Ultra low Vth MOSFET

[cyberfish]

Just wondering if anyone know of a MOSFET (N-channel) that can switch ~10A with a Vgs of 3.3V.

Switching speed is not important (kHz range).

Or will I have to use a BJT to drive it?

(just trying to save a transistor)

Thanks

PS. this is for PWM-ing a motor for a quadrocopter propeller.

[cyberfish]

And if I need a BJT, something like this will work?

[cyberfish]

Actually, I don't think I will need the second transistor, since my microcontroller (STM32) allows the output pin to be configured as open-drain.

I can just add a pull-up to supply voltage.

[Soeren]

Hi,

You're better off driving a MOSFET gate ON hard (i.e. not with a pull up/down).
You should however not rely on MOSFETs with only 3.3V to spare - go BjT.
Even better... Up the voltage. 10A @ 3.3V is gonna get you in trouble, as every little resistance in wires and contacts will seem huge.

Edit: The V_th isn't the limiting factor, it only tells you the voltage where it won't stay open any longer (after being turned on).

[dunk]

Actually, I don't think I will need the second transistor, since my microcontroller (STM32) allows the output pin to be configured as open-drain.

I can just add a pull-up to supply voltage.

are you sure?
most microcontrollers maximum voltage on an input pin very close to VCC.
check the electrical specification for IO pins in the datasheet.

for example, i happen to have the ATMega2561 datasheet open just now:

Voltage on any Pin except RESET
with respect to Ground ................................-0.5V to VCC+0.5V

Voltage on RESET with respect to Ground......-0.5V to +13.0V

so apart from the RESET pin, 0.5V above VCC is the limit.


rather than use a transistor to drive a MOSFET there are 2 other options:
1. use a dedicated MOSFET driver IC. (these are essentially just transistors but optimised for driving MOSFETs so you don't need to do as much research.)

2. use a MOSFET with a low trigger voltage. they do exist.
search for "logic level MOSFET". you are looking for one with a low Vds.
look on the datasheet for a "Saturation Characteristics" graph that will give you an idea what trigger voltages are acceptable.
Digikey actually has the filter "Logic Level Gate" on it's MOSFET page. http://search.digikey.com/scripts/DkSearch/dksus.dll?Cat=1376381&k=mosfet

here's one i've used in the past designed for TTL level logic (ie. 5V): http://www.fairchildsemi.com/ds/RF/RFP30N06LE.pdf
i'm sure if you read enough MOSFET datasheets you will find one for 3.3V as well.


dunk.

[dunk]

You should however not rely on MOSFETs with only 3.3V to spare - go BjT.

interesting.
am i reading you correctly that you would recommend using a BJT transistor instead of a MOSFET when your trigger voltage is 3.3V?
or are you suggesting using the transistor to trigger the MOSFET?

it was my understanding many logic level MOSFETs had driver circuitry built in to buffer a (lazy) designer from these issues.


dunk.

[Soeren]

Hi,

am i reading you correctly that you would recommend using a BJT transistor instead of a MOSFET when your trigger voltage is 3.3V?
or are you suggesting using the transistor to trigger the MOSFET?

If the supply is 3.3V as mentioned, I doubt that you'll be able to find a MOSFET that would work as well as a BjT (but hey, new stuff arrives all the time, so it's just based on what devices I know of).
If there is eg. a 12V line for the MOSFET, then a BjT would be the way to trigger the MOSFET.

it was my understanding many logic level MOSFETs had driver circuitry built in to buffer a (lazy) designer from these issues.

I haven't read the datasheets of all existing logic level MOSFETs of course, so perhaps you could point to one that would work? But even IR's newest Logic Level devices like IRLS3034-7PPBF (240A!) won't work at its best, even if it might be semi-useable at a measly 10A, but is it worth the cost (and below par function)?

The problem is, that the Miller plateau is usually at around 3.5V and upwards in Logic Level devices and if you cannot supply at least that voltage (plus a bit), you won't get anywhere near the R_ds_on or current handling ability that the device are spec'd at.

[dunk]

yea, i just had a quick look at some datasheets.
it appears my assumption that there are 3.3V MOSFETs out there defies physics....

o well.
there's definitely loads of TTL ones out there.


Cyberfish,
the power draw of your microcontroller will be very low compared to your motors.
have you considered just using a 5V microcontroller?


dunk.

[cyberfish]

You're better off driving a MOSFET gate ON hard (i.e. not with a pull up/down).
You should however not rely on MOSFETs with only 3.3V to spare - go BjT.
Even better... Up the voltage. 10A @ 3.3V is gonna get you in trouble, as every little resistance in wires and contacts will seem huge.

Sorry I didn't make it clear. The supply voltage is 7.4V (2 cells lithium polymer). Only the microcontroller, sensors, and radio run at 3.3V.

are you sure?
most microcontrollers maximum voltage on an input pin very close to VCC.
check the electrical specification for IO pins in the datasheet.

That is a very good point . Not sure how I totally missed that. Guess I'll have to go with BJT (or driver chips).

And yeah I am aware there are many 5V MOSFETs, but didn't find any 3.3V.

the power draw of your microcontroller will be very low compared to your motors.
have you considered just using a 5V microcontroller?

I would rather not do that because all my sensors and radio run on 3.3V and most aren't 5V tolerant.

[Soeren]

Sorry I didn't make it clear. The supply voltage is 7.4V (2 cells lithium polymer). Only the microcontroller, sensors, and radio run at 3.3V.

OK, then this will work (either of them, your choice, but not both at the same time of course):

Full resolution here.

And to answer the next possible questions... No, you cannot get rid of components here if you want it to work as it should - except for R2 and R8 if you are sure the pin will always be driven (either high or low and never go Hi-Z when the circuit is powered).

And you should still use a Logic Level device, to make sure it's driven as hard as possible (regular MOSFETs are spec'd at Vg=10V).

[cyberfish]

Thanks.

I'm trying to understand the circuit.

What's the purpose of D2? And isn't it intrinsic in Q3? Same for D4.

I'm assuming the 2 stage driver is to make switching faster?

If I only do, say, 25kHz (to clear the audible range), does it really matter?

What's the purpose of R5 and R10? Can't they be shorted? Why only pull them 1/3 way?

Sorry I've never really used MOSFETs before, so all this is first time analysis.

[Soeren]

Hi,

What's the purpose of D2? And isn't it intrinsic in Q3? Same for D4.

Protection. Don't rely on the intrinsic diode of the MOSFET for that.
If D2 dies, it shunts and saves the life of the more expensive MOSFET. But do use a power Schottky or a similarly fast diode.

I'm assuming the 2 stage driver is to make switching faster?

No, it's to get the right polarity, keeping the 3.3V side 3.3V.

If I only do, say, 25kHz (to clear the audible range), does it really matter?

I'll repeat...
"And to answer the next possible questions... No, you cannot get rid of components here if you want it to work as it should"

What's the purpose of R5 and R10? Can't they be shorted? Why only pull them 1/3 way?

Current limit and protection against oscillations.
No, they cannot be shorted... Remember... "No, you cannot get rid of components here"
They should be rearranged for best effect though (that's what happen when designing electronics stone cold sober ) and this is done now.

Full Resolution
(Refresh your cache if it looks odd).

[cyberfish]

Oh yeah I'm sure they have their uses, but I don't want to use circuits I don't understand in my system, so I'm trying to understand everything (or as much as possible anyways).

If D2 dies, it shunts and saves the life of the more expensive MOSFET. But do use a power Schottky or a similarly fast diode.

I think in that case I'm more worried about the rest of my helicopter falling out of the sky...

That aside, you mean D1? But how do we know D2 will breakdown before Q3?

No, it's to get the right polarity, keeping the 3.3V side 3.3V.

Oh yeah of course, not sure how I missed that.

Current limit and protection against oscillations.

Since I've not taken an amplifier course, my knowledge of this oscillation issue is very limited, so can you briefly explain how shorting R5 or R10 will cause oscillation?

I thought there has to be positive feedback for something to oscillate? Where is the feedback?

Thanks