what is the LED efficiency and costs to run?

[d dunthorne]

Hi Phil,

Nice to get an input from you, particularly on a subject that’s dear to my heart. Before I put in my two pennyworth, just a small personal point – Sorry boss man, this is nothing to do with signs. You were kind enough to once recommend a marvellous book by Daniel Dennett – Darwins Dangerous Idea. Returning the favour, have you read his latest ‘Freedom Evolves’, it’s mind blowing.

Returning to your values of lumens and efficacy, Philips said in 2000 that they expected white LEDs to reach 50 lumens/w by 2005 so I guess that that 25 lumens/w now is about on course. I note these values from Lumileds latest samples.

However, practical LED strips for sign illumination do not seem to have reached anything like these values. I have been carrying out some tests on sign illumination on behalf of the BSGA and the results are not particularly encouraging. Tests carried out this spring are not that much better than others carried out in January 2002.

Tests have been carried out using a straight section of sign letter with a 3mm Opal 050 face. The surface luminance has been measured using a Hagner photometer. 300 mm strips of red and white LEDs from two sources were placed in an optimum position along the centre line of the test rig. The luminance obtained was measured and compared with that obtained from 15 mm clear neon and cool white (halophosphate) tubing operating at 50 mA.

The input power (excluding gear losses) was mesuared in all cases and standardised for a surface luminance of 100Cd/m2.

The results were as follows:

White cool white tubing = 740 Cd/m2; 0,62 W for 100 Cd/m2.
Red clear neon tubing = 250 Cd/m2; 2.7 W for 100 Cd/m2

LED supplier 1 (January 2002)

White LED strips = 69 Cd/m2; 2.9 W for 100 Cd/m2
Red LED strips = 60 Cd/m2; 2.3 W for 100 Cd/m2

LED supplier 2 (March 2003)
White LED strips = 108 Cd/m2; 2.6 W for 100 Cd/m2
Red LED strips = 64 Cd/m2; 1.8 W for 100 Cd/m2

From these tests you can see that the efficacy of the white LEDs is only about 25% of that of typical white cold-cathode tubes. Bear in mind that these results are for halophosphate coatings. Those for tri-phosphor coatings will be better. Also, have you seen the Vodaphone sign in Haymarket, which uses white LEDs. These have become duller and have lost their cold whiteness in a very short time. I have just completed some further calculations on cool white tubing. You would need 6 strips of white LEDs to equal the light output of 11 mm tubing operating at 35 mA.

As you rightly say, the position for red LEDs is much better. They are generally a bit more efficient than clear neon tubing, although as the above results show, their brightness is still lower than that of even basic neon tubes.

I remain very optimistic concerning LEDs for sign illumination and no doubt better devices are with us now. However, there is still over selling of their performance by some suppliers and sign makers should be wary of false claims. One of the reasons why I am carrying out this work for the BSGA is to try and throw some light onto a quite obscure subject – brightness being expressed in a multitude of different units is just one of the problems were are up against. Another is to make sense of the various optical distributions of different devices.

If you have details of other suitable configurations – strips seem to be the best for sign work – I should be grateful to hear from you. Perhaps you could email me direct. I will keep members of this sign board up-to-date with my findings.

Best of luck. John tells me that you are working far too hard.

Regards

Don Dunthorne

[Robert Lambie]

hi phill & don..
thanks to both you guys for starting to post around here. as much as i know lots of what you say maybe over some of our heads with regards to lighting etc. i still feel that the boards can all only benifit from your posts.

i look forward to any more posts by you guys and do please keep them coming.
is there such a website that the uk boards can link to that will give an idiots guide to LEDS, Neon and the like.
im not an electrical man myself but do dabble in this side of the trade from time to time. it would be good to have somthing to refer to every once and a while.

thanks again…

oh yeh.. any of you guys got a picture? 😉 if so, feel free to send me and ill load it for you.. thanks 😀

[Phil Rimmer]

Don,

I’ve just ordered my copy of the Dennett. Many thanks for that.

As we agree, the big problem is white LEDs. The reason for me showing the lumens per dollar figure is to show the shocking cost of white light from LEDs. This high cost lies at the root of why white LEDs are proving to be so much less reliable than advertised. Put simply, users are pushing them too hard. By cranking the current up to get four lumens per dollar instead of three, the chip temperature goes up.

As you know white LEDs are based on blue LEDs covered with a broad spectrum yellowish phosphor to render a sort-of-white. These devices have a considerably lower permitted junction temperature than red, yellow and some green LEDs. Their colour balance is a precarious thing at best, but sustained high junction temperatures soon pushes them out of kilter. (Discolouration of the packaging is also implicated here.) Along with the colour shift comes the inevitable loss of lumens.

Its not so much that the LED manufacturer data sheets are dishonest, its just that the claimed (useful) life and the maximum power rating are not to be had at the same time. Signs can experience considerable internal ambient temperatures due to solar heating, as you have pointed out on many occasions. I suspect that many are not taking this into account.

Behind the scenes the LED vendors are working hard on the problem. Lumileds have been one of the earliest to see that this heat / rating problem is crucial to making white LEDs work. As yet the new thermally efficient packaging is even more expensive…..but one day….one day….

In the meantime I believe the onus is on the LED system vendors to demonstrate that they understand the problem. They must demonstrate depreciation curves for lumen output and colour balance at the rated running current and at realistic sign box temperatures.

Rob,

Again, thanks for this useful forum. Alas, I know of no good neon websites with real knowledge. I suspect the very best we could do is have some kind of web server permanently plugged into Don’s brain. As you may know Don writes a lot of the technical notes for the BSGA. This great resource is available online to members.

A good resource for all the new stuff is a site called Don’s Lighting Page (I think). Lots of opinion, most of it well informed. I’ll post a link later todayish.

[Phil Rimmer]

Rob,

I just realised “Don’s Lighting page” looks like it might be the same Don as our Don. Not so, alas. If it were it would contain only verified fact. So use with a little care…

http://misty.com/people/don/light.html

[Robert Lambie]

excellent mate.. thanks for that ill have a look! 😉

[Phil Rimmer]

Just reading my latest copy of Opto & Laser Europe (I get it for the crossword). An interview with the VP of Research and Development at Lumileds, Frank Steranka. The news looks pretty gloomy for white. He claims current costs are 10 lumens per dollar and that if the devices can get into volume, the new technologies may in a few years time get to 50 to 100 lumens per dollar.

The efficiency claims made for the mooted new technologies seem much less gung-ho than a year ago. As to white LEDs displacing fluorescent lamps he says, “I can’t say that I see a clear path for that ever happening.”

Oo-er…

[t-winks]

we have an enquiry for a 2 totems and they are asking about led illumination run with solar panels, is this possible and if so do you know where we can get information for this

[Phil Rimmer]

The big question is why do they want LEDs?

If the answer is “for energy efficiency reasons”, the next question is what colour is the illumination to be? (Well, non white may just work with some forms of artwork…)

If the answer is white, remember the size of the solar panel will then need to be four times larger than if you used fluorescent lamps, likewise the batteries. Remember also that the solar panels and batteries are by far the most expensive items on your shopping list.

(Efficient battery powered ballasts exist for fluorescent lamps e.g. emergency lighting ballasts, but watch the output current rating.)

If the answer is red, orange, yellow or green then LEDs will do well or very well.

If they want white with good lumen maintenance then again fluorescent will win, (for 15,000 hours at least). Remember that a solar powered sign is generally turned on for less time than a normal sign due to the up front cost of the electricity supply. (So 15,000 hours goes quite a long way.) As the sign will need occasional cleaning, (especially the solar panel!) lamp replacement is not such a big deal and you get your full lumen package back.

If they want white light with good lumen maintenance at low temperatures remember that Aura make low temperature T8 lamps and others can supply a full range of special, sign T5 lamps that have no cold spot and can operate at temperatures 12 degrees celsius below normal tubes.

So, what do they need?

[d dunthorne]

Couldn’t agree more with Phil’s comments on white LEDs – they really are pathetic at the moment. They are certainly nowhere near the 50 lumens per watt that Philips suggested was likely before 2005. Even then, they will still be quite a bit less efficient than standard fluorescent tubes.

In my tests, I was hoping that the poor luminous efficacy would be somewhat compensated by the fact that LED optics can be arranged to put most of the light on the front face. However, this still did not make up for the low lumen output. Typically about 7 LED strips are needed alongside each other to match the light from one line of 15 mm cool white tubing operating at 50 mA.

If you’re in the Piccadilly Circus area (my happy stomping ground!) take a look at the white!!! LEDs on the Vodaphone sign. Six months ago they started life as a nice cool white colour. A few months later they were a distinct shade of pink. Now they are back tpowards the blue – pathetic.

I would agree with Phil on the superior performance of coloured LEDs. I am currently engaged in specifying new LED matrix displays for one of the Piccadilly signs. These use red, green and blue LEDs in each pixel and we are asking for a minimum deterioration of light output and colour over an 8-year period. With careful control of the operating conditions and LED junction temperature, suppliers can achieve this.

Don Dunthorne

[t-winks]

thanks for the advice the problem was more to do with the room available for the lighting. It was going to have to be mounted between a double elipse, changed the design slightly, now have the room to look at fluorescent.

Tony