# Cannot find efficient current/voltage regulator to power LEDs

I have a 12V 4-pin common-anode RGB PWM'd signal being used to power some LED strips, and I would like to use it to power the VL-H01RGB00302 Vollong 3W RGB High Power LED - which is basically a set of 3 LEDs:

forward current and voltage:
Red:          400mA, 2.5V
Green, Blue:  350mA, 3.4V


Unfortunately heat dissipation would be a major issue if I merely used resistors: perhaps 10W wasted as heat ((12V-2.5V)400mA + 2(12V-3.4V)*350mA), which I am loathe to do even with a heatsink. I tried it on some high-power resistors and it gets really hot (mirage illusion hot).

I consequently have been looking for a few months for a very small driver/regulator circuit (perhaps at most a cubic inch) that is very efficient and therefore produces almost no heat, and can be powered from the 12V without requiring its own power supply. I imagined that I might be able to use a switching (or other highly efficient, e.g. 90-95%+ efficiency, non-linear) voltage or current regulator, and just pretend the PWM signal is the power supply. (I would assume it would work if it had a fast power-on/off-time and no built-in anti-ripple capacitors or other mechanisms.)

I would prefer a component with long pins I can just carefully bend and screw-mount onto a binding post, and something with minimal wiring. Though the holy grail of this would be a current regulator, I've also considered a 4.5V voltage regulator... but that still forces one to dissipate something like 2W across the resistors.

The issue is that I have not been able to find such a device anywhere. Almost all regulators specsheets I see hide their efficiency (or do not provide the information necessary to determine it for various operating ranges). The efficient ones I've found specsheets of I cannot find where to order from (goggling for the part name reveals nowhere I can buy them). I've also looked for efficient DC-DC converters and haven't been able to find any (except one, which I bought which unfortunately would not work with PWM).

So my question would be: What is the easiest way to do this? Are these the components I am looking for? Where does one get such components?

(I could make a complicated circuit perhaps that uses the PWM as input to modulate a brand new 3.4V signal, but have minimal experience with complicated designs, and due to wiring issues (have to hang this), I'd rather just keep the 12V 4-pin common-anode RGB PWM as the only electrical input to the system. I would consider all simple solutions though.)

• electronics.stackexchange.com/questions/56819/… might help. Also, the reason the regulators "hide" their efficiency, is that the efficiency is a subjective thing, based on input voltage, output voltage, current draw, oscillation/filtering, etc. They normally have graphs that show you what the efficiency would be. Some might be 95%, when VIN - VOUT is 2~3V at 80% current, or it could be 60% at VIN-VOUT = 12V at 20% current draw. It's not a fixed amount. – Passerby Jun 12 '13 at 8:27
• (I had seen Passerby's URL previously, but unfortunately it was more geared towards the abstract high-level details of designing such a circuit.) – ninjagecko Jun 12 '13 at 11:15

Try using a PT78ST106V-ND Through hole SIP-3 Module it's a 1" x 1" three terminal integrated switching regulator (ISR) which has internal short circuit and over-temperature protection as well as good line and load regulation.

http://www.digikey.com/product-detail/en/PT78ST106V/PT78ST106V-ND/323539

Number of Outputs 1

Voltage - Input (Min) 10V

Voltage - Input (Max) 38V

Voltage - Output 1 6V

Voltage - Output 2 -

Voltage - Output 3 -

Current - Output (Max) 1.5A

Here's the PDF http://www.ti.com/lit/ds/slts059a/slts059a.pdf

• I don't think there are many switching regulators that perform well when supplied with PWM-ed input power... – JimmyB Jun 11 '13 at 11:59
• A quick search showed up some switching devices though that still require somewhat constant input power but can be PWM controlled by an external signal on a dedicated pin, see for instance LT3492. – JimmyB Jun 11 '13 at 13:46
• It's pretty tricky finding through hole packages, I've been hunting Linear's site. Most of their newer stuff like the LT3492 mentioned by Hanno looks like a good choice. Also check out the search Search Page I suggest you give some of the TSSOP packages a shot, as the DIP packages are harder to come by on digi-key. – crackhaus Jun 11 '13 at 15:33
• Sadly the form factors from Linear were 0.5mm-between-pins or so, and required a separate power supply (I imagined perhaps I could play a trick with a diode and capacitor to ground to create a stable Vin). The digikey product linked in the answer was obsolete and unavailable at the places I checked online. [continued] – ninjagecko Jun 12 '13 at 11:23
• [continued] However from your terminology, I googled for through hole led current (regulator OR driver) and found a recently-released(?) small-form-factor component called sparkfun.com/products/9642 with only Vin and Vout, that probably automatically senses drawn current and can be PWMed. I've ordered it and will revisit this question (possibly with a separate answer) if it works. Thank you! – ninjagecko Jun 12 '13 at 11:24

The reason the regulators "hide" their efficiency, is that the efficiency is a subjective thing, based on input voltage, output voltage, current draw, oscillation/filtering, etc. They normally have graphs that show you what the efficiency would be. Some might be 95%, when VIN - VOUT is 2~3V at 80% current, or it could be 60% at VIN-VOUT = 12V at 20% current draw. It's not a fixed amount.

Find a regulator board, look up it's regulator part number and check what the efficiency would be like from 12V In to 4V out at 1.2 Amps.

As for what you can do, it depends on your 12v source. If it is like most common RGB controllers/power supplies, the 12v is an unregulated in from the power supply, through the LED strip's common anode, and then goes to 3 individual transistors or mosfets. That's it. The PWM control is done by modulating the transistors' base pin. The PWM control section has it's own regulator to bring it down from the 12v input to 5v or 3.3v depending on the circuit they used. Pretty simple stuff actually.

Knowing this (And VERIFYING through opening the controller up), you could use a smaller voltage regulator, maybe 7V (Depends on the PWM control section's regulator), and work from there. You wouldn't need to modify much.

Other options are using three of the same led in series (since it looks like the one you posted has individual pads for each color, not common anode).

Another issue is that you are calculating the wattage incorrectly (or I am misunderstanding). Say you use a regulator down to 4V. P = I * V. For the Red LED: VIN 4V - 0.2V Transistor Collector-Emmiter At Saturation Drop (Nominal) - 2.5V LED Forward Voltage Drop = 1.3V Across the Resistor * 0.4A = 0.52W For the Blue and Green, 4V - 0.2V - 3.4V = 0.4V * 0.35A = 0.14W (x2)

That's only (0.52 + 0.14 + 0.14) 0.8 Watts across 3 resistors. Using 1 watt resistors for each, you would never even notice them getting hot.