Using a Mosfet for a series pass transistor.

[InductorbackEMF]

Hi first things first the reason i am not using a non Mosfet transistor is due to two things that being.

1. I want to keep the control circuit current draw as low as i can.

2. (Really i just like Mosfets)

So with than in mind i will say i have tested the circuit and it works fine,
 but it would be nice know any really bad reason not to use a Mosfet. (Say in the long run)

Thanks
Alex.

[bloguetronica]

Hi,

Are you using a MOSFET as the pass element on a regulated power supply? What configuration are you using? It is P-channel MOSFET or a N-channel one? The question seems a little vague. However, your point number two shows that you already made up your mind.

Kind regards, Samuel Lourenço

[IDEngineer]

Briefly... for an analog pass transistor application, you need an FET optimized for operation in its linear region. Most today aren't, since FET's are so commonly used in PWM circuits where they are either fully off or fully on and thus minimize their dissipation. The manufacturers chase the sales volumes. But they're available if you dig a bit.

[InductorbackEMF]

Hello,I am using the MOSFET to regulate the voltage of a unregulated supply to drop the voltage to about 12v from 25v and keep it there. Yes i know the huge amount of power loss in the fet part of the "Fun" in building it. Oh and the MOSFET is a Nchannel right now however this may be changed to a P channel later.

All in all sorry about the mess of text had no sleep
I will post a circuit when i wake up.

Ps The reason i am not using a voltage regulator IC is due to really high currents of about 6 to 8 amps

Thanks bloguetronica for your comment as well!

Alex.

[schmitt trigger]

There was a recent thread (perhaps not in this forum though) that explained in great detail exactly what IDE engineer mentions.
Essentially switching Mosfets are constructed using a MESH or similar, which means thousands of small Mosfets wired in parallel.

There was an Ixys app note which explained what this meant. Briefly and from memory, it said that the process is optimized to obtain the lowest RDSon and Miller plateau charge, but not matching individual transconductance.
Operating a transistor in a linear fashion may mean that some individual cells carry more current than others, resulting in hot spots and eventual failure.

Bottom line: if you are using a switching transistor in linear mode, you have to provide it with very generous SOA.

[MiDi]

Briefly... for an analog pass transistor application, you need an FET optimized for operation in its linear region. Most today aren't, since FET's are so commonly used in PWM circuits where they are either fully off or fully on and thus minimize their dissipation. The manufacturers chase the sales volumes. But they're available if you dig a bit.

This information can be found under SOA (Safe Operating Area), if there is a curve given for DC it is rated for linear mode, but there are very few modern Mosfets that are rated for this.

IXYS has a series of special linear mode optimized ones, but not cheap.
Mosfets for audio amplifiers should be suitable as well.

You have to cool ~100W and let the mosfet stay quite cool to be in SOA.
A tough challange if you have to deal with higher environment temperatures.

To get an impression look at the teardowns of electronic loads...

[David Hess]

Unfortunately the safe operating area (SOA) curves given in datasheets for power MOSFETs intended for switching applications often do not show the region of instability where the temperature coefficient of gate threshold voltage reverses.  They also often do not show the DC SOA at all.  But if you are only dropping 25 to 12 volts, then only the DC SOA will matter; the instability is a problem at higher drain voltages.

The big disadvantage of power MOSFETs compared to bipolar transistors is that power dissipation is limited by area and power MOSFETs simply cost more for the same area so you always pay a premium.

Also do not forget that you have to drive the total input capacitance.  The static current requirements are low (essentially zero) but power MOSFETs have considerable input capacitance which has to be driven under dynamic conditions.  This will become a larger problem if you have to use several in parallel to meet your power dissipation requirements.  Bipolar transistors are a little more forgiving in this regard.

[InductorbackEMF]

Hey guys sorry for not replying i was very tired hehe,once i get back to my "lab" i shall post the circuit.
Oh and i forgot about Mosfets being rather poor about being used for a pass element i may rethink my plan.
Thanks for all who replied!

Alex.

[David Hess]

At lower powers the difference in price between power bipolar and MOSFET transistors may not be significant.

Another advantage bipolar transistors have is their low and consistent base-emitter voltage versus the higher and more varied gate-source voltage of a MOSFET.

If you use a PNP bipolar or p-channel MOSFET for a high side regulator, then the voltage gain will vary with output current and load impedance which makes frequency compensation more difficult.  An NPN bipolar or n-channel MOSFET makes frequency compensation easier and more consistent.