11

I have a question about MOSFET switching operation.

According to an article:

In order to operate a MOSFET as a switch, it must be operated in cut-off and linear (or triode) region.

According to another article:

MOSFET in saturation region is preferred to make it work as a switch.

I am very much confused about the operating region of MOSFET to be used as a switch.

Should I operate the MOSFET to "Turn ON" in a (linear/ohmic/triode) or saturation region?

Peter Mortensen
  • 1,676
  • 3
  • 17
  • 23
Rivon
  • 127
  • 1
  • 1
  • 8
  • 4
    The best way to use a BJT or a FET as a switch is to overdrive the control terminal, whatever people choose to call the operating region it ends up in. – Neil_UK May 09 '20 at 09:53
  • 2
    Yep. The word "saturation" when used for bipolar and field transistors mean almost exactly the opposite thing. People mistake them a lot. – fraxinus May 09 '20 at 21:57

5 Answers5

23

When your article says this (wrongly): -

MOSFET in saturation region is preferred to make it work as a switch.

It's because it's written by someone who thinks that the name of the equivalent section of the BJT's characteristic is 100% transferable to MOSFETs.

To clear this up: -

  • When a MOSFET is operated as an on-switch it works in the triode or ohmic region
  • When a MOSFET is operated as an off-switch it works in the cut-off region
  • When a MOSFET is operated as a controlled current device it works in the saturation region
  • "Saturation" refers to the channel being saturated

enter image description here

  • When a BJT is operated as a switch it works in the saturation region and cut-off regions
  • "Saturation" in the case of a BJT refers to the saturation of the base in that both PN or NP junctions are (somewhat) conducting

enter image description here

Should I operate the MOSFET to "Turn ON" in a (Linear/Ohmic/Triode) or Saturation region?

Answer: the linear/ohmic/triode region

Andy aka
  • 434,556
  • 28
  • 351
  • 777
  • 1
    +1 as the answer is better than me ;) – Jess May 09 '20 at 07:52
  • And now you are seeing papers and guides written by people who learned electronics from a CMOS perspective, and they are using the MOS-terminology to discribe the regions of a BJT, only making it more confusing. – Joren Vaes May 09 '20 at 08:04
  • 1
    All saturation regions are equal but some are more equal than others (George Orwell misquote). – Andy aka May 09 '20 at 08:07
  • 1
    I had never spotted that this terminology was used differently for BJTs and FETs. I would have made the same mistake of saying a switched on FET was saturated. I don't like it when words are context dependent, you're always correcting people who misuse them. Perhaps I'll just describe the low VCE or low VDS region as hard on, and take a risk on it being read in context. – Neil_UK May 09 '20 at 09:41
  • 2
    @Neil_UK I'm pretty sure the term "hard on" might raise the odd guffaw. – Andy aka May 09 '20 at 10:03
  • And you see "saturation" used in a third incompatible sense as the thermal current through a reverse-biased diode junction. It scales the forward current. This shows up in models of BJTs and MOSFETs because both contain diodes. – John Doty May 09 '20 at 16:34
8

If you want to use a MOSFET as a switch, you probably want to have a low VDS, as an ideal switch has no voltage drop across terminals, but it can have an infinite current through it.

So according to the characteristic curves of a MOSFET, you have to be operating in the linear region. As you can see, if you need 10 A (supposing the scale is in ampere), the VDS will be lower if you increase VGS.

Enter image description here

Your confusion may come from the fact that for a bipolar transitor, the name of the different regions is not the same:

Enter image description here

So if you use a bipolar transistor as a switch, you should use it in the saturation region.

Peter Mortensen
  • 1,676
  • 3
  • 17
  • 23
Jess
  • 2,694
  • 1
  • 18
  • 46
1

A MOSFET transistor can be used as an on switch in both triode and saturation regions, but it gives us different advantages.

In saturation, a higher current can be obtained, but in triode, because of its lower resistance, lower losses can be achieved.

Normally in the digital circuit design, the triode region is more common.

Peter Mortensen
  • 1,676
  • 3
  • 17
  • 23
arsalan ghasemian
  • 112
  • 1
  • 7
  • You cannot get much more drain current in saturation. You may get a thermal runaway instead. It is BJT that can be used as a switch both saturated or not with different advantages (power loss vs switch time). – fraxinus May 09 '20 at 22:00
  • Dear Fraxinus, as I said normally triode region is being used but in case you wanna use MOSFET switch to drive something sometimes saturation region is being used. – arsalan ghasemian May 10 '20 at 04:21
-1

Be it mosfet, fet or bipolar, for switching purpose, it is only in one definition: it is in saturation or cut off. Saturation is min voltage across device, and Max current. In cut off or open state, all the voltage across the device, but negligible current. In this switching mode, in both the above cases, the dissipation in the device is minimum, and Max power is controlled. To ensure saturation, you may have to overdrive, within limits, and similar under-drive, or even negatively drive the devices to ensure full cutoff. Go by the datasheets and don't overdo it. The only time these devices are used in the active region, or linier region, is in amplifiers (not even all).

Bringing "triode mode" in to this is meaningless. NOT to mention, who said triodes don't have cutoff and saturation? It is advantageous to use 'English' sometimes when you want to describe the physical picture, but there are no cmos language, triode language etc. Common sense does not depend on language. Engineering does not preclude common sense.

Vishy Nathan
  • 1
-4

Given a FET differential pair, you can use that differential pair as a fine current-steering switch, yet both transistors will be in FET saturation mode.

Before sub-micron FET transistors came along, the current-steering method was used in bipolar technologies (emitter coupled logic, or current-steering DACS). Avoiding the charge storage in bipolar transistors, if operated in forward-biased base-collector junction, was the mindset.

============================

If you are switching currents, the FET differential pair may work just fine.

I once communicated across boundaries on a chip, wishing extreme isolation so as to achieve excellent spurious performance, using small differential currents (I stole the idea from someone else).

Peter Mortensen
  • 1,676
  • 3
  • 17
  • 23
analogsystemsrf
  • 33,703
  • 2
  • 18
  • 46
  • Actually, I want use Mosfet as a switch, not as an amplifier. Please explain me the best operation region to be used to turn the Msofet " ON" . – Rivon May 09 '20 at 07:16
  • 1
    What is "DACS" (in this context)? Not [DACs](https://en.wikipedia.org/wiki/Digital-to-analog_converter)? – Peter Mortensen May 09 '20 at 20:01