LED Light Strip Control Circuit not working once assembled as a PCB

Question

My project is to create a controller for some warm and cool LED light strips. The parts I used were an Arduino SAMD21, two FZ44N MOSFETs, and two 10 kΩ resistors. The Arduino sends a PWM signal to the MOSFETs, and the MOSFETs controls the ground of the LED lights.

Technically this is a lie. The prototype had three MOSFETs and three resistors. I had thought that I could control the overall brightness using one MOSFET to control the source of the other two MOSFETs. This didn't work, so I left brightness control MOSFET at 100% PWM so that the other two MOSFETs always had access to ground.

I created a prototype using a protoboard and it worked perfectly.

I then designed and purchased a manufactured PCB with pre-soldered components (this time with only two MOSFETs). The components are, as far as I can tell, identical in functionality. The resistors and MOSFETs are surface mount instead of through-hole. The MOSFET that I used were FZ44NS. There is also a variation in which pins I used on the microcontroller, but they are still analog PWM according to the spec sheet. The pins I used on the prototype were A8 and A9, and on the PCB one I used A1 and A2.

Now the issue:

At any duty cycle except for 0%, the MOSFETs (both of them) constantly pass ground through the source to the drain. I am measuring the voltage between my 24 V power source and the MOSFETs drain and it's always around 23 V (±1 V).

I have checked:

• The microcontroller
  • The voltage on the output pins to the MOSFETs vary depending on the PWM value I pass in code
• The resistors are 10 kΩ verified using a multimeter.
• The continuity of all the components; they lead exactly where they are supposed to.

The root cause of the issue (again, as far as I can tell) is the MOSFETs. I have either missed something big, or they are both faulty. I have some next steps to confirm, but I thought I would reach out for help in case it was something else.

Here's a table of the voltage measurements under no load:

Vg	Vd
3.3 V	24.3 V
.01 V	23.8 V
0 V	0 V

I'm going to attach images of:

• PCB schematics
• PCB's copper layer
• Manufactured PCB
• Tinkercad of the original prototype
• A photo of the original prototype

I am at a total loss as to what I'm doing wrong. I would love help. Thank you.

Answer 1

The 3.3V digital output of the SAMD21 seeduino is not sufficient to enhance the IRFZ44 MOSFET, take a look at the gate voltage to drain current chart below.

At 4 volts (the lowest the chart goes) the slope is very high - as I see it the implication is that there's no guarantee it will conduct below 4V. Note as well the listed threshold voltage of 2V to 4V is for a drain current of 250 uA - not enough to get any real work done in your application.

I saw that you are concerned your parts may be defective - I'd try applying 5V to 9V (from a convenient supply or 9V battery) from gate to source to see if you can turn on the MOSFET and confirm if your parts work or not.

So why did it work before?

I offer two possibilities:

• operating so close to the threshold voltage leads to unpredictable behaviour batch to batch or even fet to fet, one could conduct enough current to light your LEDs while another will not.
• the previous parts you received are fakes with a different threshold voltage, vendors on ebay, amazon and aliexpress are nortorious for fakes.

What to do?

A quick search for discreet MOSFETs in the D2PAK package with Vth<2.5V on digikey yielded quite a few results that may work. Find a FET from a reputable vendor of your choice that carries current in excess of your load current at a Vgs of 3.3V. Here's my quick search: https://www.digikey.ca/short/5w97w3n5

All images sourced from IR's datasheet for the IRFZ44NS from here: https://www.infineon.com/dgdl/Infineon-IRFZ44NS-DataSheet-v01_01-EN.pdf?fileId=5546d462533600a40153563b43a5220f