🚀 go-pugleaf

On Fri, 4 Oct 2002 09:14:21 +0100, Keith Wootten
<keith.w@ntlworld.com> wrote:

>In message <3d9d452f.4128065@news.xs4all.nl>, "Wouter van Ooijen 
>(www.voti.nl)" <wouter@voti.nl> writes
>>>Have assembled an array of LEDs for use as a 3rd brake light in my car,
>>>configured as 4 rows of 9 jumbo LEDs, as of yet unwired. Of course the power
>>>supply will be 12 volts nominal (13.8 volts, maximum).
>>>
>>>What is the best method to power these from an intermittent (whenever I press
>>>the brake pedal) power source?
>>
>>The crucial point is that when you put a resistor and N LEDs in
>>series, the LEDs will take roughly a constant voltage so all voltage
>>variation from your battery will be over the resistor, thus causing
>>current variation.
>
><Snip>
>
>You can mitigate this to a useful extent by using a small filament bulb 
>instead of a resistor.  The positive temperature coefficient of the bulb 
>will help to keep the current variation reasonable as the supply voltage 
>changes.  The bulb should be selected to run at more or less normal 
>brightness at maximum supply voltage while passing maximum LED current.
>
>
>My bicycle rear light is an array of 50 LEDs from RS components which 
>are arranged in a series/parallel configuration with a nominal Vf of 9V5 
>IIRC.  Using two small paralleled 5V lamps (can't remember the type - 
>selected from a junk box) the current swings 2:1 for a 14V-10V supply.
>
>In my case, it's also a useful low battery warning - the filament lamp 
>goes out long before the LEDs.

In the auto situation, the simplest mitigation is to feed the
resistor/LED chain thru an LM317 configured as a constant current
source.  Just the IC and one resistor added, and no voltage variation
issue.