Hi all,

I'm trying to drive a RFP30N06LE N-Channel MOSFET with TC4420 MOSFET Driver. The MOSFET is used to switch a [1W LED][1]. The input to the MOSFET driver is provided by the 5V from an Arduino Nano. But my circuit is not working: it seems that the VDD somehow leaks to the input of the MOSFET driver, causing it to switch on constantly. The Arduino side switch has no affect.

The VDD in the circuit is 16V, and VCC is 5V coming from an Arduino Nano. The VCC goes into the MOSFET driver's input to turn it on.

I find my LED operates best at a current draw of 291 mA. With a VDD of 16 V, I put the LED in series with a 56Ω resistor.

Here is my circuit:


And photo of the breadboard implementation (LED and resistor are off board and not shown here). Power rail runs VDD.


I'm providing the VDD with a bench power supply. I have my scope hook to 1) yellow-MOSFET driver input 2) Blue-MOSFET driver output 3) Purple-Between drain and source of the MOSFET.

You can see in this picture when I first turn on the VDD power supply at 0.5V. The yellow signal in the scope raises above the others by a little bit. I measured ~0.4V on the input pin of the driver. This does not really make sense to me because the input should be isolated from the VDD. The LED (in the foreground of the photo on the left) does not turn on.

Now when I set power supply to 4.1V. The LED turns on and it starts to draw about 0.47 A. This should not happen since I'm not giving anything to the input of the MOSFET driver (the switch between Arduino 5V and the MOSFET input is open). Yet the scope shows the voltage on the MOSFET input is as high as on its output, turning the MOSFET on. See photo below.



I suspect there are some shorting between the MOSFET input and VDD so I check with my DMM but they are open.

I'm really at a lose as to what to do. Please let me know I didn't make something clear or if you want more information. Any insight would be greatly appreciated!

 

I don't see on your breadboard any power/VCC connected to the Nano, unless it is hidden under the board.
You need a separate power supply for the Nano, you cannot feed the 16V to the Vin, must be 12V or less.
Without power, the Nano pins may be high impedance, no current passing anywhere.
Or are you powering by the USB port?

 

Have you checked you button/switch S1. It might be 90 degrees rotated.

Gate drivers hare fast power monsters that want bypass capacitors very close.
I drew two red lines where I would put capacitor(s) from power to ground.
The purple line is the length of the power loop which is toooooo long.

 

I don't see on your breadboard any power/VCC connected to the Nano, unless it is hidden under the board.
You need a separate power supply for the Nano, you cannot feed the 16V to the Vin, must be 12V or less.
Without power, the Nano pins may be high impedance, no current passing anywhere.

Thanks @sagor for your reply! I'm powering the Arduino from from the USB cable during the testing, which are not shown in my initial posting. Here's an updated breadboard connection with annotations.

 

Do you have signal at the Gate Driver input?
What does the MOSFET Gate look like?

 

Have you checked you button/switch S1. It might be 90 degrees rotated.

Gate drivers hare fast power monsters that want bypass capacitors very close.
I drew two red lines where I would put capacitor(s) from power to ground.
The purple line is the length of the power loop which is toooooo long.
View attachment 270560

Thanks @ronsimpson for the insight. You are suggesting using two bypass capacitors for each of the VDD pins of the driver. Could you clarify why a power-hungry device like the driver wants to have bypass caps close to their VDD?

I got help from another platform suggesting I add a pull-down resistor to the driver's input, so it can be pulled low when no input is given.

Also, could you please explain why a long power line is bad? I can't think of any other than causing clock skews.

 

If the Gate Driver to Gate resistor was 0 ohms this driver can push/pull 6 amps to the Gate.
Even with a very high 220 ohms the current can be very high or 20nS. (maybe 2 amps)
Wires have a delay. Maybe only nS. Current to the ICs needs to come from storage, like a capacitor. It is possible for a wire to have 16V on one end and 5V on the other for 2nS. Your scope might not be able to see this fast. The IC will not like having the supply falling to a very low voltage on every switch.

 

I modified the circuit with:

1. added a 10K Ω pull-down resistor
2. using two bypass capacitors for each VDD pin of the MOSFET driver, per @ronsimpson's suggestion
3. reduce the length of the power loop, again per @ronsimpson's suggestion

Here're the new schematics and breadboard connection



Here are the signals I'm getting now @ronsimpson; when I set the bench power supply to minimal at 0.05V, I'm able to get a trigger on the driver's output with a rise time of 20.5 µs. Please see the following scope photo. The yellow (ch1) measures the driver's input, and the blue (ch2) is the driver's output. The purple (ch3) is the drain-to-source voltage of the MOSFET.


However, when the input still rises and falls with VDD despite adding a pull-down resistor at the input. The behavior has not changed. Here's a video showing it not working. You can see when I turn up the VDD from the power supply. Both yellow (driver input) and blue (driver output) rise, and the LED turns on. Pressing the button will give the driver input a small rise, which I don't quite understand. But the button does not affect the switching. The MOSFET is always on. Any suggestion on what might be going on?

 

It appears there is not LED current. Is the LED backwards? Can you put R1, D1 across the supply and see if it lights.

 

It appears there is not LED current. Is the LED backwards? Can you put R1, D1 across the supply and see if it lights.

Thanks for the follow-up @ronsimpson. I double checked the LED connection and connected the D1,R1 series directly over VDD and GND. It lights up as expected so this is checked.

 

It appears there is not LED current. Is the LED backwards? Can you put R1, D1 across the supply and see if it lights.

Continuing from my last, in my test with LED and R1 directly across VDD. I set the VDD to 4.4V and measured 2.3mA passing through D1 and R1.

Then I tried putting the LED back into the MOSFET drain. With VDD at 4.4V, I got the exact same measurement. So the MOSFET is on. The scope also shows high on the input. In fact, the input is at the same level as VDD. The driver input always follows the VDD. This is the most confusing part, and I checked there is no short between VDD and driver input. I swap couples of TC4420 by now, but each with the same behavior...

 

Leave R1 connected to the MOSFET-G.
Pull out the other end of R1 and connect to ground. (no light) and connect to any voltage more than 4V. (should light LED)
Simple MOSFET test with out Gate driver and no computer.

 

Leave R1 connected to the MOSFET-G.
Pull out the other end of R1 and connect to ground. (no light) and connect to any voltage more than 4V. (should light LED)
Simple MOSFET test with out Gate driver and no computer.

@ronsimpson I tried your suggested test for the MOSFET. It works. Here's the trigger plot from my scope.


I have blue connected to the gate and yellow measures drain-to-source. The MOSFET opens OK when applying a VDD above 4V, and the LED lights up.

It was brought up that the MOSFET driver might be defective. But I'm not sure how to just the driver just by itself...

 

The driver does not work if its supply is less than 4.5 volts.

 

The driver does not work if its supply is less than 4.5 volts.

Thanks for the reminder @ronsimpson. I put together this circuit to test the driver with a supply of 9 volts.

Given that my connection is correct, I think the driver might be defective. This is what happened when I connect the input to 5V:

Again, purple is the driver input (between the pull-down resistor and GND), and blue is the driver output. When the input is connected to 5V, the voltage is pulled low to give to the input. However, the blue output is not affected! I got my TC4420s in a package of 4. All of them have more or less the same behavior. The blue never moves a bit. Additionally, I noticed the driver gets hot just with its GND and VDD connected. It gets worse when there's input.

I think this means my package of TC4420s are bad?

 

Quiescent current = 450uA
Thermal resistance of package = 125°C/Watt
For a 5V supply, that's 2.25mW, which would result in a temperature rise of 0.28°C

 

Problem solved!

My new TC4420 arrived today, and I swapped my old one out with the new one. It switches like a chime.
With the old ones, when I close the gate, the Vout doesn't have any change (see my previous post).
And here is the switch event of the working new drivers. The gate is in purple, where I connect it to Arduino 5V to turn the driver on. The rising edge of Vout (blue) from the driver follows closely. VDD is at 10 V.



I got the old one from Amazon. Now look at them closely. They don't have the little dot indent at the lower left side.

I see this dot in both my new drivers and in the datasheet. Maybe the old ones I have is a different packaging, and I may have messed up the pin layout. I'm not sure.

Anyways, problem is solved for now. Thanks all for you help!