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

Thanks, Wallace

MichaelG.

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

Cheers, John

MichaelG.

Yes, it is too easy! You really need a series resistor as well; even more so if your voltage source isn't particularly well regulated. Let's consider the white LED I am planning to use for my traction engine lamps, as I have the figures in my head. The voltage drop for the typical operating current of 20mA is 3.4V, but worst case the voltage drop can be 4V. However, the maximum DC current is specified at 25mA, no far above the 20mA operating current. So you really need to be sure that the LED is operating at the current you think it is! Hence the series resistor. A series resistor along with a voltage source simulates a current source.

I didn't follow the comment by Neil on resistor values. If you're simulating a current source with a voltage source and series resistor then the larger the resistor the better the compliance of the simulated current source. So larger value resistors are better. Of course if you're using the resistor to simply sense the current and using that to control the voltage by external means then the resistor can be small. Usually limited by noise performance and the amplification needed. Common-mode can be a problem too, if the current sense resistor isn't referenced to 0V.

Regards,

Andrew

Thanks Andrew, seems I don't know as much about electricity as I thought, and that's not much. Simulating a current source with a voltage source ?. All new to me, and I got a national certificate in electrical eng. about 50 years ago! Anyway- I've just taken my 6 LED headlight to bits- it has 3 dry cells total voltage 3.5v ( cells a bit old ) can't get at the LEDs or electronics, if any. These must be wired in parallel to give 3v at each LED ? Think I'll give up and buy something ready made. Starting to rebuild battery drill pack, think I can manage that.

.

Gordon,

I'm re-posting this, in case you missed it

... I'm sure you will find it helpful.

MichaelG.

Thanks I did look at the site, may well use it later. Just had a head banging moment and realised that these things are diodes, and not simple filament lamps.

>I didn't follow the comment by Neil on resistor values.

because the Vf for the LEDs ensures the voltage to be dropped is quite small, you only need a fairly small resistor to limit the current, making the approach relative efficient.

You will often encounter a single, small white LED run off a stack of three button cells, or 20 or so in parallel run off three alkaline cells, no dropper resistor. Despite breaking the rules the internal resistance of the cells is sufficient to limit the current to (Vbat-Vf)/Rbat.

Well, I know what I mean, anyway!

Neil

Thanks for the update, Wallece

That lamp with the three 1W beads looks tidy.

I must profess to being a little concerned by the spikes on your 'scope trace ... Could you try a faster timebase? the detail might be interesting.

MichaelG.

.

 Let's hear it for these Guys 

Edited By Michael Gilligan on 07/10/2014 18:54:58

NewsFlash : eMail from Screwfix, this morning

I haven't tried any of their LEDs, but Screwfix' price, service, and returns policy, are generally good.

MichaelG.

Regarding the EU legislation on efficacy:

Here is a useful and reaonably intelligible leflet from Sylvania.

MichaelG.

Edited By Michael Gilligan on 13/10/2014 09:43:37

Just replaced 4 GU10 Halogens with 3W leds. Although 24months and 20,000 hrs warranty, after 15 secs., half leds (30 out of 60 )in one unit go out. Turn off and let the afterglow run down, switch on again and fully lit for 15 secs and drops out again. Rotated unit through 180deg. and same half drops out.

Regards Ian

Ian,

It would be helpful if you could "Name 'n' Shame"

MichaelG.

Hi Ian,

Unlike GU10s, worth taking back for a refund/exchange. Hope you have your receipt!

When I was a boy and helped out in my Dad's shop, every time I sold a bulb (any voltage or rating) I had to test it and show it was working on a big board he had made covered with sockets and rails etc.

I always assumed that was why they used to be in packing with the contacts ex[posed, so you could test them without taking them out of the packet!

Neil

THG GU10 3W DayWhjite 60 SMD 240V   Bought via T'internet, E-mail out waiting for reply, dispatched from Scotland.

Regards Ian.

Edited By Circlip on 14/10/2014 13:37:59

.

Thanks Ian

Hope the outcome is good.

MichaelG.

Here's an odd one ...

Plugged a little USB 'gooseneck' LED light into the Mac, just to check it.

Mac popped-up a warning message, telling me that the device required too much current, and was therefore being switched off ... Odd, because the USB port on the MacBook Pro is capable of delivering One Amp, and the single LED should only draw a few milliAmps.

Thankful that the Mac had protected itself ... I had a look inside the lamp:

Note that the insulating sleeve is just a little shorter than it needs to be.