0
\$\begingroup\$

I initially used a 74HC04 to drive the MOSFET gates. This was initially because I wanted 5V on the gate: by putting the pwm 3v3 through the 74HC04 and then re-inverting the pwm came out st 5V. I chose a 60V 20A part manf:NTD5867NLT4G. Initially this did not fully turn on. RDS_{on} was prob about 3 ohms.

enter image description here

After chaining three 74HC04 outputs (using first one to invert signal) and separating them using 33 ohm resistors (i.e 3 33 ohms to a final 33 ohms that went to Gate, to stop them back driving, etc.) the MOSFET behaved well but the first 5 μs was a mush, kind of a soft start. The 74HC04 output only gives out about 7mA so three together is enough to get the gate capacitance under control.

This obviously stressed/heated the MOSFET so I looked at the microchip TC4427 driver. This took the 3V3 PWM signal and very cleanly and HARD turned the MOSFET on. However when run at over about 24V for the heater circuit, the hard TURN OFF caused a > 60V spike. This spike eventually killed the MOSFETS. (Drain to Source typically 20 to 2 ohms: not good for my heater!)

  • I have a 100k from gate to ground.
  • I have a 33 ohms from PWM to TC4427 input.
  • I have a 10nF series 100 ohm snubber.

enter image description here What else can I do to make this circuit reliable? Do I need a better rated MOSFET even though my voltage is 37 (UN-REG 24V D.C. into 10,000 μF). Do I need other protection devices and if so are they suitable for long term reliable operation?

enter image description here

\$\endgroup\$
6
  • 3
    \$\begingroup\$ Please show a schematic. If the picture show the actual circuit, I would say there is a big possibility that it is a big part of the problem. \$\endgroup\$ Commented Sep 10 at 14:30
  • \$\begingroup\$ Hi, added schematics for the logic MOSFET driver (which seemed to work) and the TC4427 that seemed to kill the mosfets with the turn off spike. \$\endgroup\$ Commented Sep 10 at 14:56
  • 2
    \$\begingroup\$ What frequency PWM are you using? Heaters generally don't need high frequency, 10 to 0.1 Hz (or less) depending on the construction is usually fine. If you reduce the frequency you reduce the switching losses proportionally. \$\endgroup\$ Commented Sep 10 at 15:34
  • \$\begingroup\$ spehro 'spell' Pefhany: Hi yes a 5mS duty cycle. Using an STM32G474 at 168MHz so the PWM is pretty fine grained (out of 10,000). currnetyl applying 42W. Also colleague here suggested a 15R between output and gate. That has improved the turn off spike no end, might put 100R on it to slow it even more. that MOSFET driver is very fast and hard (1.5A max I think) \$\endgroup\$ Commented Sep 10 at 15:54
  • 1
    \$\begingroup\$ @Robin when responding to another user, just type @ followed by the first letter of the user name, and a list of one or more names pops up for you to choose from ... that user will then get a notification ... you can also type the user name without spaces and prefixed by @ \$\endgroup\$ Commented Sep 10 at 17:30

1 Answer 1

Reset to default
2
\$\begingroup\$

There is way too much inductance in a breadboard solution. You need tight loops to turn ON and OFF the mosfet. Long leads will generate ringning and unpredictable behavior of such a circuit.

If you have inductive loads, you need a freewheeling diode in parallel with the inductor to limit the kickback.

\$\endgroup\$
8
  • \$\begingroup\$ OK point taken , but the heater in this case will have to go through a cable run. So I am going to have to cope with inductance. At the moment the microchip driver has solved the ON 5uS delay problem but introduced the sharp TURN OFF. I am going to try a diode from the drain to the power supply next I think. \$\endgroup\$ Commented Sep 10 at 14:37
  • \$\begingroup\$ OK I have a diode back to the plus side of the inductor now and while I am still seeing a spike at turn off it looks much shorter in duration, zooming in on the scope its a ringing of about 100V but duration 1uS \$\endgroup\$ Commented Sep 10 at 15:12
  • \$\begingroup\$ the diode goes from Drain to 37 V d.c. (heater supply) \$\endgroup\$ Commented Sep 10 at 15:22
  • 1
    \$\begingroup\$ Tip: 's' = second. 'S' = siemens (unit of conductance = 1 / ohm). Capitals matter! \$\endgroup\$ Commented Sep 10 at 15:22
  • 1
    \$\begingroup\$ @Robin--Also connect a 100nF ceramic cap from the point where the diode is connected to the supply, to the MOSFET ground. \$\endgroup\$ Commented Sep 10 at 16:42

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.