Re: Connecting strings of LEDs in parallel

From: Ben Bradley (ben_nospam_bradley_at_mindspring.com)
Date: 08/05/04 

Date: Thu, 05 Aug 2004 10:10:24 -0400

On Wed, 04 Aug 2004 16:10:32 GMT, Mike Harrison <mike@whitewing.co.uk>
wrote:

>On Wed, 4 Aug 2004 23:30:11 +1000, "Jason Burton-Woods" <jasonbw@[eyeeye]net.net.au> wrote:
>
>>Hi all,
>>
>> There's a project going on the www.ducatimonster.org forum about how to
>>convert an incandesant tail light, into a bright LED light... but we could
>>really do with some help from the experts?
>>
>>The guys are standardizing on the following:
>>
>>Using a LM7812 to regulate the variable voltage from the bike down to a
>>constant 12v, make 8 strings of 3 LED's and connect those strings in
>>parallel... if you cant follow me then have a look at this schematic..
>>http://www.ducatimonster.org/ImageFolio3_files/gallery/Misc/LED_tail_light_circuit_diagram.JPG
>>
>>There are two resistors used, on 150 ohm connected to the Brake light +,
>>the other 150 ohm to the running light, these are in parallel so when the
>>brake switch is not in operation then the total resistance is 1/2 that
>>compared to when the brake light switch is operated.
>>
>>So, heres the question...
>>
>>>From a number of difference web sites we've been looking at, we've found
>>people saying its important to wire an individual resistor into each
>>string.... and of course as the web would have it, we came up with a few
>>that said its not necessary... Think you could clear this up for us??
>>
>>Thanks all in advance.
>>
>>P.s. Link to the thread is here:
>>http://www.ducatimonster.org/cgi-bin/yabb/YaBB.cgi?board=parts;action=display;num=1090891230;start=15#27
>
>Using a seperate resistor in each string helps ensure the same current flows through each string
>regardless of small variations in LED characteristics due to manufacturing tolerance, aging and
>temperature. In practice, if the LEDs are all the same type, and not sublect to substantially
>different temperatures between strings, you can probably get away with paralleling the strings.

   Has anyone ever got 100 or 1,000 new LED's and measured the voltage
across each with a fixed current through it? I wonder what to expect
for variations.
   I'd go for overall high efficiency, put all the LED's in a big
series string and use a switching regulator set up to put a constant
current through it, though this isn't fault tolerant for a LED opening
up (do LED's ever fail that way?). Since this is "safety critical"
application, I might go for two or three regulators, each powering 1/2
or 1/3 of the LED's.

>However resistors cost almost nothing, and having seperate ones per string improves fault tolerance
>- any single component failure will at worse lose 1/3 of the LEDs. With paralleled strings, a LED
>failing to lower voltage will decrease brightness of the other 2 strings.
>
>Using seperate resistors also spreads out the heat dissipation over a wider area.

  A series current-limiting resistor for a LED or even for the current
through three shouldn't cause significant heat (unless these are tiny
0804 SMT resistors).
   Of course, a higher resistance wastes more power. To know how low a
series resistor can be, you need to know the possible variations in
voltage across different LED's.

>Therefore there is rarely a good argument for not using seperate resistors per string. COst is the
>only one, and compared to the cost of the LEDs, it's insignificant.

-----
http://mindspring.com/~benbradley

Loading