P-Mosfet & Mosfet Driver problem

Go To Last Post
11 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Dear Freaks,

I have a hardware problem. It's my 1st time I am using a medium voltage power mosfet. The mosfet is the IRFR9024:
http://www.datasheetcatalog.org/...

I need to drive the Mosfet using a micro (Vcc = 5Vdc) in switching mode with 25us period(~40KHz) and 8 bit resolution duty cycle.

The mosfet's sourse voltage is 55Vdc and as the datasheet says the Vgs must be in the range of +/-20Vdc.

Looking at Fig.3 I see that for -Vgs < 4V the -Id = 0 (aprox.) and for -Vgs > 10V the -Id = high (Low Rds).

So if I understand it well I need a Gate Voltage:

For given Vs = 55V
- For Low Rds (Mosfet saturation): 35V < Vg < 45V
- For High Rds (Mosfet cut off): 51 < Vg < 55V

So the Gate (Vg) pulse must be like this:

55V (cut off)   -------        -------        ---    
               |       |      |       |      |       
               |       |      |       |      |       
35V (satur) ___|       |______|       |______|       

1. Please confirm if I am saying it right.
2. What do you suggest me to do in order to drive the mosfet from the microcontroller pin?

For Vds = 55V(max) the Mosfet's Gate capacitance doesn't exceed the 1000pF and as I said before the Period is 25us and one step of duty cycle times ~10ns.

Thank you.

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Take a gander at figure 4. This should be of some help.
irf.com/technical-info/appnotes/an-940.pdf

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hello, Michael,

Looking at the Gate charge characteristics in figure 6, you will find a flat, horizontal portion which is called the Miller Plateau. During the time the device switches on or off, the Gate voltage is clamped to that plateau voltage and remains there until sufficient charge has been added to or removed from the gate capacitance in order for the device to switch.

A MOSFET responds instantaneously to changes in gate voltage and begins to conduct when the gate reaches the threshold voltage (V_gs). So, you will need a gate driver to provide sufficient drive current to quickly pass through the Miller Plateau region of the power-MOSFET’s switching transition.

This chart can help you estimate the Gate driving requirements, because it indicates the voltage of the plateau and the required charge for the device to switch.

The gate charge is:
Q_gate = I_gate * t_sw

The gate current is:
I_gate = (Vcc - V_plateau) / R_gate

Q_gate is the charge difference between the ends of the plateau. In your calculations mind also the additional Miller capacitance charge Q_gd.
t_sw is the Gate switching time, during the builting up or removing of the gate charge
R_gate is the inrush current limiting in-series Gate resistor and helps obtain the required t_sw
Vcc is the peak output voltage of the gate driver.

In your case, the plateau voltage is about 6.9V for |V_ds| = 48V, and the required gate-source charge Q_gs is 9.2 - 3.0 = 6.2nC. It is also obvious that a 5V micro is insufficient to drive this device directly.
You are halfway, now, because you will also have to calculate the additional gate-drain Miller charge Q_gd (mind its polarity because the Miller effect is what predominantly limits the device's switching speed)...

-George

I hope for nothing; I fear nothing; I am free. (Nikos Kazantzakis)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jaydhall,

Thanks for this pdf. It seems to be very helpfull.

George,

It is nice to comunicate with you again. Thnaks for your profesional analysis. Actually the mosfet I am using for my tests is not the final device. I have ordered some pcs from an IXYS part. I don't remember it's name now, but I will do the same calculations for that P-Mosfet.

I was thinking of using a npn (BC817-40), as a level shifter, with a voltage divider connected to it's collector. Something like the circuit shown at figure of the previous .pdf
www.irf.com/technical-info/appno...

I am just wondering about the power dicipation of the resistors. Because as I see, in order to charge and discharge the gate at 10ns (1 step), they might be some handrends of Ohms.

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you are concerned about switching speed (especially with high capacitive gates), you could use a photovoltaic gate driver and use it to switch a capacitor to the gate of your mosfet.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thank you for your kind words, Michael,

You are right worrying about the driver section, because that specific level shifter will be unable to provide more than a few mA of gate drive current, though more than 1A/2A peak gate charge/discharge currents are not unusual in high speed switching of high voltages, where the Miller effect becomes a very considerable factor.

In my opinion, it always pays to use N-channel power MOSFETs everywhere, due to their superior speed and Vds & Rds(on) characteristics, and their lower cost compared to their P-channel counterparts; especially for use as high-side power switches.
Just look at that 75V, 6.5mΩ, 120A/480A, 310W little guy: STB160N75!

For ultra fast gate driving (especially if there were multiple devices to be driven simultaneously) I would use gate drive transformers driven by push-pull emitter-followers; maybe with some additional asymmetrical gate biasing circuit at the secondary to speed up gate discharge. The galvanic isolation of the power section is a bonus in this case!
Otherwise, a cheap high-side gate driver chip (these ones generate the additional gate drive voltage using a charge pump), that can deliver more than 1A peak gate drive current, would be more than sufficient for almost any case.

-George

I hope for nothing; I fear nothing; I am free. (Nikos Kazantzakis)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I am agree with George but if you need to drive continuous it is not easy with a N type.
charge pumpe or transformer not work in "DC" drive.
Thierry

Thierry Pottier

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Guys,

The circuit I am trying to do, is a step down converter that can provide any voltage lower than the 55V input. The maximum load current will not exceed 250mA (max). Don't mind about the desired output voltage and the current limiting circuit.

The method of doing a step down converter is standard. It typical, has to do with the passive elements values calculation. But my problem is how to drive the P-Channel's gate using a micro's pin, while the Mosfet's -Vgs cannot be more than -20V.

In my mind, the only way of doing this using a N-Channel Mosfet is to use a transformer, instead of a simple coil, to transfer power to the step down circuit. Please correct me if my knowledge level is poor on this.

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

To use a transformer or a coil is a design decision depending not on the transistor type, how difficult is to drive that transistor, or if there is a microcontroller involved or not.

The most important requirement is if your load need to be isolated regarding the primary power source. If is the case, then you need to use a transformer, and the driving transistor is simpler. If such a feature is not needed, than you still have both choices.
Next, what is the ratio between output versus input voltage ? If the ratio is a very small one, than you should go for a transformer. If the ratio is high, lets say between 0.3 and 0.9, then is better to go with an inductance.
I think the topic is about using an inductance. The cheapest way to drive that transistor (a N channel one) is as someone mentioned, to use a small transformer driver. But again, you can not use it in dc, nor in a very high duty cycle way. In the primary side you can use pnp - npn emitter follower circuit. The 5V will be a challenge though, being quite small. If you go with a higher voltage for the emitter followers, will complicate the circuit.
The best solution, technically speaking, is to provide a separate power supply for your transistor driver circuit. Probably your 55V power supply, can afford an extra winding in this regard.
If you come up with more details, people here can give you more accurate advices.

And by the way, 20V gate-source is the maximum one, in practice is used something between 12 and 15V.

As a conclusion, forget using P channel on a step down converter, at 40kHz, 55V input.

George.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I was going to suggest some special high-side MOSFET drivers designed for static or lower frequency operation (for example, the LTC1155 and other ones from Maxim, ST, etc.), and the well known in the automotive industry Infineon PROFETs which are intelligent power switches widely been used as solid state power relays. Or, the more commonly used high-side MOSFET/IGBT gate drivers that use a bootstrap supply.

But, after realizing that the subject of discussion is a high voltage buck converter, I will insist on using an N-channel power MOSFET (with its drain at the +55V input and its source as the power output) instead of a P-channel one, driven by a small toroidal gate drive transformer (which is no bigger than a few mm in diameter), since gate drive of the power MOSFETs becomes challenging at higher voltage and frequency conditions due to the Miller Effect.

Additionally, there is plenty of information out there for anyone who wishes to design and construct their own gate drive transformers. Their design is similar to the forward-mode power transformers and their ratio is usually 1:1 or 1:2 (with single, dual or multiple secondaries).
There is also a couple of external components required, not always mentioned:
1. To prevent saturation by the presence of any DC component at the primary, the transformer should be driven by a coupling capacitor, whose value should be at least ten times the value of the total gate capacitance it is expected to drive (assuming a ratio of 1:1).
2. There will also be a need for two 15V back-to-back Zeners between gate and source, for transient protection.

The transformer-coupled gates are good for duty-cycles from 0..50% (which is the case in forward converters). At larger duty cycles there might not be enough voltage in the secondary to turn or keep the MOSFET on, because t_on gets longer in expense of the output voltage amplitude (the volt-seconds product is ct. for a transformer).
The answer to such problems are the DC restorer circuits.

Finally, though it took me a while to locate it on line, here is that famous paper, called: Design and Application Guide for High Speed MOSFET Gate Drive Circuits, by Laszlo Balog.

-George

I hope for nothing; I fear nothing; I am free. (Nikos Kazantzakis)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thank you all guys.

Michael.

User of:
IAR Embedded Workbench C/C++ Compiler
Altium Designer