MOSFET understanding

DarrenY

I'm trying to understand and use MOSFETS for the first time in a switching application.
My current understanding is that P & N channel MOSFETS operate in the following manor......

P-Channel MOSFETS default behaviour is that it allows current to flow between drain and source. If voltage is applied to the gate, then the mosfet will stop current flow between drain and source.

N-Channel MOSFETS are the opposite in the it doesn't allow current to flow between drain and source UNTIL voltage is applied to the gate.


I am planning on using an LT1529 voltage regulator and controlling it via it's shutdown pin. When the shutdown pin is grounded, the regulator goes into standby, when open then the reglator wakes up. So, in this instance, I would be looking on using a P-Channel MOSFET to control the SHDN pin on the regulator. When voltage is NOT applied to the gate of the P-Channel Mosfet it allows current to flow between drain and source and effectively connects the SHDN pin to ground, thus turning off the output from the LT1529.
This will allow me to control the LT1529 from an output pin from the prop connected directly to the gate of the P-Channel MOSFET.
i.e Prop pin high == LT1529 SHDN pin high impedence(open), Prop pin low == LT1529 SHDN pin grounded.

I can use N-Channel MOSFETS to turn on/off things like GPS, Bluetooth modules etc.
Connecting a Prop output pin to a N-Channel MOSFET will be Prop pin high == MOSFET allows current flow (ON), Prop pin low == MOSFET stops current flow (OFF)

Are my descriptions above correct?
What do I need to consider when selecting the MOSFETS?
Is it OK to connect prop pins direct to the MOSFETS gates?

Mike Green

Not quite. First of all, look at the Wikipedia articles on MOSFETs (search internet for "wiki mosfet").

Most MOSFETs these days are enhancement mode. They conduct current when the gate has voltage on it and don't conduct when there's no voltage on the gate.

N-Channel MOSFETs have one polarity where the source is connected to ground, the drain is connected to the load and the other end of the load is connected to a (+) voltage source. When the gate is connected to a positive voltage, the MOSFET will begin to conduct.

P-Channel MOSFETs have the other polarity where the load is connected to a (-) voltage source (relative to ground) and, when the gate is connected to a negative voltage, the MOSFET will begin to conduct.

Because of the way MOSFETs are constructed, P-Channel devices tend to have lower gain and lower voltage ratings than the equivalent N-Channel devices.

Leon

Depending on the current you need to control, and the MOSFET you are using, you will probably need to use a driver between the Propeller output and the MOSFET gate. 3.3V might not be enough to turn the MOSFET fully on.

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Mike Green

Back to your original application ... controlling an LT1529. If you read the datasheet for the device, you'll notice that the shutdown pin threshold is a maximum of 2.8V which is well within the switching range of a Propeller output. Why not just connect a Propeller I/O pin directly to the shutdown pin of the regulator. When the Propeller is first started, the I/O pin will be an input and will float, so you'll need a pullup resistor if you want the regulator to power up on or a pulldown resistor if you want the regulator to power up in a shutdown state.

If you plan to power the Propeller from the regulator, you'll need a more complex circuit to hold the regulator in a shutdown state while the Propeller is powering down.

DarrenY

I'm using the N-Channel MOSFETS to turn on/off a GPS and bluetooth modeul - each take max around 50-70mA.
N-Channel seems to be what i'm looking for in this instance, on the ground side of each module.

I'm still confused on how I would control the SHDN pin on the LT1529. I need to toggle this pin from open, to grounded under control of a Prop output - perhaps a MOSFET is not what I should use in this instance?

Mike Green

Just connect it to a Propeller I/O pin. When the I/O pin is in input mode, it's essentially like an open circuit. When the OUTA bit for the I/O pin is set to zero and the DIRA bit is set to one, the I/O pin is essentially grounded. You don't need an MOSFET. To play it safe, I'd use a 1K resistor between the I/O pin and the LT1529's shutdown pin. That helps protect the I/O pin from misconnections and short circuits.

Mike Green

It's usually not the best idea to switch the ground lead because it increases the threshold for logic signals. You're better off using a PNP switching transistor in the (+) power lead so you can directly ground the (-) lead of the device. Connect the emitter to (+), the collector to the switched device. Connect a 470 Ohm resistor between the base and a Propeller I/O pin. Use a 10K resistor between the base and (+) to hold the transistor off until you're ready to turn it on. If the Propeller I/O pin is in input mode, the transistor will be off. If you make the I/O pin a low output, it will switch on.

DarrenY

OK, now I have it, thanks ALOT for your help Mike

Cluso99

Using Mosfets, there are 2 parameters you should be aware of. Firstly, the Ron resistance as this governs the amount of power dissipated in the Mosfet. May not be a problem for 75mA. This also governs the voltage drop in the mosfet. You can now get mosfets with 3-5mOhm (0.003 ohms). The second, usually to fully turn on an N-channel (most common) you need about 7V on the gate above the drain voltage. You will have to look at the graph of the mosfet to determine what the Ron will be for a given gate and drain voltage. IRH is a good source (manufacturer) of mosfets and specifications.

evanh

Cluso99 said...
... you need about 7V on the gate above the drain voltage.

Shouldn't that be Source rather than Drain? Otherwise there would be is a distinct lack of consistency of the somewhat fixed 4 volt (Lower for logic level mosfets) switching level of the Gate wrt the Source. If it was wrt the Drain then wouldn't there be a dependence on the initial Source-Drain voltage for the switching level?

I'll add that the gate capacitance is also an important parameter. It dictates the switching frequency. And also the amount of load reflection, this part is why a series resistor is a standard addition on the Gate.

It's quite cool to watch a mosfet gate voltage on a scope during both turn-on and turn-off to really see the capacitance in action. The line flattens right out at the switching level for the duration of the rail-to-rail transition of the Drain. Non-linear galore!

In addition, for turn-offs, there is the flyback and resultant ringing to deal with. This has a wonderfully bad influence on the Gate because of the capacitive coupling. Whatever is driving the mosfet better have some give in it (The series resistor) or damage and latchup conditions will ensue.


Evan

Leon

Cluso99 said...
Using Mosfets, there are 2 parameters you should be aware of. Firstly, the Ron resistance as this governs the amount of power dissipated in the Mosfet. May not be a problem for 75mA. This also governs the voltage drop in the mosfet. You can now get mosfets with 3-5mOhm (0.003 ohms). The second, usually to fully turn on an N-channel (most common) you need about 7V on the gate above the drain voltage.

Shouldn't that be the source voltage?

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Post Edited (Leon) : 8/10/2008 4:48:43 PM GMT

Cluso99

I think I am correct because I needed 21V on the gate to switch 14V on the drain (when off, this is the rail voltage thru' the load). Many high power mosfets have a max voltage of 20V. This meant that I could not fully turn on the mosfet quickly because I had to limit the gs voltage to 20V.

However, that said, it may not be necessary for logic levels and low current - you will need to check the datasheet.

evanh

Cluso99 said...
I think I am correct because I needed 21V on the gate to switch 14V on the drain (when off, this is the rail voltage thru' the load).

That's definitely wrong. Mosfets can switch hundreds of volts with the same 12 volt driver. Have you even got that transistor wired up right?


Evan

Cluso99

My apologies. I think I am confusing my attempt to use P-channel mosfets where I had to turn them off (in a bridge situation using 2 x N and 2 x P-channels for 14V @ 20A for forward and reversing a motor).

Most likely you will require a Vgs of more than 3v3 to turn on an N-channel mosfet adequately. I just looked up the IRF1503. See www.irh.com for datasheets.

evanh

Cluso99 said...
My apologies. I think I am confusing my attempt to use P-channel mosfets where I had to turn them off (in a bridge situation using 2 x N and 2 x P-channels for 14V @ 20A for forward and reversing a motor).

Cool, haven't been down that road myself but, yeah, I can see that being a way to counter the various forms of inductive bounce on both the power rails and the storage caps will cause unwanted voltages on the gates. The sort of thing that destroys otherwise okay looking circuits.


Evan

cheapskate

For switching a power rail to other modules, have you considered high-side switches.
They are more expensive than standard MOSFET (though cheap if you can get them on ebay)
I use the Micrel Mic2514 which can switch 3 to 13.5 v with a off i/p of 0 to 0.8v and a on i/p 0.8 to 2.0 so can be directly driven by the propeller. ith a 3v supply they typically can switch 0.5A
On top of this they have a current limiter and thermal shutdown which makes them almost fool proof

pmrobert

Where do IGBTs fit into the logic level switched, high current devices spectrum?

-Mike

evanh

pmrobert said...
Where do IGBTs fit into the logic level switched, high current devices spectrum?

Consider them as the modern darlington. They have the same double transistor arrangement, the difference being in that the darlington uses two BJTs where as the IGBT uses one MOSFET driving one BJT.

The target market of the IGBT is high voltage/high power starting from say 100 volts. Again, I don't know the physics but it goes something like BJTs, being diodic, can be fabricated to the same current capacity no matter what the voltage is. A MOSFET, on the other hand, has to trade current for voltage when it's fabricated.

At low voltage/high current, though, MOSFETs reign supreme in switching apps. They are way more power efficient than an IGBT.


Evan

Cluso99

@DarrenY - I am sorry, I was way off topic from what you initially asked. If you cannot connect directly to a prop pin then I suggest an npn transistor like the P2N2222A as used in the propplug.

MOSFETS: Overnight I was recalling my initial problem - it was with an N-channel mosfet in the upper lines of an H bridge driving a motor, and also in the positive power line. The use is onboard boats (such as mine) to control an autopilot motor. The boats use AGM batteries and a recharging regime of 14.4v bulk charge. The upper lines in the bridge (and the power input) are n-channel mosfets and in order to fully turn them on (20A) it was necessary to drive them as close to 20V as possible. The reason for this is that the load is actually in the source, not in the drain, and this was the source of my confusion.

evanh

Cluso99 said...
The upper lines in the bridge (and the power input) are n-channel mosfets and in order to fully turn them on (20A) it was necessary to drive them as close to 20V as possible. The reason for this is that the load is actually in the source, not in the drain, and this was the source of my confusion.

Ah, yup, common enough except the high side driver circuits are usually isolated. Each with it's own small power supply.

Just thinking about your circuit. I'm quite intrigued by the idea that you could reliably go to 24-28 volts, or even higher with a higher motor voltage, on the driver supply rail. I say this because of the way Vgs naturally clamps itself to within 4 volts until Vds is zero. After which the Source has risen all the way to the 14 volt high-side motor supply rail and the Gate will then freely rise to the 24 volt driver rail, giving you the recommended 10 volts for Vgs.


Evan

evanh

And vise-versa, when turning off the MOSFET, Vgs won't drop below that 4 volt point until Vds has stopped increasing. Think of the 4 volt crossing as a bit like a charge pump. The harder you kick it the swifter it kicks back. You can't beat it without destroying the MOSFET.


Evan