LEDs … The Good, The Bad, and The Ugly

[MuzzerParticipant @muzzer]

There are 2 power ratings being discussed. One is the bare LED component itself and the other is the complete lamp assembly that runs off 230Vac. Michael's lamp that is marked 9W probably only has a few W of actual LED power and the rest is distributed as losses around the driver gubbins inside the housing. I wonder what the conversion efficiency is from AC input to LED connections? I suspect pretty poor.

Inside a modern car's rear LED cluster, the LEDs are usually mounted on IMS to act as a heat spreader and heatsink. There is heat to be got rid of but it's a lot less than the original incandescent bulbs they replace.

You'll notice that some (most?) LED car lights are pulsed, giving a strobe-like effect with movement.

Murray

[MuzzerParticipant @muzzer]

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

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Muzzer on 05/10/2014 21:34:08:

There is heat to be got rid of but it's a lot less than the original incandescent bulbs they replace.

.

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

[Michael HornerParticipant @michaelhorner54327]

.

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].

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.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Michael Horner on 05/10/2014 23:49:21:

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

.

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

[Michael GilliganParticipant @michaelgilligan61133]

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

MichaelG

[Oompa LumpaParticipant @oompalumpa34302]

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.

[Michael GilliganParticipant @michaelgilligan61133]

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

[Oompa LumpaParticipant @oompalumpa34302]

Posted by Michael Gilligan on 06/10/2014 09:09:19:

Thanks Graham … Yes, been there.

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

… I will re-visit.

MichaelG.

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.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Oompa Lumpa on 06/10/2014 09:24:17:

… but it is all the same principle and the knowledge is interchangeable really.

.

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

MichaelG.

[Oompa LumpaParticipant @oompalumpa34302]

Posted by Michael Gilligan on 06/10/2014 09:30:24:

Posted by Oompa Lumpa on 06/10/2014 09:24:17:

… but it is all the same principle and the knowledge is interchangeable really.

.

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.

[Neil WyattModerator @neilwyatt]

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

Neil

[Gordon WParticipant @gordonw]

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.

[Anonymous]

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

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Neil Wyatt on 06/10/2014 10:24:24:

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

.

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.

[Gordon WParticipant @gordonw]

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

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Gordon W on 06/10/2014 12:05:54:

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

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Oompa Lumpa on 06/10/2014 09:24:17:

Some of the runtimes of these run in to weeks, not hours, at a low setting. Absolutely invaluable in certain situations.

.

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.

[john swift 1Participant @johnswift1]

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

[Neil WyattModerator @neilwyatt]

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

Neil

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Michael Gilligan on 24/09/2014 13:05:49:

If you are interested in failure modes of LED clusters … this is worth a look.

.

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

[WALLACEParticipant @wallace]

I just put a scope across the output of the pcb in a white 61 led MR16 bulb….

There is ripple or spikes on the output – but it’s very little – just a couple of millivolts. The oscillator works at around 180kHz..

Annoyingly, I can’t do any current measurements as I left my meter at home….so I’ll probably leave running it on ac until I have it available.

Pics will go in an album when I get better internet connectivity.

W.

[Michael GilliganParticipant @michaelgilligan61133]

Thanks, Wallace

MichaelG.

[john swift 1Participant @johnswift1]

Hi Michael

typing finger and brain not in sync

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

should of been – breaks the rules having LEDs in parallel with a common dropper resistor"

 

the forward volt drop across LED's will not be identical

the LED's with the lower Vf will take a larger share of the current , run hotter

resulting in the Vf dropping and the LED current increasing

this runaway condition will destroy the LED

 

the regulator in the     http://www.instructables.com/id/LED-Replacement-Fluorescent-Tube/     project

monitors the current through the 3 LED's and adjusts the supply voltage accordingly to keep the current constant

simple but has a problem if one of the first 3 LED's fail

the circuit on page two of the ON semiconductor application sheet    http://www.onsemi.com/pub_link/Collateral/AND8109-D.PDF     looks like solution

John

 

 

 

 

 

 

Edited By john swift 1 on 06/10/2014 15:58:00

[Michael GilliganParticipant @michaelgilligan61133]

Cheers, John

MichaelG.

[Author]

Posts