[RE-wrenches] DC power for LED lighting

Dan Fink buckville at hughes.net
Sat Dec 17 08:11:38 PST 2011


From: Dan Fink <danbob88 at gmail.com>
To: Hugh Piggott <hugh at scoraigwind.co.uk>
Cc: danbob at otherpower.com, RE-wrenches <re-wrenches at lists.re-wrenches.org>
Date: Sat, 17 Dec 2011 08:59:58 -0700
Subject: Re: [RE-wrenches] DC power for LED lighting
Hullo Hugh;

Perhaps I could state it better,  that each string of series LEDs needs 
to have its own current limiting device (a resistor or regulator).

The reason I advise that is that all LEDs vary a bit in their 
volt/amp/lumen performance curves, especially between batches, but even 
within a batch.  If you were to put 2 "identical" LEDs in parallel 
sharing one current limiting resistor, one LED is *always* going to draw 
more current than the other and be brighter. If the difference is too 
great, the one drawing more current will burn out first. If you are 
lucky with your matching, it might not be problem for a very long time, 
but I've also seen burnout happen between batches in just a few minutes.

It's a positive feedback loop (thermal runaway). As an LED heats up, 
forward voltage drops and current increases.

So, by current limiting each string you avoid both the variable 
brightness problem and the thermal runaway problem.

This is a pretty good article on the topic:
http://www.ledsmagazine.com/features/4/8/1 
<http://www.ledsmagazine.com/features/4/8/1>

It's nice that we can all purchase decent LED lighting products these 
days and not muck about with breadboards and soldering irons like we had 
to a few years ago. A good inexpensive compromise is to regulate all the 
strings with a single DC power supply, then current limit each string 
with a resistor.

-- 
Dan Fink,
Executive Director;
Otherpower
Buckville Energy Consulting
Buckville Publications LLC
NABCEP / IREC accredited Continuing Education Providers
970.672.4342 <tel:970.672.4342>(voicemail)


On Fri, Dec 16, 2011 at 11:03 PM, Hugh Piggott<hugh at scoraigwind.co.uk 
<mailto:hugh at scoraigwind.co.uk>>wrote:

    H Dan,

    On 17 Dec 2011, at 00:02, Dan Fink wrote:

     >  parallel connections are BAD with LEDs, and the string with the
    most voltage will eventually fail first, so best practice is
    regulate every string.

    This statement puzzles me so I wonder if you could clarify it for
    me?  If you connect two strings of LEDs (or batteries) together in
    parallel then their voltage will be the same.  So how come you are
    talking about one having more voltage?

    If the answer is that it's absolutely impossible to make the
    connections and leads have the same resistance then this same logic
    applies to two strings that are connected each to its own regulated
    supply.  If a tiny difference in voltage matters (which I doubt)
    then you will have the same issues with a single string.

    What am I missing?

    thanks

    Hugh


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org/attachments/20111217/7e31e2d2/attachment-0004.html>


More information about the RE-wrenches mailing list