LEDs ... The Good, The Bad, and The Ugly

It would be interesting to run a dismembered LED lamp from an isolation transformer and measure the input and output voltage and current using a DMM to get a rough idea of the efficiency. Sounds as if some of them may even work on a lowish voltage DC power supply. It doesn't seem to be a measurement that is commonly mentioned out there yet it's a fairly obvious one.

I'd almost have a go myself but I'm moving house and have packed all that gear up. I'm sitting expertly in the armchair here, metaphorically speaking.

Murray

.

Murray,

What you say is true, but [as I noted earlier]; incandescent lamps put a lot of their heat into radiated infra-red [as part of the broad-band light output], whereas the light output from an LED is narrow-band, and almost all the heat is conducted away via the base. ... My 10W LED on it's test heatsink produces a cold [i.e. physical temperature, not colour-temperature] light, and a very hot heatsink.

I'm still waiting for the thermometer, but my guestimate is 60° to 70°C after ten minutes at 10W ... whereas a 100W incandescent bulb would be nowhere near that; because it radiates most of its heat within the light.

Yes; the quantity of heat to be got rid of is much smaller, but it needs to be conducted away from the device.

All interesting challenges.

MichaelG.

.

P.S. When I finally get to grips with keeping the 10W chip cool, there is a 50W device waiting to go into a microscope lamphouse [no Fans permitted, it needs to be Passive cooling].

Edited By Michael Gilligan on 05/10/2014 22:18:38

Hi Michael

Can you use liquid cooling like I believe the use on some high power PC CPU's?

I think I read somewhere pulsing the supply (PWM) is not an option for you.

HTH

Cheers Michael.

.

Yes, that's a possibility that I'm considering, Michael

Trouble is: Most of the PC CPU Liquid Coolers seem to use pumps [which, like the fans, I'm keen to avoid] ... If I recall correctly you need pretty large diameter pipes before natural convection works properly in a liquid circulating system.

Definitely worth having a go some-time though!

... after that it's probably "Heat Pipes" and/or Peltier Elements.

This is all getting rather more exotic than I had hoped; but it's still easier and safer than running a Carbon Arc Lamp, or a [continuous] Xenon tube.

MichaelG.

Edited By Michael Gilligan on 06/10/2014 00:23:48

This morning's find; a company that seems to know about passive cooling of LEDs.

MichaelG

Michael, this is where you need to be at for your LED info: Candlepowerforums

The guys on there seem to have tried it all. There are some really interesting threads comparing the different types of LED and there are useful tips on how to get rid of the heat. Living in the sticks I have a couple of home made flashlights, one of which is exceptionally bright despite it's very diminutive size.

graham.

Thanks Graham ... Yes, been there.

At the time; most of the threads I saw seemed to be about "flashlights"

... I will re-visit.

MichaelG.

.

Edit: I've just tried a search on the word 'microscope', with rather meagre results

... Will try a few other search terms.

Edited By Michael Gilligan on 06/10/2014 09:16:43

Well, yes it is Michael, but it is all the same principle and the knowledge is interchangeable really. They do seem to have a handle on which LED manufacturer's product to buy and which to avoid. Some of the runtimes of these run in to weeks, not hours, at a low setting. Absolutely invaluable in certain situations.

graham.

.

Fair comment, Graham ... I didn't mean to sound negative.

MichaelG.

No, I know that Michael, the comment was more for general consumption as I find a good many people find themselves bewildered by all the various choices options and permutations. I am an "early adopter" of the technology and I have made my own fair share of mistakes when choosing different bulbs. I think the biggest single mistake most people make is to get LED's that are far too blue. In other words of a high Kelvin rating and this puts them right off as it is not a comfortable frequency to work by.

My good lady for instance won't buy LED's because she went down the LED road early on and the lights were all "too blue" so I am surreptitiously going around the house swapping the bulbs out, slowly she is coming around but when swapping out the 12v halogen bulbs I have hit snags, especially with wire wound transformers and even more so with lighting dimmer circuits. And I have a bit of a clue as to what I am doing so average Joe will find themselves going back and forth in some instances.

graham.

What makes LED's dimmable? Do the ones with 60 leds and no SMPSU work off a simple triac dimmer?

Neil

I've been reading all the info' about LEDs and now know even less, not your fault ,it's mine,I have a number of LED devices, chargeable torches and work lights and battery powered torches and headlamps, all are very good. What I want to know is how to buy and power new systems, eg work lights. eg.- what voltage does a single LED run at, or there different types ? Are the many different devices wired with LEDs in series and/or in parallel to suit requirements? What sort of voltage variation can be used ? This is very simple stuff for most of you, I know , but all the different acronyms mean nothing to me. Any help welcome.

It's all in the name; LEDs are diodes, so the voltage/current curve is exponential, exactly the same as ordinary silicon diodes. Ideally an LED should be driven by a current source, as small changes in voltage can lead to large, and ill-defined, changes in current. The quoted voltage drops of an LED for a given current depends on the work function of the semiconductors and dopants used in manufacture, but are typically in the range 2-4V versus the 0.6-0.7V normally quoted for a silicon diode.

While LEDs can be run in parallel or series, parallel is not normally used, as equal current sharing cannot be maintained. Consequently LED brightness will vary from LED to LED. All the LCD backlighting applications I have designed for use LEDs in series. If there is more than one series string then they are driven from separate supplies.

Best Regards,

Andrew

.

Neil,

I believe that the "Consumer" classification is effectively that "dimmable" LED lamps do not have Constant Current drivers, whereas "non-dimmable" ones do.

Therefore: Yes, I think "dimmable" means it will dim with the triac dimmer.

MichaelG.

Thanks Andrew, So if I buy 6 LEDs I wire them in series and run at 12Volts? Seems too easy.

.

Gordon,

Be aware that different colours of LED have different forward voltages.

White ones are more typically about 3.2v

MichaelG.

.

P.S.   .... This site has useful calculators and info.

Edited By Michael Gilligan on 06/10/2014 12:14:14

.

Very, Very, True ... but perhaps I should explain my interest in the 50W Lamp.

I have a couple of microscope lamphouses that use this type of Xenon lamp ... These lamps are expensive, fragile, and run at lethal voltages; so I am looking to devise a safe and serviceable substitute.

MichaelG.

Hi Micheal

have a look at this MR11 lamp

breaks the rule of having LEDs in parallel with a common dropper resistor

(more pictures in the album)

the GU10 mains lamp has 20 LED's in series

C1   the 0.68UF capacitor acts as a Wattless dropper instead of a resistor

(the resistor  R1 provides a discharge path for C1 )

this DIY 4" tube replacement could give you ideas

I like the fact it runs of 12V , usefull when the mains is lost

John

Edited By john swift 1 on 06/10/2014 14:43:24

It's worth observing how small the current limiting resistors can be and still be effective.

Neil

.

John,

Thanks for the schematics, but I'm not sure what you mean by "breaks the rule of having LEDs in parallel with a common dropper resistor" ... It's certainly not a good rule. [see link above]

There do seem to be many variants on the theme ... which is why I thought it worth starting this thread.

... Keep 'em coming !!

MichaelG.

Edited By Michael Gilligan on 06/10/2014 14:59:59