• 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.

Power MOSFETs as Source Followers

Status
Not open for further replies.

BobW

Active Member
Some time ago, I needed a power output buffer for a circuit. Normally I would have used a bipolar emitter follower, but didn't have any available that could handle the required current. But I had a bunch of power MOSFETs and so I figured, what the heck, and tried one. Even though these are normally intended for on-off switching, it worked very well. The only disadvantage that I can think of is that VGS offset may be quite a bit higher than the VBE of a bipolar emitter follower. I never thought to measure it at the time. Next time I drag out that circuit, I'll remember to measure it.
Of course power dissipation is a concern, since the device is somewhere in it's linear region, and dissipating a lot more power than it would when it operates as a switch.
Can anyone think of any other problems that could be encountered using these as a source follower?
 

AnalogKid

Well-Known Member
Most Helpful Member
Power MOSFETs designed to be switches are different than those designed to be amplifiers. In particular, they have very different Safe Operating Areas (SOA). Make sure yours is operating "under the curve" at all times.

ak
 

JLNY

Active Member
Sure, there is no reason that a power MOSFET can't be used as a source follower in the linear region. MOSFETs are routinely used as the output transistors in class AB amplifiers and the like. Is this buffer just a high or low side follower or a complementary pair?

As you mention, the high Vth is a major issue, so without some kind of feedback from the buffer output there will be a larger offset than the Vbe of a BJT or even a Darlinton pair.

In addition, the higher Vgs will also limit the maximum output voltage swing of the buffer, as even if the gate voltage can swing all the way to the top or bottom rail, the maximum output at the source will max out at VDD-Vgs (or 0+Vgs for a P-type FET).

Power MOSFETs also tend to have very high input capacitance, often as much as several nanofarads, so that is a factor to consider if fast rise times or high frequencies are needed. This is typically listed in the datasheet.

Many power MOSFETs have an integrated body diode which should protect against back-EMF spikes, but you may also need dedicated schottky flyback diodes to suppress voltage spikes if you plan to drive highly inductive loads or in switching applications that will see repeated spikes.

Power dissipation also needs to be considered, but really this applies regardless of the transistor type being used.
 
Last edited:

BobW

Active Member
Is this buffer just a high or low side follower or a complementary pair?
The application is a simple vacuum tube curve tracer. I used it mainly for low power tubes, and the grid voltage driver didn't have to supply any grid current, since I always kept the grid supply negative with respect to the cathode. More recently, I needed to characterize some power tubes, and had to get data operating them Class A2 (positive grid current), and so I had to add a buffer that could source current to the grid. I used an NDP603 N channel MOSFET, with D connected to VCC, G connected to the original grid drive output, and S connected to –VCC via a 5.6k resistor, with S being the new grid drive output. So, the sink current is limited by the 5.6k resistor, but it's insignificant, while the source current is whatever the MOSFET can supply, which is more than high enough. The high VGS isn't a problem, because I have a zero adjust pot that is adjusted during setup, and I haven't needed to go anywhere near the positive rail. Actually, the high VGS is probably an advantage, since it gives a bit more of a bottom end which is often more useful.

The high gate capacitance shouldn't be an issue, since the frequencies involved are very low.
 

crutschow

Well-Known Member
Most Helpful Member
I know you don't need it for your application, but here's an interesting circuit for your bag of tricks that uses a couple of BJTs to reduce the input-output offset of a source-follower to typically only a few tens of millivolts:

upload_2017-5-5_20-46-56.pngupload_2017-5-5_20-50-28.png
 

tomizett

Active Member
Just to expand on what's been said by AK and JLNY in #1 and #2, the MOSFETs typically used in class AB amplifiers are different to those designed for switchmode applications. Generally, "lateral" FETs are used in linear applications, while "vertical" types are for switching. Even within vertical types, there seems to be wide variation in the SOA - Mosaic posted a link to an interesting paper recently if I remember rightly. I've also seen another article somewhere on the topic, I think, called "Making Linear Mode Work" - I'm sure you'll find it with google.
As AK said, though, just as long as you stay within the SOA you should be fine.

On the matter of Vgs, a couple of things to remember are that this can vary a lot between parts - far more than Vbe for a BJT - so your circuit is not as "repeatable" if you build it again. Also, I believe MOSFETs generally have a lower transconductance than BJTs, so you'll see worse inherent regulation of output voltage over current.
 
Status
Not open for further replies.

Latest threads

EE World Online Articles

Loading
Top