CREE Design Circuit 12Volt Power Supply Diagram

DrewE wrote:

theoldwizard1 wrote:
FYI - your title says "12V power supply". With 12V in you are probably only get a maximum of about 11V out, probably less depending on the transistor used.

With three white LEDs being powered, output voltage would be in the vicinity of 9-10V; their forward voltage is generally a bit over 3V apiece. This circuit should be designed to act as a constant current supply, rather than a constant voltage supply, for this application. I can only assume that the circuit they suggest would indeed behave this way, though I haven't made any effort to verify that.

With a 10V output, efficiency would be around maybe 75%, with a 25% loss to the linear regulator. Switching regulators do better, of course, but it's not like this is impossibly inefficient.

As for suppliers for the components, any reputable electronics supply place should give you good, non-counterfeit parts, or at least be as likely to as you can reasonably get--DigiKey, Mouser, etc. Avoid ordering directly from random unknown Asian companies/places. LCSC may be okay; they're Chinese, but not so much random and unknown.

you hit the nail on the head on the power supply. LED's should be driven by a current source NOT a voltage source. As they heat up junction voltage drops, and if driven by a voltage source, current goes up, heat goes up junction voltage goes down.... and your on to thermal runaway.

theoldwizard1 wrote:
FYI - your title says "12V power supply". With 12V in you are probably only get a maximum of about 11V out, probably less depending on the transistor used.

With three white LEDs being powered, output voltage would be in the vicinity of 9-10V; their forward voltage is generally a bit over 3V apiece. This circuit should be designed to act as a constant current supply, rather than a constant voltage supply, for this application. I can only assume that the circuit they suggest would indeed behave this way, though I haven't made any effort to verify that.

With a 10V output, efficiency would be around maybe 75%, with a 25% loss to the linear regulator. Switching regulators do better, of course, but it's not like this is impossibly inefficient.

As for suppliers for the components, any reputable electronics supply place should give you good, non-counterfeit parts, or at least be as likely to as you can reasonably get--DigiKey, Mouser, etc. Avoid ordering directly from random unknown Asian companies/places. LCSC may be okay; they're Chinese, but not so much random and unknown.

Gdetrailer wrote:

Although there are switching regulator chips you can buy, not really recommended for the amateur DIY person since in the wrong hands without proper circuit design you can end up with a very bad RFI transmitter..

I think you might have described what i did!!! Total amateur without a clue. I made some LEDs designed for RV use and was getting weird noise at times. Wasnt smart enough to know the problem. I had asked on some electronic forums. Ended up buying ferrite rings in hopes for a cure. But again i didnt have a clue what i was doing. The lights work great but at certain dim levels i can here a noise. My girlfriend swears she cant hear it but i do. The fix was easy enough to get by. I run them full blast or very dimmed. Been a good 3 or 4 years i dont even think about them anymore. I did use CREE leds though. Still have boxes of material and completed lights laying around.

FYI - your title says "12V power supply". With 12V in you are probably only get a maximum of about 11V out, probably less depending on the transistor used.

MEXICOWANDERER wrote:
Yes to all
This is the reason why the power supplies have been failing on you. As I stated earlier I only use Meanwell power supplies. DC to DC and AC to DC. They use proprietary ICs. Sort of a benevolent trick GDEtrailer but this is why I use my more expensive option. Avoid 120vac CREE devices. Single unit 120vac current drivers are flawless when double checked for current (thermal) combatibility. Did I forget to mention the discrete device above is Crees own design?

Cree didn't "design" that circuit..

That is taken directly from TIs own circuit samples, sure they may have "tweaked" the values but "designed", no..

Giving Cree a bit more credit than they deserve.

Most manufacturers instead of designing simple things from scratch they use things like that regulator and a sample diagram to build on, saves time and R$D money..

Although since for yrs Cree like any other LED manufacturer have been in the efficiency wars would use higher efficiency switching power supply regulator chips instead of the energy burning fire breathing and wheezing Linear regulators..

Don't get me wrong, I HAVE been known to break out a few Linear regulator circuits when time and money or noise are at stake..

Example, to save on RFI like my homemade 30A 13.8V power supply for Ham equipment where RFI IS a major consideration. Big heavy honking transformer, heavy duty bridge rectifier, regulator designed with 12V 1A linear regulator with several diodes in series lifting the ground on the chip and the chip is driving 4 paralleled 2N3055 in series pass regular mode.. The output in theory could handle 60A but the transformer is only good for 30A.. Cjeap, effective, Built proof and stone dead quiet in the RFI department but at a cost of generating a bunch of wasted heat in idle mode..

Yes to all
This is the reason why the power supplies have been failing on you. As I stated earlier I only use Meanwell power supplies. DC to DC and AC to DC. They use proprietary ICs. Sort of a benevolent trick GDEtrailer but this is why I use my more expensive option. Avoid 120vac CREE devices. Single unit 120vac current drivers are flawless when double checked for current (thermal) combatibility. Did I forget to mention the discrete device above is Crees own design?

Mex,

You do realize that a LM338 is a "linear" regulator?

PER HERE

Linear regulators waste a lot of power in the form of heat, pretty much dissipating the same amount of heat as a plain old resistor.. In some respects using a resistor in series with the LEDs will be more energy efficient than a linear regulator..

LM338 is not a more efficient switching regulator.

That thing will eat your lunch and ask for more, pretty much negating the energy savings of a LED..

Although there are switching regulator chips you can buy, not really recommended for the amateur DIY person since in the wrong hands without proper circuit design you can end up with a very bad RFI transmitter..

There are plenty of variable output switching regulators prebuilt for reasonable prices to bother with building a linear regulator from scratch now days.