• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

how much voltage to gate a pFET: Vsource = 5v, Vgate = 0v/3v

Status
Not open for further replies.

mikewax

Member
so i have a pFET, IRF9Z34N, Vsource = 5v, and a DIO pin that goes 0v and 3v. the pFET has a threshold voltage of Vgs = -2 to -4v, and an input capacitance of 620 pF. will the pFET turn all the way off at 3V? and should i have a pullup between the source and gate? (the DIO is a 5v tolerant pin). i need to switch it on and off at 500kHz.

thanx, mike
 
Last edited:

crutschow

Well-Known Member
Most Helpful Member
Not it won't reliably turn off at 3V (which is at the Vgs 2V threshold voltage)..
If the source is at 5V then the gate should be switched between 5V (ON) and 0V (OFF).
But depending upon how much current you want the transistor to switch, 5V may not be enough.

Here's is the line from the data sheet that specifies the Vgs you need to fully turn on the IRF9Z34N MOSFET:
upload_2017-1-16_21-46-1.png

Since you don't have 10V available you may need to get a "logic-level" type MOSFET that has the on-resistance specified at a Vgs of 5V, if you want to switch amps of current.
 

Pommie

Well-Known Member
Most Helpful Member
To get that FET turned on properly you need the gate to be around -3V relative to ground. Also, at that frequency you need a hefty driver for the gate. I'd use something like the Microchip TC4431 high side driver. I'd also switch to a logic level mosfet.

Mike.
Edit, looking at farnel I couldn't find a suitable one but one that's close is SPD18P06
 
Last edited:

mikewax

Member
ok the TC4432 data sheet says "Input Supply Operating Range: 4.5V to 30V" that means that the 4432 can operate with VDD = 5?
and it says "High Output Voltage VOH = VDD - 0.8 " so that means that it will raise my 3v input to 4.2v output? so i'll get a logic level mosfet, a IPP45P03P4L-11, which has a VGS threshold of -1.5.
and if the IPP45 pfet has Vsource = 5, and Vgate = 4.2, then it will turn off fast enough for 500kHz?
 

Pommie

Well-Known Member
Most Helpful Member
That's a very impressive mosfet and should work fine assuming you're not switching 10s of amps. You will need some serious decoupling capacitors on that microchip part.

Mike.
 

Pommie

Well-Known Member
Most Helpful Member
Why are you using a chip that can handle 100 times the current? That is complete overkill. Choosing a smaller fet will simplify the circuit as the gate capacitance will be much lower and a lot easier to drive.

Mike.
 

mikewax

Member
several reasons:
1- i don't really know what i'm doing
2- electronic components are REALLY hard to come by. the ONLY way i can get them is by mail order from digikey, which means that it takes about ten days to pick out a different transistor, stick it in my circuit and see how it works, and if it doesn't i'm SOL and just have to find a way to make due with what i got on hand, which ain't much
3- finding transistors at digikey with the right criteria is a *****! it's hard to use SMDs, can't afford to order a dozen different ones, and i don't entirely know what i'm looking for. i'll try that si7461dp it looks like it might work better
i dunno if it's feasible to drive a pFET directly with a 3V IO pin, but until i get that TC4432 mosfet driver, i gotta try. I scrounged around and the best pFET i could find is a FQP7P06 which has a Cinput of 225p and a Vgsthreshold of -2 to -4.
guess i gotta go back to digikey and search through the haystack again.
i also gotta find a nFet for a booster circuit and see how to turn the thing on with only 3 volts.
anyway THANX for the info :)
 
Last edited:

spec

Well-Known Member
Most Helpful Member
Hi MW

500KHz is quite a high frequency so it would be best to go with a gate driver that has plenty of current and voltage capability.

One of the problems with low-gate-threshold MOSFETs is that they have a very high effective gate capacitance.

Don't forget to put a 10 Ohm, or so, gate stopper/shaping resistor as close to the MOSFET gate terminal as possible.

Is there any chance you could go for an NMOSFET rather than a PMOSFET, as that would simplify gate driving?

You could use an enormous range of MOSFETs with the appropriate gate driver.

spec
 

Pommie

Well-Known Member
Most Helpful Member
The OP needs to tell us what he is trying to do so that we can solve the problem rather than debug the solution he has come up with. Using a 45A mosfet to switch 450mA is just crazy.

Mike.
 

spec

Well-Known Member
Most Helpful Member
The OP needs to tell us what he is trying to do so that we can solve the problem rather than debug the solution he has come up.
Yes, the big picture would be a big help.:)

spec
 

mikewax

Member
yeah i didn't think of that. the whole circuit is:
circuit2.jpg
i did order a gate driver but i'm in a difficult situation. it takes over a week to get the part. now i've partly solved the problem by finding another power supply, so now the MCU has 3.6v. maybe that's enuf to get something goin. the FQP7P06 pFET is the best transistor i have on hand, so i'm gonna see how high a frequency i can get with that, and i'll look at digikey again and see if i can find something better. the real limitation is that i need through-hole components to make my prototype, and that really cuts down on my choices. but there's no point in designing the final product with SMDs until i get it working on a breadboard.
i'm not sure what you mean by a shaping resistor. Is that the R2 resistor? for rounding off the square input?
As for using an nFET, i thought of that before and figured that in the final design, when i'm using all SMD components, it would be better to have a high-side switch because a low-side switch will have to handle way more current.
i'm just gonna experiment and study the whole thing and wait till i get the gate driver.
 

Attachments

Last edited:

mikewax

Member
DAMN i screwed it up again! the current sense signal is measured at the capacitor. hold on....
OK, it charges the supercap and monitors the current going through the circuit and continually adjusts the PWM duty cycle so as to limit the average current draw from V2 to 450mA. this circuit already works fine when it's connected to an arduino board controlling the PWM, but that's a 5v board. now i'm using a discovery evaluation board http://www.st.com/content/st_com/en...tm32-mcu-discovery-kits/32f0308discovery.html and today i modded the board so now it runs on 3.6v.
pardon my muddled brain please. i'm under a lot o' stress.
 
Last edited:

mikewax

Member
well i just did another tedious search on digikey. there are over 2 dozen mosfets that are better than the FQP7P06, but they're all surface mount units. there are no through hole ones that are better than the FQP, so that's what i gotta use for now.
 

Pommie

Well-Known Member
Most Helpful Member
Assuming you're making your own PCBs then why not mount the (smd) mosfets on the underside?

Mike.
 

mikewax

Member
no i'm using a breadboard and i'm still working on the first prototype. i gotta get the whole circuit working on the breadboard before i start the PCB design. it's company policy. o_O
 

spec

Well-Known Member
Most Helpful Member
Hi MW,

i'm not sure what you mean by a shaping resistor. Is that the R2 resistor? for rounding off the square input?
The gate resistor performs a number of functions but, in essence what you say is correct. The gate resistor should be physically as close to the gate terminal of the MOSFET as possible, ideally connected directly to the gate terminal of the MOSFET.

On a separate subject, the PMOSFET gate drive. As has been said, a driver chip is really the only way to go at the frequency and current in your circuit but, at the moment, going by your schematic of post #14, you are driving the gate of the PMOSFET from a 3.3V MCU.

Can I suggest that, until you have time to get a dedicated driver chip, you make a driver with a high conductivity NBJT; a BC337 would be ideal.

(1) Connect the emitter of the NBJT to 0V
(2) Connect a 220 Ohm resistor from the collector of the NBJT to 5V
(3) Connect a 330 Ohm resistor to the base of the NBJT
(4) Connect the NBJT collector to the PMOSFET gate shaping resistor
(5) Connect the free end of the 330 Ohm resistor to your MCU output pin.

Of course, you can flip the home made driver and use a PBJT (BC327).

The edge when the home made driver BJT is conducting will be the faster and this may affect your choice of driver sex.

Note though, that the home made driver will invert the phase of the driving waveform.

Also, on your schematic of post#14, can I suggest that you make the 15 Ohm resistor 4.7K and the 10uF capacitor a 100nF disk ceramic capacitor. That will protect the current sense input to the MCU better and also improve filtering.

spec
 
Last edited:
Status
Not open for further replies.

Latest threads

EE World Online Articles

Loading
Top