PWM drive MOSFET not working properly.

I have an NTE2987 N-Ch Logic Level, Enhancement Mode High Speed Switch that is acting strangely, probably because it's not wired properly: http://www.nteinc.com/specs/2900to2999/pdf/nte2987.pdf

  • Arduino Uno board running at 5V.
  • 12VDC power going straight to a LED with integrated current limiting resistors. (Tested the LED with a normal setup and they work fine at 12V)
  • PWM coming from Arduino pin 10 goes to the left Gate pin on the MOSFET. This gate is also shorted via a 10kohm to the Arduino ground.
  • MOSFET Drain pin hooked up to (-) LED side.
  • MOSFET Source hooked up to Arduino ground.

At times, the gate seems to saturate and the PWM signal has no effect on the current going through the MOSFET, and even killing the Arduino keeps the light going like if the gate was permanently open. So I added a pulldown resistor.

Now that that resistor's there, using a Fluke I verified the voltages across the MOSFET. Since I'm using a variable power supply, I can see that when I increase the voltage across the resistor, it spikes for a second, then goes back to zero.

Question 1: Can the MOSFET and the gate's pulldown resistor share a ground on the GND Arduino pin? If not, do I hook it up to the power supply's 12V negative terminal? Should I just put all grounds on the power supply (-) port? The power supply has a black -, green gnd, and red +. Do grounds go on black or green?
Question 2: am I using the proper pulldown resistance?
Question 3: is my MOSFET dead?

Should I just put all grounds on the power supply (-) port?

Yes, connect the Gnds.

Add a 220 ohm between Arduino and gate to avoid any risk of overloading the
Arduino pin.

There is only one source terminal so it has to be the common point between control
and power circuits, ie connected to both grounds.

The symptoms you describe sound more like the MOSFET source was
connected to 12V ground only, BTW, so a bit uncertain what's happening.

How would that happen? Could the MOSFET send a spike down the gate to the pin?

Schematic?

How would what happen?

Imgur

Imgur doesn't work here? Anyway, there it is.

Connecting all the grounds got it to work. Now I hooked up an o-scope to the MOSFET, and the voltages seem to be "upside down" (ground probe to ground, and bigger probe to drain. The PWM signal seems to be inverted, at 0% duty cycle, the voltage is at max (12V), and at 100% duty cycle, there is no voltage. Is it because I have the probe on the wrong side? I guess it makes a little bit of sense that while the light is out, there is no voltage potential across the source and drain. How do I probe the circuit to see the signal?

dgcaste:
I guess it makes a little bit of sense that while the light is out, there is no voltage potential across the source and drain.

Wrong. What you are seeing is correct. When the light is turned off, all the voltage is across the source and drain of the MOSFET. When the light is turned on, all the voltage is dropped by the light/resistor, and the MOSFET should be dropping very little.

I SWEAR I mean to say "on" rather than out. Anyway, thanks for helping me clear this up. So, the only way I can look at this square wave is "inverted"?

dgcaste:
Connecting all the grounds got it to work.

Imagine.

dgcaste:
How do I probe the circuit to see the signal?

Provided that your scope and circuit are "isolated" (your FET circuit is on batteries or the DC source is not referenced to Neutral/Chassis) then you should be able to place the scope across the LED and the resistor.

Hi, with the MOSFET you have to realize that basically you are using a voltage on the gate-source to switch a current through the drain-source.

Tom..... :slight_smile:

dgcaste:
How would that happen? Could the MOSFET send a spike down the gate to the pin?

Driving a MOSFET gate is driving a large capacitive load. Logic signals normally
cope with upto 50 to 100pF no problem, but MOSFET gates can be upto 20nF depending
on size/power - thus the Arduino 40mA current limit will be exceeded for brief periods
on each switch over - the datasheet doesn't say anything about brief overcurrent
limits on I/O pins, so play safe and use a gate resistor to limit it.