High power LED strobing - What's limiting my current?

Hi all
(noob to this forum… apologize for hand-drawn schematics… )
Ive got a project working to synchronize high power LED strobe to a motor, with the help of a Hall sensor sampled by the Arduino. Both the motor and the LED bank are driven via IRF520 MOSFETs whose gates are directly driven by digital outputs. Works nicely BUT… Cannot get sufficient power from the LED, and wondering where my bottleneck might be. Here are some details on the LED side of the circuit:

  • LED bank consists of 12 parallel connected 10W COB LEDs (each one rated at 12V 0.8A)
  • Code activates an LED flash of 1.5 milliseconds, once per revolution, or at about 15Hz
  • The LEDs do flash brightly; but not enough. To check the pulse current, I connected an 0.1 Ohm resistor in series with the IRF520 and read it’s voltage with scope. The pulses read about 0.43Volts, meaning 3 Amps, for about 1.5 msec
  • …However - NOT 10 Amps, which is the current draw I was expecting from 12 LEDs connected in parallel.
  • Tried to improve it by increasing VCC (whilst also increasing the series resistor on the Nano’s Vin pin to protect it), all the way up to 19 Volts. But, this increased the pulses from 0.3V to only around 0.5V, so only 5Amps.

Cable leading up to the LEDs is relatively thick and about 1 meter long. Placing a 2000uF cap across the power supply didnt help.
Any ideas… ? ? ?
Thanks

Try measuring the voltage drop across the MOSFETs. It has a gate threshold of 4V but it is not fully enhanced. I expect you are getting about a 4V or greater drop across the MOSFET with your current configuration. Fully enhanced you should get about 2.7 volt drop across the MOSFET. If you measure the gate source voltage during the pulse that will let you know how close to enhancing the MOSFET you are. I would suggest a different logic level MOSFET with a higher current rating. I expect if you leave it so the LED stays on the MOSFET will get smoking hot.

IRF520 requires more than 5V Vgs to be fully open. Use something like IRLZ44N instead

Check this:


Even at 50V Vds you’ll get around 3Amp if you control IRF520 directly by Arduino pin (5v Vgs)

irf520_t.jpg

driven via IRF520 MOSFETs whose gates are directly driven by digital outputs.

Mosfets have 3 pins. Your schematic shows something with 6 pins? Please post a link to that. It might be a module containing an opto-isolator or something. In which case the advice given so far, in good faith, might be mistaken.

Other errors I spotted:

22R in line with Nano's Vin pin. Bad idea and uneccessary.

Pot connected to 3.3V. This won't be able to give an analog reading much above 675. Should be connected to 5V I suspect. Could this account for your low brightness somehow?

PaulRB: Mosfets have 3 pins. Your schematic shows something with 6 pins? Please post a link to that. It might be a module containing an opto-isolator or something. In which case the advice given so far, in good faith, might be mistaken.

According to TS description he is using this module (most likely)

Great help - Wonderful thanks! @gilshultz - Will check the voltage across the IRF520 S-D when open. @alesam - That's very very insightful and very instructional. Will try to maybe add a transistor stage and amplify the digital output slightly to around 7V. Kudos! (And btw yes - Using the module you sent an image of) @PaulRB - thanks. Agree about the pot. But this is related only to the range of motor speed control, so not worried about it. Why is the 22R resistor inline with Vin a bad idea? (Im wary of approaching the Nano's 12V max Vin rating. Unnecessary risk, there....)

noamcee: @alesam - That's very very insightful and very instructional. Will try to maybe add a transistor stage and amplify the digital output slightly to around 7V. Kudos! (And btw yes - Using the module you sent an image of)

Going around it the wrong way. :roll_eyes: Just use the correct FET.

Use this module: |500x500

noamcee: Why is the 22R resistor inline with Vin a bad idea? (I'm wary of approaching the Nano's 12V max Vin rating. Unnecessary risk, there....)

Because it is useless.

You are right to be wary of operating the Nano at 12 V. It is not rated for 12 V. (If the Arduino website was not "down" at present, I would cite the link!) You need a proper 5 V regulator - switchmode "buck" converter - fed to the "5V" pin.

According to this page

Input voltage 7-12V

So: 12V is on, but within, the limit. In this circuit, there is little to draw current from the 5V pin other than the Nano itself, so it should be ok to be on that limit. But if you extend the circuit so that more current is drawn...

The resistor is not needed, but not going to cause a problem in this circuit because there won't be much, or much variation in, the current drawn by the Nano. But in general using a resistor is not a suitable way to reduce input voltage because in most circuits, the current could vary significantly depending what peripheral components are drawing at any time. This changing current will result in a different voltage drop across that resistor. As more current is drawn, the input voltage to the Nano will drop. The Nano's regulator will be able to cope with that, up to the point where the input voltage falls below 7V. However, the Nano's regulator would burn or shut down before that point was reached with a 12V input.

Back to your led brightness problem: now we know what those MOSFET modules contain, I agree they are the problem, because 5V input signal will not fully switch them on. Why are such modules sold for use with Arduino? Because profit can be made by selling them to the inexperienced.

LOL wonderful Thanks so much @PauRB