High end P type switch.

As Max will tell you, I never really use FET's..

However I have the need... Input voltage can be 12v, 24v or 48v..


The P Mosfet needs to be a TO92 type (Microchip does loads) But as it only powers a opto switch I can get away with a bit RDon resistance.. I picked this

What I need to know is, what happens with the gate voltage.. I've seen designs that implement a Zener to protect the gate Nominally this will be on 24v systems

The display has a separate on / off switch, so I'll be using the switched 5v to turn on / off the opto switch ( it stays powered at the moment..)

Also 30v... Should I go higher?

The little MOSFET has a max Gate to Source voltage of 20V. You need to keep this G-S voltage in the 5V to 19V range.
Could put a 12V Zenner across G-S to help limit the voltage. (will work with out the Zener)
I added a 4.7k Emitter resistor to the NPN transistor. When the switch is on the Emitter will have about 4V. 4V and 4.7k is about 1mA. This will translate to about 8V G-S on the MOSFET.

The key is the emitter resistor to keep the NPN from turning on full.

Either there is a Zener or a resistor divider network to keep the gate voltage within the maximum (20V Vgs usually) difference from the 12~48V supply (V+). I've seen zeners in parallel with the gate resistor, but you may still need a resistor at "X" to limit the current through the Zener. Other designs, I've seen just two resistors from V+ to the transistor as a voltage divider.
If your V+ is less than maximum Vgs (like a 12V supply), you can skip everything, and connect directly across "X".
Any zener has to be less than Vgs, and are usually 12V or 15V, enough to fully turn on the gate of the MOSFET. If the supply voltage actively varies between +12V to +48V, then a resistor divider will not work because there will be too much variation in the resultant gate voltage, you have to use a zener in that case.

Cheers Ron.. I have to cobble a circuit onto the existing board ( hence the To92 requirement..

I have seen zeners in where I put my big X so That's why I did that



Am I good to go??

In this application I can't see the advantage of a p-MOSFET over a PNP bipolar transistor. You will need a resistor to limit the base current, and another one between base and emitter to make sure the PNP isn't turned on by leakage current in the NPN transistor.

You could just use an opto-relay and the only other component you need is a current limiting resistor for the input of the relay.

Also, if you are driving an opto-relay, why does it have to be a high-side driver?

I would swap the 4.7k resistor and the NPN transistor .... have the NPN emitter to ground and the NPN collector to one side of the 4.7k and the other side of the 4.7k to the MOSFET Gate.

Beau Schwabe said:

I would swap the 4.7k resistor and the NPN transistor .... have the NPN emitter to ground and the NPN collector to one side of the 4.7k and the other side of the 4.7k to the MOSFET Gate.

Click to expand...

That's the way I'd do it as well, I don't see the reason for a resistor in the emitter, you're wanting a digital 'switch' not a linear analogue control to the gate of the FET.

I'm currently using small 8 pin SM IC's which contain a P gate and N gate MOSFET, they are both fairly high current (7A or so) and with a couple of resistors I'm doing basically the same as Ian wants, but I'm not sure the ones I'm using are high enough voltage? (I'm only switching 4-5V, and I think they are rated at 30V).

So we're going with this?

Like I said... A lot of circuits have the Zener in the place of the collector resistor... I have never studied fet gates.. They are a mystery to me.

Ian Rogers said:

So we're going with this?
View attachment 123419
Like I said... A lot of circuits have the Zener in the place of the collector resistor... I have never studied fet gates.. They are a mystery to me.

Click to expand...

I would suggest you probably don't need the zener as long as you pick the right values for the resistors - as a potential divider to give no more than 12V across the top resistor. By adding the zener it will work across a much wider voltage range though.

My SM circuit uses two 4K7 resistors, and the bottom transistor is an N channel FET with a 10K gate to source.

I'm with you about FET's though, I find them a mystery as well - but you hardly ever came across them in TV's or domestic electronics.

My circuit is powering a GSM modem module, so the VERY low Rds of the FET is very important.

Your gate drive current will be very low with a 12VDC supply (sub-milliamps) - is that enough for your Mosfet at your desired switching speed?
Also, the 4.7k current limiting resistor will need to be 0.5watt or more if you run at 48v supply. I think it will see 0.3 watts.

gophert said:

Your gate drive current will be very low with a 12VDC supply (sub-milliamps) - is that enough for your Mosfet at your desired switching speed?
Also, the 4.7k current limiting resistor will need to be 0.5watt or more if you run at 48v supply. I think it will see 0.3 watts.

Click to expand...

On or Off Speed!! Once twice a day... The opto switch (photoelectric sensor) consumes 40mA ish.... I think a bog standard mosfet should work okay... I'll test out the circuit on Monday..

Ian Rogers said:

On or Off Speed!! Once twice a day...

Click to expand...

then it's not a problem. if you were trying to switch some kHz PWM system, then much more of an issue.

Ian Rogers said:

So we're going with this?
View attachment 123419
Like I said... A lot of circuits have the Zener in the place of the collector resistor... I have never studied fet gates.. They are a mystery to me.

Click to expand...

With a 12V supply, Vgs will only be -2V.

Mike.
Whoops, read that as 47k not 4k7. Ignore above.

Ian Rogers said:

So we're going with this?
View attachment 123419
Like I said... A lot of circuits have the Zener in the place of the collector resistor... I have never studied fet gates.. They are a mystery to me.

Click to expand...

you need a 10v zener - standard value for mosfets.

I always thought gate voltage was key.. Gate current was negligible..

gophert said:

you need a 10v zener - standard value for mosfets.

Click to expand...

Great.. I have bucket loads of 10v zeners...

Ian Rogers said:

I always thought gate voltage was key.. Gate current was negligible..

Click to expand...

Correct - and as I said choosing the correct resistor values means no need for a zener.

Nigel Goodwin said:

Correct - and as I said choosing the correct resistor values means no need for a zener.

Click to expand...

"...and as I said", you say? Please read the main reason for this thread. Line 2 of post

Ian Rogers said:

...voltage can be 12v, 24v or 48v..

Click to expand...

Please let us know the values of resistors that can achieve the proper gate voltage (10V) and still accommodate the design constraints.

gophert said:

"...and as I said", you say? Please read the main reason for this thread. Line 2 of post



Please let us know the values of resistors that can achieve the proper gate voltage (10V) and still accommodate the design constraints.

Click to expand...

I refer you to post #9, where I addressed that possible issue. However, he's never made it clear if the requirement is for one unit to work across a wide range of voltages, or for individual units to be built for different fixed voltages. Both situations have been addressed here.

In post #5 another solution was suggested, which would be easier all round - use a bipolar PNP transistor, as there seems no reason whatsoever to use an FET?, and it would easily work across a much greater voltage range with no issues, or zeners. I'm presuming the optocoupler is fixed to ground?, otherwise a single NPN is all that would be needed.

Nigel Goodwin said:

However, he's never made it clear if the requirement is for one unit to work across a wide range of voltages, or for individual units to be built for different fixed voltages.

Click to expand...

I doubt the OP missed your posts. They were dismissed.

Basically!! Fork trucks are a) 12v B) 24v or C)48v... BUT!! 24v are the most common..

When I put 2k2 on top of a 10v Zener the tiny circuit (555 timer) gives 7 volts when using 24v and 4 volts at 12v which is good enough to drive the 555 this is set up as a monostable and has worked as expected for 12+ years.. When I use the 48v supply, there is a link, which when cut, ups the resistance to 4k4.. so I still get 7 volts on the 555... Maybe not the best way, but without sophisticated power supplies, works..

The old opto switch was stand alone... BUT!! needs 12 ~ 30v to work so I doubt we'll use this with a 48v system... I now have it integral to the board and it needs powering down when my system does.. Hence the need.. Any mosfet will have a lower on resistance than a bipolar...

Nigel's comments are now fixed in my head and have been absorbed… Calculating Zener current is something I can do.. But ! as I said.. I need to read up on fet gate requirements as I have seen too many false theories... One of my issues, I will now check, is upping the Zener resistor will screw up my RDon So I may have to "play" a bit to get what I want... As Nigel also said... Dropping the zener may be better (or) make it linkable for the higher voltages...