0
\$\begingroup\$

First off, my electronics knowledge is really basic. I learned some at school and very simple projects with Arduino. I'm a scale modeler making a specific diorama with LED lighting.

I'm trying to design an Arduino Mega shield that will allow me to turn on and off six 12v RGB LED strips fading in and fading out. After some research, I found that the best way to do so was to use N-channel MOSFETs as switches, instead of Darlington transistors, due to heat issues. The current taken by each strip is between 100mA and 400mA.

Also, I might be interested in varying the color in each LED strip, depending on the need.

I came out with this circuit (x6), where D2 is an Arduino pin sending the PWM signal, SV1 is the connector to the LED strip and the potentiometers are intended for color adjusting. I used 2SK3918 MOSFETs (I know they are not logic level and I can change them for logic level ones). R1 is the pull-down resistor to ensure 0V at the MOSFET when initializing. I've read somewhere that there's a need for a Gate resistor, but somewhere else I read it's not needed.

enter image description here

I made a prototype board, connected it outside the arduino but couldn't make it work.

enter image description here enter image description here

I connect 12v and GND to X1. 0-5v (testing up to 20v) power supply to D2 and GND and a test LED strip to the corresponding terminals at SV1.

No matter what voltage is applied to D2, the LED strip is continuously on. There's 12v all the time at each pin in SV1.

I know I'm missing something, but I don't know what it is.

\$\endgroup\$
4
  • 1
    \$\begingroup\$ Why didn't you simulate first? Why are both power rails at GND? \$\endgroup\$
    – Andy aka
    Jan 5 at 22:24
  • \$\begingroup\$ I didn't simulate first because I didn't have any of the parts and it was cheaper to make the prototypes than buying the components for simulation. About both power rails at GND... I don't know. Probably I saw somewhere that was a good idea. Is it preferably one at GND and the other at 12v? \$\endgroup\$
    – PSedney
    Jan 5 at 22:28
  • 1
    \$\begingroup\$ By "simulation" we mean on a computer with something like LTSpice or Falstad: falstad.com/circuit/circuitjs.html \$\endgroup\$
    – Polynomial
    Jan 5 at 22:29
  • \$\begingroup\$ OH! I didn't know about those tools. I only knew Pspice which was paid software \$\endgroup\$
    – PSedney
    Jan 5 at 22:30

1 Answer 1

Reset to default
0
\$\begingroup\$

You've connected the top side of the potentiometers to GND, and the bottom side of the potentiometers to the MOSFETs, which then also connect through to ground.

When you plug the lighting strips in, there's a path from the 12V supply (pin 1 of SV1) through the lighting strips, back through SV1, through the potentiometers, and back to ground regardless of whether your MOSFETs are on or off.

It isn't clear what those potentiometers were supposed to do. You should remove them and just have the lines from the LED strips going straight back into the drain pins of the MOSFETs.

I used 2SK3918 MOSFETs (I know they are not logic level and I can change them for logic level ones).

If you're using 5V logic levels, the 2SK3918 is fine. You're way above its Vgs(th) at 5V and you can see from the forward transfer characteristics graph in the datasheet that the forward current (Id) is more than 10A at a Vgs of 5V.

That said, these MOSFETs are massive overkill for your design and I would recommend using smaller ones in future.

\$\endgroup\$
14
  • \$\begingroup\$ Thank you very much for your answer! Well, each 4 pin connector should drive one independent strip. Each strip will be lit in a different color. I thought about pots as a way to change colors in the strips, but not mandatory. I'll try removing the pots. \$\endgroup\$
    – PSedney
    Jan 5 at 22:35
  • \$\begingroup\$ Technically you would've had the right idea there you hadn't connected pin 3 of the potentiometers to ground. However, you'd be dissipating a lot of power in the potentiometers (at 400mA the pots would dissipate several watts, which would quickly turn them to molten slag) so it's not really the correct approach. If you used separate PWM signals for each MOSFET, you could independently control the channels. \$\endgroup\$
    – Polynomial
    Jan 5 at 22:38
  • \$\begingroup\$ For now, if you desolder the pots and bridge pin 2 to pin 1 on the PCB, leaving pin 3 disconnected, that should get your circuit up and working. \$\endgroup\$
    – Polynomial
    Jan 5 at 22:39
  • \$\begingroup\$ Yes, I wondered they were too much, but I thought smaller ones like SOT-23 wouldn't be able to take the current (I need to understand the datasheets better :) ) \$\endgroup\$
    – PSedney
    Jan 5 at 22:39
  • 1
    \$\begingroup\$ hi! Finally the new boards arrived. I tested them and work perfectly. In this new version, I removed tge potentiometers and added gate resistors. Thanks for your help and comments. Cheers! \$\endgroup\$
    – PSedney
    Jan 28 at 22:43

Your Answer

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

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