours is day is it on for?

@ Three hours a day = 350 days. Call it a year.

5W LED vs 11W CFL

Difference = 6W or 0.006kW
Cost of electricity = 0.14p/kWH
Cost of LED = £6.60

Hours = 6.60/(0.006*0.14) = 7857 hours

@ Three hours a day is a little over 7 years

or leave them on 24 / 7 and they "pay for themselves" in less than a year

And ignores the reduction in heating demand created by heating effect of bulbs.

And that most bulbs are at almost ceiling height and warm air rises.

<blockquote><cite>Posted By: martin.n</cite>HollyBush, a 5W LED is nowhere near as bright as an 11 watt CFL. </blockquote>

It sure can be.

There is no such think as _a_ 5W LED. You can buy 5W LEDs that produce anywhere from 150 to 400 Lumens so it very much depends which 5W LEDs you are talking about. Best avoid specifying LEDs by their wattage. Allways look for Lumens in the data and go by that. Use the wattage to work out the efficiency once you have found LEDs that are bright enough.

I recently bought a light meter and measured several LEDs in the range 3W to 6W that claimed 350 to 450Lumes and compared them with an 11W Megaman I was using. The better LEDs were brighter than the Megaman and I won't be buying CFL ever again. LEDs producing >350 Lumens are now down to around £7-8 on ebay.

See the thread I started on the subject..

http://www.greenbuildingforum.co.uk/forum114/comments.php?DiscussionID=7769&page=1

I have recently replaced all the 50W halogens in my kitchen with LEDs (mix of brands all using the 5050 LED die and claiming 350-450 Lumens) and the result is brighter!

One big difference is the beam angle. The beam angle of the halogen is narrow and the LED very wide - so if you measure or look at the brightness directly below a single halogen it looks much brighter than below a single LED. eg The light output of the LED is spread more thinly. However when you replace a grid of them the spot below one LED is also illuminated by the other fittings and that seems to restore the brightness. It also seems to reduce shadows.

The two outstanding issues for me are the reliability and dimming. I haven't yet looked for the latter.

Would be interesting to know how they failed? Did they gradually get dimmer? Suddenly go out? etc.

<blockquote><cite>Posted By: qeipl</cite><blockquote><cite>Posted By: CWatters</cite>Would be interesting to know how they failed? Did they gradually get dimmer? Suddenly go out? etc.</blockquote>

They went 'pop' and tripped the MCB at the CU. When I removed them there was evidence of combustion on the ceramic(?) cap between the terminals.</blockquote>

Sounds like the driver circuit may have failed. My limited experience has been that the LEDs themselves rarely fail, but the driver electronics in the base often do, probably because of over-heating (or maybe just poor design). The same seems true for cheap CFLs.

It's the reason I prefer 12 V LEDs, as they don't have to deal with the problem of converting 240V AC to the low voltage and current needed by the LEDs themselves.

If anyone has any dead LEDs and they like to know what the probable failure cause is then I'd be happy to conduct a few post mortems on them and publish the results here. It might help understand what is likely to make them fail and maybe show whether or not it's heat build up in the base that's the problem.

Thanks Malcolm, email sent.

I shall endeavour to publish photos of the post mortem here.

Malcolm kindly donated a dead 2.2W GU10 LED so that I could have a go at doing a post mortem on it and find the likely cause of it failing.

The pictures below pretty much illustrate in order how I took the thing apart. The first shows the LED as it arrived, with clear signs that it had gone "pop" quite dramatically near the contacts.

The base (which contains a small switched mode AC-DC converter) was held on by two small cross head screws and came off easily. The red and the black wires go directly to the LED puck mounted in the reflector/heatsink, the AC-DC converter is in the plastic base.

I snipped the LED wires and twisted the base pins to break the crimp connections to the AC-DC converter module. This then pulled out fairly easily and the source of the black soot could be seen, a burnt out inductor on the 240V mains input connection (marked in red).

The close up shows that the inductor seems to have been shorted out against the circuit board and there is evidence that solder nearby has been melted. It seems possible that this component overheated, the solder melted and this then caused a short that blew the wires off. There's no indication that the other bits of wire on this inductor overheated, so it doesn't look like an overload failure.

Finally I connected the LED puck to a power supply (via the white dropper resistor in the photo) and confirmed that the LED itself still works OK.

This failure mode is very similar to the LED I took apart last year, so it seems that there is some limited evidence that it may be overheating of the AC-DC converter circuit that's the probable cause of GU10 failures. Finding fittings that allow good ventilation around the base might possibly improve reliability, by keeping the critical bit a little cooler (although this is only really guesswork).

I'll happily do post mortems on any other LEDs if it would be of interest.

Malcolm's raised a very good point in an email, that this unit (and others of the same type) failed as they turned on, so overheating seems unlikely to be the cause.

I've just taken a closer look at the point of failure and it may just be down to poor manufacturing. The small toroidal inductor is wound as an interference suppressor right on the 240 V AC input. One winding is the live, the other is the neutral, and the thing is intended to help prevent radio frequency electrical noise from the switched mode converter getting back into the mains and causing interference.

Looking closer, it seems that two very short bits of wire from this, one on the live side and one on the neutral side, have been vaporised, much like a fuse. I've drawn in (in red) on the photo below where the wires used to run. At their widest spacing, where they join the circuit board, they are just about 2mm apart. At their closest they are virtually touching. I suspect that the very thin enamel insulation on the wire may have been damaged during manufacture, possibly because it's a self-stripping polyurethane enamel that vaporises during soldering.

If the wires are damaged, then it's probably just luck as to whether they stay sufficiently far apart as to not touch and short. At turn on there would be an in-rush current pulse to the AC-DC converter input capacitor, and it may be that this just slightly moves the toroidal inductor, perhaps just enough to create a short. Once the input wires short the unit just fuses by vaporising the inductor wires, which probably blow at a lower current than the lighting circuit trip.

I have run the LED for a while on the bench (from a current limited DC source) and found it gets to around 62 deg C at the point where the reflector joins the plastic rear housing. This was in open air though, without the added heating that would come from the AC-DC converter. Even so, I don't think that it could get hot enough internally to melt solder, although it may well reach pretty high temperatures that might impact on long-term reliability.

Is the finned part of the case plated plastic or metal? Hard to tell from the photo.

glad to see you are back posting JSH

Jeremy
Would it be worth looking at an old one that has not blown, see if there is any difference in the circuitry?

This mostly confirms my opinion that GU10 mains LEDs are crummy engineering. 62C with the lid off doesn't sound good.

I've made a nice current-driven XPG-based LED lamp to test my ideas on how it should be done. Busy right now but will post some pics soon.