1
\$\begingroup\$

I am having trouble getting some MOSFETs I have to behave like I expect, but I am just testing them with an oscilloscope, battery, and appropriately high voltage digital output on a microcontroller.

Do I need to attach a load such as a motor, solenoid, or light bulb to bench test a MOSFET?

I don't know how a MOSFET works physically, so if there is some sort of capacitance that needs to be charged inside the MOSFET for it to work (i.e. current received and charge stored from the source), then the oscilloscope might not be supplying this.

The MOSFET I am using is this.

EDIT:

This is the setup I have: The battery is 6V, and the Gate is a wire. If I hook up the wire to ground (drain pin) or 6V, the meter reads 0V and 6V respectively, so it does seem to be switching now that I removed the microcontroller. This 6V is stable, and does not revert to 0V until the Gate is discharged to ground. Interestingly, if I hook up the Gate to the source pin, I get 5.5V output, not 6V. This 5.5V is stable (and needs to be "reset" by touching Gate jumper to ground).

schematic

simulate this circuit – Schematic created using CircuitLab I found out that the strange stuff occurs while using my microcontroller to switch the MOSFET. If I hook up the ground from my Teensy3.1 3V microcontroller to the MOSFET drain, the meter jumps up to 6V, as if it was switched on. This 6V is also noticeably more noisy than the 6V of the battery. This 6V is not stable, and the meter immediately drops to 0V when the microcontroller ground is disconnected.

enter image description here

If I instead connect the digital output (3V) of the microcontroller to the Gate of the MOSFET, the output of the meter jumps from 0 to 4.5V, with 1.5V square pulses I am sending from the microcontroller riding on top. 4.5V + 1.5V = 6V obviously, but confused. I expected the pulses read on the meter to be 6V in amplitude, from a baseline of 0V. I checked and the control voltage pulses I am sending from the microcontroller are 3V as expected.

enter image description here

If I now connect the microcontroller ground to the drain of the MOSFET, the voltage meter output jumps to 6V, and these 1.5V square waves riding on top of the 4.5V disappear. This effect is reversible.

\$\endgroup\$
  • \$\begingroup\$ Your set up is incorrect. I'm looking over what you said and your circuit. I don't mean to be rude, but your schematic drawing, is awful. Why is that voltmeter so off to the side with awful angles for connections ? You don't need to label terminals for common components like this, the people who answer your question will know whats what. \$\endgroup\$ – efox29 Aug 14 '15 at 3:14
  • \$\begingroup\$ I never used this drawing tool before. If it is easy for you then can you correct it? Otherwise I could take pictures tomorrow. \$\endgroup\$ – user391339 Aug 14 '15 at 3:16
2
\$\begingroup\$

It will switch, but you won't be able to see it, since the oscilloscope can only measure voltage. You need to give it some voltage to measure, and the easiest way to do this is to connect a power supply to the drain of the transistor through a resistor. For example, if you have a 5V power supply, a 1000Ω resistor will pass 5mA, and you will easily be able to see the difference between MOSFET on (near 0V) and MOSFET off (5V).

\$\endgroup\$
  • \$\begingroup\$ You could even use the internal pull up resistors in the arduino as a load, \$\endgroup\$ – W5VO Aug 14 '15 at 4:16
2
\$\begingroup\$

First, you're setup is wrong.

This is what you want to have.

schematic

simulate this circuit – Schematic created using CircuitLab

You have an N channel MOSFET. The mosfet will begin to conduct when Vgs > Vgs(th). According to the datasheet, that's about 2.3V. So if you feed in 6V, great.

Notice how the circuit above and yours is different.

The first being, I have load (R2). Resistors are good. They limit current. You want to limit current when you are doing stuff, unless you have a very good reason not too.

Second, the source is connected to ground. Typically in switching applications using NMOS, the transistor is used in this configuration and is also called a low side switch. It switching the load to ground (hense the name). There is also something as a high side switch, but that's for another time.

Third, I have a resistor on R1. As you found out through your experimentation, the gate voltage remains until you tough it to ground. The gate of a Mosfet is a capacitor. When you charge it, like all capacitors, it will hold its charge, until something causes it to discharge or through natural leakages (since nothing is perfect). The resistor is there to allow for the transistor to turn off when SW1 is not connected to the battery.

Fourth, the voltmeter is in parrallel. This is how you measure voltage and its connected to the load (R2).

This setup should give you predictable results for your testing. I don't know where you got that circuit from before, but its not good.

You can now modify this circuit and connect it to your microcontroller by removing SW and connecting it directly to your microcontroller. 3.3V would be enough to this to switch on.

Added after comments

schematic

simulate this circuit

This would be the setup you would be trying to go for. The pin of the microcontroller connected directly to the gate and the ground of the microcontroller connected to the source.

\$\endgroup\$
  • \$\begingroup\$ Thanks for your help. I will try your circuit when I go back tomorrow & give an update. \$\endgroup\$ – user391339 Aug 14 '15 at 3:44
  • \$\begingroup\$ Your circuits 10k resistor helped with turning the MOSFET off when gate disconnected from +. Still having strange problem with adding MCU. Connecting MCU ground to the circuit ground trips the MOSFET. Connecting MCU + (alone) to Gate introduces noisy baseline with slight positive bias. \$\endgroup\$ – user391339 Aug 14 '15 at 22:11
  • \$\begingroup\$ @user391339 is your connection like the second schematic ? \$\endgroup\$ – efox29 Aug 14 '15 at 22:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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