mosfet; p-channel - basic question

[tester43]

MOSFET: IPP120P04P4L-03
Datasheet: https://www.infineon.com/dgdl/Infineon-IPP_B_I120P04P4L_03-DS-v01_01-EN.pdf?fileId=db3a30432f69f146012f783b3b5a2e3f

Maximum Vgs = +-16v

Voltage I will be controlling: 24V peak

Does it mean that I can't just tie the Gate to GND because it will create 24V Vgs which is more than 16V and will break the device?

[Zero999]

Yes, you'll kill the device, if the MOSFET is in the +24V supply and the gate is connected to 0V. If speed isn't an issue, connect a 10V zener between the gate and drain and connect that to 0V, via a resistor. If this is being run off a microcontroller, then the zener can be avoided: use BJT amplifier in common base configuration.


V2 = V_input
V_GS = (V_input-V_BE)*R2/R1
V_BE = 0.6V

V_input = 5V
V_GS = (5-0.6)*1000/470 = 9.36V.

Why not switch the 0V side? It's much easier and can be done with an N-channel MOSFET.

[tester43]

Hi again
Thanks for this solution.
Regarding "why not N-Channel" part: i'm still thinking which way to go with my soldering station power.
I have all the info I need: AC Zero crossing interrupt signal, opto coupler to drive the gate etc. The only thing that is bugging me is the idea of doing all measurement with AC voltage present at the connector. I know that it will be in zero (or close to it) but this concept is somehow.... counter-intuitive
I like this idea of having P-Channel oriented FET. With it I dont have any AC on my load when I need to do the thermocouple reading.
Load is T12/T15 hakko tip:
http://dangerousprototypes.com/forum/download/file.php?id=12167&mode=view

http://dangerousprototypes.com/forum/download/file.php?id=11012&mode=view

[StillTrying]

If the 24V is raw rectified DC, one minor problem you'll have with these simple methods of switching the P-Channel is that it won't switch on and off at the zero crossing point, but every time the 24V rises and falls through 2V - 4V. So its near-zero dissipation will be a bit higher than near-zero.

[tester43]

hi, yes - it's because of threshold voltage - if I remember correctly.... Would need to go to -10V at 0 AC.

[Zero999]

Yes, if it's rectified, unfiltered DC, then that could cause a problem, when the voltage is low, since the MOSFET won't turn on properly, although the current will also be low at this point, so the power dissipation might not be too high.

Why not simply use a 24VDC switched mode power supply, rather than a transformer, to power the whole circuit? They're not expensive and using DC, rather than AC, will make things easier.

[tester43]

i could also use capacitors and filter power to the complete DC but.... I want to learn AC and mosfets.
I could dump the concept of positive ac and do the back to back mosfets. For the moment I am thinking how to produce kind of "dynamic" gate voltage that would be always 10V less than source. Do we have any known patterns how to do it easily?
I am thinking on building a negative supply (in relation to ac zero), and use linear regulator (lm##10) with GND pin attached to src - will it work?

[StillTrying]

For the moment I am thinking how to produce kind of "dynamic" gate voltage that would be always 10V less than source. Do we have any known patterns how to do it easily?

You could easily create a zener shunt supply 10V below the 24V raw DC, and use an opto to switch the mosfet at the zero-crossing points, very little current is needed.
But from the the simulation I've just done using a Si4425BDY mosfet(12mR 3.8V) it's not really worth it.
With it's lower on resistance and threshold voltage a IPP120P04P4L-03 mosfet should be even better than this.

The 1st plot is the voltage across a 9R load, Yellow switching at zero crossing and Red switching on and off as the 24V crosses 3.8V.

The 2nd plot is the Watts in the 9R load, the small difference at the bottom of the waveform averages only about 100mW in 25W across a full cycle.

The 3rd is the dissipation in the Si4425BDY mosfet, the Red peaks the one switching when the 24V crosses 3.8V.

[Zero999]

i could also use capacitors and filter power to the complete DC but....

The trouble with that is it would increase the output voltage to much higher than 24VRMS: near the peak voltage of the transformer, if the capacitors are big enough. It would also need a considerably larger transformer, since the nasty pulsed waveform, drawn by the rectifier, causes more heating in the transformer, than a resistive load. A compromise could be a small capacitor, to prevent the voltage from falling below 5V or so, which should be enough to keep the MOSFET on enough, to prevent excessive heating.

I want to learn AC and mosfets.
I could dump the concept of positive ac and do the back to back mosfets. For the moment I am thinking how to produce kind of "dynamic" gate voltage that would be always 10V less than source. Do we have any known patterns how to do it easily?
I am thinking on building a negative supply (in relation to ac zero), and use linear regulator (lm##10) with GND pin attached to src - will it work?

It's possible, but would require a negative power supply.

Scrap the P-channel MOSFETs. Use N-channel MOSFETs: they have a much lower on resistance, for the price. The power can be coupled to the gates, via a small transformer (left), or capacitors (right). If the frequency is high enough, the internal gate capacitance of the MOSFETs is sufficient to keep them on. As the voltage is only 24V, not mains, the transformer is overkill and ordinary ceramic capacitors can be used, rather than special Y1 capacitors.


https://www.eevblog.com/forum/projects/thermostat-control-relays-vs-triacs/msg71441/#msg71441