strange behavior from a mosfet

[apchar]

I'm getting some wierd behavior on the gate of a simple n-channel MOSFET circuit. The circuit is:
A TTL input from a function generator putting out a single 5usec pulse.
That feeds a TC4427 MOSFET driver chip running at 12V.
That feeds the gate of an NDP4050 MOSFET through a 24ohm resistor.
A 200k resistor shunts the gate to ground.
The source of the MOSFET is connected to ground.
The drain is connected to a 6ohm wirewound load resistor. There is a freewheeling diode across the wirewound resistor in case of inductive kick.
The other end of the load resistor is connected to Vdd which may be as much as 30V.
A battery powered floating scope is looking at the gate voltage Vgs, with a good x10 probe.
That's it. Not much to it.
When Vdd < 5V Vgs looks fine. Above that, Vgs rings for a bit at a very high frequency on the trailing edge of the pulse. The higher Vdd the longer the ringing. Above about 12V Vgs starts freely oscillating after the input pulse. The MOSFET starts heating up and would probably burn itself up if I didn't kill the power.
I tries another NDP4050 and got the same behavior. And it is very consistent.
What would cause the gate voltage of a MOSFET to go into oscillation??

 

[Grossel]

I guess that your TTL is type open colector output. Therefore you have to connect ttl output to a resistor to 5V. If not, there will be high impedance output whenever the ttl output goes high.

 

[silverxxx]

You should probably decouple the power supply, the leads are most likely acting as inductances.

 

[audioguru]

The gate of a Mosfet needs a low value resistor (10 ohms to 47 ohms) in series with it at the Mosfet to stop ringing and oscillation.

 

[smanches]

I would remove the 200k pull-down resistor as it's redundant and just wasting power. The TC4427's are totem pole drivers, so they will pull the gate low automatically.

Do you have any snubbers across the mosfet drain and source terminals? Not having a snubber can affect gate behavior as well.

 

[apchar]

Smanches: The 200k resistor is just for the safety of the MOSFET. There is no snubber across the MOSFET. There is an internal zener diode.
audioguru: There is a resistor in series with the gate (24 ohm).
silverxx: The power supply is decoupled.
grossel: The output of the TC4427 is not open collector.

 

[mneary]

The ground on your scope probe should be connected to the circuit ground as close as possible to the MOSFET.

 

[shortbus=]

Your driver to mosfet gate distance should be as short as possible. On one App. note it said less than one inch.

 

[apchar]

mneary: the scope probe is right on the grounded source of the MOSFET.
shortbus: distance between driver and MOSFET is < 1inch.

 

[Electronworks]

Have you decoupled the MOSFET driver close to its power supply pins? Indeed have you looked at the power supply at the supply pins of the FET driver? Is this oscillating in sympathy? These devices kick 1.5A into the gate and that current has to come from somewhere i.e. its power supply. The root cause of your gate oscillating is the Cgs of the FET and any inductance on the track feeding it. However, if your supply line is oscillating too, this could enforce the oscillations on the gate and the whole thing tries to run away with itself. 100uF tant in parallel with 100nF ceramic close to the pins should be a good place to start

 

[Electronworks]

...also, what is your board layout like? Build the whole circuit on a solid ground plane (not vero board or a birds nest) with short wire lengths and you should see an improvement in performance

 

[smanches]

Smanches: The 200k resistor is just for the safety of the MOSFET. There is no snubber across the MOSFET. There is an internal zener diode.

Definitely place a snubber cap and resistor (if needed) across the drain and source, as that high voltage ringing is often a cause of failure.

However, as mentioned, the mosfet drivers do need good decoupling with enough cap to handle the current needs, and that very well could be the issue. What does the waveform of VDD into the drivers look like?

 

[apchar]

electronworks: It's on solderless breadboard. I know a lot of people hate the stuff but I've never had any problems at such low frequencies. Just the same, I'm going to rebuild it with good old dead-bugs and point-to-point to see if it helps.
smanchez: I'll try a snubber and I'll pile more capacitance on the TC4427. But it's just wierd to see a MOSFET switch oscillate. As far as the waveform goes, for Vdd < 5V the gate voltage looks like an ordinary 12V pulse. For Vdd>5 the oscillation grows on the trailing edge of the pulse and continues for a few milliseconds. The length of the oscillation varies randomly but the higher Vdd, the longer the oscillation. Above about Vdd>12 The oscillation never stops.
Apchar

 

[audioguru]

Solderless breadboards cause most of the problems with circuits on these forums. It is for DC circuits that have a battery, a switch and a buzzer.
Usually the contacts are intermittent.

The high inductance, high capacitance and the many long connecting wires acting like antennas cause problems with AC circuits.

 

[smanches]

What is the source of Vdd? Could the current through the mosfet be causing a high voltage drop on the power source which is causing the oscillation? This happened to me once when I was trying to build a high current (30A) super cap charger. I had to put a diode before my control circuit power supply to prevent those input filter caps from being drained by the main switches. It didn't quite cause an oscillation like you're getting, as the drop was so bad the circuit would just stop functioning.