Cor van de Water wrote:
>I do not know if you actually measured a white LED recently,
>they have a very large internal resistance. The regulation
>is not as bad as you would like us to believe.
>Still, because I am a tinkerer, I built my own regulator.
>Since LEDs are current-controlled devices, I built a
>It is pretty simple to create a current source that drops
>less than one Volt and is pretty stable in its controlled
>characteristic - a resistor to measure the voltage drop
>caused by the current and you have a pretty solid value
>that only changes a few percent when ambient temp changes.
>(To be precise: 1% for every 3 deg Celcius change)
>If you want to make it perfect (temp compensated) it costs
>you more than two diodes and two resistors plus a transistor.
>Such a current regulator will keep the brightness of the LEDs
>constant and you can easily use 3 white LEDs in series and
>have no impact from fluctuating voltages, so your LED will
>not burn out prematurely or dim unnecessary for all reasonable
>voltage levels (say, 11 to 15+V)
>See the crude drawing below (watch with fixed font)
>The component between the e-b-c is a NPN transistor
>with at least Hfe = 100, preferably higher.
>The good old BC-547 will do perfectly.
>R is the current sense resistor and will drop
>approx 0.7V, dependent on the type of diode D used.
>If you use two schottky diodes, the total drop across the
>diodes is less than 1V, minus 0.7V for the b-e drop, so
>the resistor will then get less than 0.3V
>The R-bias is just to get some current through the diodes
>and open the base of the transistor. Around 33k Ohm should
>be enough to get everything to work, assuming the LEDs
>do not need more than 50 mA (I actually use two parallel
>strings of 3 LEDs in series to regulate 6 LEDs in my
>For the indicated 0.7V and 20 mA LED current, the resistor
>R needs to be: 0.7 / 0.020 = 35 Ohm. (33 Ohm nearest value)
>NOTE that the efficiency and the voltage drop for this
>regulator is determined by the drop across the R and the
>e-c drop of the transistor.
>Since many transistors are happy to go down to 0.2V, the
>total drop of this regulator can easily be less than 1V
>and with two schottky diodes, it would even be around 0.5V.
> R ____ e ___ c 3 white LEDs
>| 0.7V b| |
>| D D | ____ R-bias |
>|__o negative 12V supply positive o
>Cor van de Water
>Proxim Wireless Corporation http://www.proxim.com
>Email: [email protected]
>Skype: cor_van_de_water IM: [email protected]
>Tel: +1 408 542 5225 VoIP: +31 20 3987567 FWD# 25925
>Fax: +1 408 731 3675 eFAX: +31-87-784-1130
>Second Life: www.secondlife.com/?u=3b42cb3f4ae249319edb487991c30acb
>From: [email protected]
] On Behalf Of Danny Miller
>Sent: Thursday, October 04, 2007 12:42 PM
>To: Lee Hart; Electric Vehicle Discussion List
>Subject: Re: [EVDL] LED driver that runs off of 12V
>Because your solution does not have adequate regulation.
>At 14.6V, the LEDs will take over 3x the current. At 11v current drops to 1/6th. A regulated voltage source helps, but most regs drop quite a bit of the voltage, and there is still a regulation issue due to the thermal coefficient of forward voltage in the device as well as the
>device's inconsistent Vf. In practice a 3.6V LED may actually be
>2.9V-3.8V (depends on spec sheet), so even with a regulated 12V the current could anywhere from 20ma to 110mA so we don't know if it will be bright enough to be seen or if it will burn itself out from overcurrent. If you only use 2 3.6v LEDs in series and increase the ballast resistance then it becomes much better regulated, but less efficient.
>Cost is not as big as it seems for mass producers. Ever take apart the base of a $3 110V CFL tube? Holy crap there's a lot in there. They kept the cost down (well in part by not using ASICs like this LED drivers, but still).
>ZXLD1360 is a good, simple LED regulator but it's only 36V IIRC.
>There's gonna be a ZXLD1362 out eventually with a much higher voltage rating.
>HV9910 is a great reg, high voltage rating, but you need to buy an external switching MOSFET with it so it's a bit more complicated and expensive.
>Lee Hart wrote:
>>From: Rod Hower
>>>Thought somebody might want to roll their own,,,
>>It always amazes me how complicated they can make driving an LED.
>>This circuit uses dozens of parts and ICs to drive some 3.6v LEDs from a 12v battery. A plain old resistor for each three LEDs in series also does the job nicely. The resistor will drop 12v - (3 x 3.6v) = 1.2v which means the overall circuit is 90% efficient -- higher than this switching regulator circuit. It automatically draws no current if the battery goes dead because the LEDs stop conducting.
>>Yes, the light output varies with battery voltage; but how often is that a problem. Car tail lights also vary as the car's electrical system wanders around 12-14v. An ordinary 3v flashlight bulb could be used as the series resistor as a crude constant current source, for a 1-component solution.
>>"Excellence does not require perfection." -- Henry James
>>Lee A. Hart, 814 8th Ave N, Sartell MN 56377, leeahart-at-earthlink.net
>>For subscription options, see
>For subscription options, see
>For subscription options, see
For subscription options, see