I've deployed a homemade strobe system using Cree 10W LEDs in my homebuilt aircraft. My implementation uses a simple PIC controller to do the flashing, an IRL2505 mosfet (driven by an NPN transistor) for the switching, and a linear regulator (LM338) to limit the current to 5 amps. The system is using 7 Cree's in parallel for the landing light, and 8 in parallel for each wing strobe unit, each driven by an independent IRL2505/LM338 5 amp drive circuit.

Details can be found in this instructable:


After adding a protection diode to manage inductive field collapse on the wing mounted strobes, the system is working fine... however, more brightness is always better!

Here's the schematic: enter image description here

So, my specific question is this... to make this brighter, I need to go beyond 5 amps. What options are there for a more efficient, higher output, physically small current regulator?

Some implementation constraints:

  • There is little to no space in the wingtips or nose for the electronics, and heat is also an issue... I can't afford to have something melting my wings! I've had to work hard to control the heat from the LEDs and also from the linear regulator/mosfet. So... space and heat are my enemies...!

  • Switching regulators may be the right option, but I've been reluctant to try designing one because by reputation they are inherently electrically noisy... I depend on my HF radios, GPS, bluetooth and Wifi (all currently in the airplane) - so adding electrical noise could be an issue.

  • Commercial LED based strobe systems for homebuilt aircraft are out there... and I've looked at them closely. I don't see current limiters. At all. And, they are plenty bright. What gives??? What don't I understand??? Are they just keeping the duty cycle so low that the LEDs can't overheat?

ANY thoughts, suggestions, comments, or ideas would be welcome.


enter image description here


2 Answers 2


You can get rid of the mosfet by controlling the LM338 directly. The lower pin on that device is just a sense pin and can be used to force the current from the device to be a mA or less I reckon. This saves on the fet and you can use twice as many LM338s.

  • \$\begingroup\$ Are there any issues I'd see with having two LM338's in parallel? Any configurations to avoid? \$\endgroup\$ Oct 16, 2014 at 17:25
  • \$\begingroup\$ You don't have to put them in parallel - have one driving half the LEDs and the 2nd driving the other LEDs. Having said that I think they'll be OK in parallel because they are current generators and they can exist in parallel. \$\endgroup\$
    – Andy aka
    Oct 16, 2014 at 17:28

Try this circuit for your LEDs. There is no reason to use a regulator for the diode string. It is just wasted heat and power. Also paralleling LEDs without a series resistor on each LED will cause them to not share the current equally. The one with the lowest forward voltage (based on manufacturing variability) will take the most current. This circuit will give you the same light as your current circuit for 1.4 amps and generate way less heat. I used 2 resistors in series to share the heat and I needed 3.6 Ohms, which is not a standard value.

You can use more strings for more light or if you use a PWM signal with a maximum duty cycle of 50% to drive Q1 and change the resistors to 0.7 Ohms, you will get almost twice the light for the same average power. Make sure you switch at a reasonable frequency like 200Hz-1KHz and make sure your heat sink is adequate. It also may be a good idea to add a good low ESR cap between the 14V and ground to prevent any switching noise. The value of the cap will depend on your frequency but make sure it is at least 50V.

On a side note you do not need the NPN transistor to drive the MOSFET if you use one with a logic level gate. Make sure your MOSFET has a breakdown voltage of 55v or greater to prevent damage from the load dump that occurs when your starter motor turns off. This can cause very large voltages on the 14V buss. I always put a unidirectional TVS diode between the 14V and Ground (Cathode to 14V). Anything in the 24v-30v range should work. It will also prevent damage from static electricity. Good Luck!


simulate this circuit – Schematic created using CircuitLab


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.