Is there such a thing as limiting max GS voltage on a MOSFET with a Zener diode? I am contemplating protecting the MOSFET like this.
I think that works, but your P-channel is rated at 25V GS voltage, so it wouldn't need protection from a 12V supply.
My concern is this will be going in a car, which I understand has noise and spikes in the power supply.
Diagram 4 on this page.
Leo..
Thanks. I think I understand most of that except 2 things.
One is why is the 330 ohm current limiting resistor on the emitter and not on the collector side of the transistor, is there some advantage to that?
And the other is what is the name of the characteristic in the datasheet of the zener diode that describes the needed current to activate the zener diode, and how do I make calculations to ensure I'm meeting that value should I want to change the power supply voltage and resistor values?
Google "emitter follower".
The emitter is always a diode-drop (0.65volt) lower than it's base.
I you drive the base with a 5volt Arduino pin, then the emitter is 5 - 0.65 = 4.35volt.
Current through the emitter resistor is 4.35volt / 330ohm = ~13mA.
Base current = 13mA / Hfe = insignificant.
That makes current through the collector about the same as through the emitter. Which means that with a 1k collector resistor voltage across that resistor can never be more than 13volt.
The 10volt zener (only a safety net) clamps that to 10volt, but even without the zener source-gate voltage stays below 13volt. The only thing with this circuit is the ~4.5volt loss across the transistor, which lowers gate drive voltage to 12 - 4.5 = 7.5volt on a 12volt supply. Still ok for most fets.
Advantage is that supply (spikes) can rise to 50volt+ without any danger.
Leo..
I google emitter follower and am studying what I find, but able to understand only parts.. I am now comparing these diagrams of Nick Gammon's
Diagram 3
Diagram 4
Diagram 4 introduces the zener to clamp the voltage to 10v, and at the same time moves the current limiting resistor from the base to the emitter. Could we have added a resistor on the collector instead? Is it just a matter of saving on the number of components because it means I don't need a resistor on the base if I have a resistor on the emitter?
I would like to understand it well enough to know using the 2N7000 MOSFET instead of the NPN BJT would work and that I'm using suitable value resistors.
Edit: it seems Nick Gammon has disallowed external references to images on his website, I am including my own hand-drawn diagrams 3 and 4 here.
In diagram 3 the transistor is used as a switch.
Which means that voltage across the collector resistor (source/gate) is the same as the supply,
and that could peak above the safe source/gate voltage of the fet on a noisy car supply.
You could split that single resistor into a voltage divider, but then you have the problem of reduced gate drive and still have the spike problem that you must kill with a zener diode.
Diagram 4 is more clever, and you could replace the NPN with a 2N7000 (as you could with diagram 3).
I would then lower the 330ohm to 220ohm, because the fet has a higher Vgs(th) than Vbe of a bjt.
Both circuits could be made to work, but I would prefer #4.
Leo..
The resistor in the emitter makes it into a constant-current driver; the current into the Zener is now independent of the supply voltage.
OK I replaced the NPN with the 2N7000 and lowered the resistor value, and prototyped this. It appears to work well for my purposes.
Is this still called a constant-current driver, or is it called something else because it's based on an N-Channel MOSFET?
Yes.
Thank you folks for the conversation and education. Happy new year.
1k to ground not necessary on a FET gate (or more correctly, on an Arduino pin - if you use a series resistor it should go after the pulldown).
10k or 47 k would be adequate.
Funny, after not being here for awhile I forget that the latest post is always the end of the converstation and you have to choose between reading only the latest posts or going back
to the beginning and reading all the posts. The Subject really is a just a generalization.
Has anyone asks why you want to do this ?
Can you state your design criteria ?
OK, thanks, I will consider the value. I found some sort of pulldown necessary because during boot up the gate was floating and flickering my test LED that was attached to my output.
Nobody asked why yet. Design criteria is kind of formal way to put it. I'm just throwing it together, winging it.
I'm working on a relay and accessory controller for my truck. The input is going to be a button. Depending on patterns I press on the button, it will switch a variety of outputs. Some of the outputs will sink to ground to turn on something using an N-channel MOSFET, others will send power to turn on something using a P-channel MOSFET, and others will power relays using a MOSFET and those relays will do something. The sketch detects what I want to do by how many button presses I do and how long the button presses last, kind of like morse code. I've got the sketch done and now working on the hardware.
If you have any suggestions or ideas to help with that, please let me know.
And indeed that is the reason for the resistor, but since it is a high impedance point and the gate will be discharged at the moment of switch-on, you should not need a particularly low value to hold it down.
Design Criteria is a statement of the circuit objective formatted in a way that allows selection
of circuit details that meet the criteria. It starts by stating operating voltages and listing all inputs and outputs and all load specs. This would include stating whether a switching device is
to be a high side switch or s low side switch. Typically high side switches are P-channel and low side are N-channel but this is not a black snd white issue, as either function can be realized with either device, if done properly. Loads are listed by voltage and current (ie: 12V/2A). When necessary turn on and turn off times are specified, and rise time may even be a requirement.
If you're using button presses you need hardware or software debouncing. If the loads are generic enough it is only necessary to present one schematic example of each type. If you are posting to obtain a schematic that meets your criteria you must state explicitly that you
need help creating a schematic because we usually expect you to provide it. Personally, I prefer to use gate driver chips , (like the TC4427 )), rather than discrete drivers like you are proposing. Why the need for current limiting ? (given that I = V/R)
Are you expecting excessive current ? If you're worried about sudden voltage spikes that can be suppressed with decoupling caps or LPF on the drive signal input.
Thanks very much! The design criteria is sprinkled throughout this thread and could be better organized. This thread is mostly limited to find out if the idea I had for using the zener diode was something that is sometimes done, or if it was a wacky idea. I learned some additional fantastic things, too.
I'm going to build this project to operate generic accessories in the truck, because I'm not sure what all the things are that I want to do with it. When I pick devices to control that are beyond the load carrying specs of the MOSFETs FQP27P06 and FQP30N06, then I'm going to use the MOSFETs to switch relays. So I'm going to build the project and make an instruction sheet to include the limits and rules, and hook it up.
The debouncing is handled by the sketch, and I've already got that done. I am very pleased with the sketch.
Yes, that's right, I'm going to use the P-channel MOSFETs for high side switches and N-channel for low side. These MOSFETs are great, they have such low resistances they don't need heat sinks for what I want to do, and they are extremely inexpensive. I've used them to switch 4 amps and can't even get them to go above room temperature.
Thanks for the tip about the TC4427. I'm going to look into that. At first glance at the datasheet it doesn't look like it already includes the zener to protect the MOSFETs gate by limiting max GS, so that function will most likely still need to be done with discrete components. I'll look into it. For this project I will most likely proceed with the components I have on hand. But I may order up some TC4427s to have. It seems like a beautiful thing.
The TC4427 Vdd is the voltage you want applied to
the mosfet gate. The TC4427 takes a 5V input signal. The output signal is the GS applied (Vdd).
I've used them with the sane mosfets you're using
for H-Bridge drivers. I suppose you could add
zener diodes on the output.