# Is the main purpose of cascoding is to increase gain in FETs?

In my recent studies, I have come across the concept of cascoded amplifier using Common Source and Common Gate Configuration. What exactly is the reason of cascoding?

• Reduction of Miller effect is one such reason. Extending the operating voltage is another reason. Commented Oct 6, 2016 at 18:23
• The wiki page has a detailed list of advantages.
– user98663
Commented Oct 6, 2016 at 18:23
• @Wossname The Wikipedia has a one-line two-sentence list of advantages. That doesn't strike me as particularly detailed. It would be nice if we can expand on that. Commented Oct 6, 2016 at 18:26
• exactly. I looked up on wikipedia and the advantages of cascoded were still not clear. If somebody could list all the advantages. Commented Oct 6, 2016 at 18:28
• Good grief guys, read the entire wiki article will you, there are many sections that list the specific uses and reasons for using it over other methods.
– user98663
Commented Oct 7, 2016 at 6:45

It is mainly done to increase the output resistance and or reduce the Miller effect.

An example of increasing output resistance is the cascode current mirror.

An example of Miller effect mitigation is a normal cascode amplifier. Of course this also increases the output resistance, but a driving reason is to mitigate Miller effect.

The gain does not increase (although the bandwidth due to mitigated Miller effect does go up) because the first transistor has no voltage gain (voltage gain of 1) but provides a current gain. The next transistor provides the voltage gain. If you do the math it ends up (ignoring parasitics and making approximations) matching the gain of the regular common-emitter or common-source amplifier: $R_L/R_{ee}$

• Doesn't increasing output resistance means an increase in gain as Av=-gm * R. and hence the gain is raised? Commented Oct 6, 2016 at 18:29
• It increases the output resistance. For example a cascode current mirror. But the gain is not higher than a simple common emitter. It is still R{load}/R{ee} Commented Oct 6, 2016 at 18:51
• No because gm comes from the transistor below the cascode and R is the load resistor. It is however possible to use a higher value load resistor and then that would increase the gain. But only adding a cascode does not increase the (low frequency) gain. Commented Oct 6, 2016 at 18:54

In a common base circuit, the Miller capacitance no longer has any effect on the high frequency gain being rolled off by negative feedback. So the 2nd stage of the cascoded pair is the best you can get. The clever thing is that the emitter of this output transistor is held near enough at a constant voltage and, this means that the first transistor also has miller effects dramatically reduced.

I know of no other reason other than reduction of Miller effects on both transistors for using a cascode amplifier be it bipolar or FET based.

• I know another reason: isolation. If for example you'd want to prevent signals present at the output of the cascoded stage to couple back to the input of the amplifier. Cascoding can help. This is sometimes applied in amplifiers where the output current is fed to a mixer and you do not want the LO signal (at which that mixer is switching) to couple back to the input. Commented Oct 6, 2016 at 18:59
• I'm thinking that the isolation problem is in the main due to miller capacitance. Commented Oct 6, 2016 at 21:01

Cascodes are:

1. used as the gain elements in amp stages when the Miller effect is an issue.
2. used in current sources and as non-linear loads where the output resistance of a single BJT isn't enough.
3. able to nearly eliminate Early Effect in one of the BJTs by holding its collector at a fixed voltage. As smaller devices tend to have smaller Early voltages, so cascoding these smaller/faster devices provides an option for recovering some otherwise lost performance.

I have created a cascode vs. single transistor amplifier comparison circuit in this free on-line simulator:

https://www.systemvision.com/design/compare-cascode-vs-single-mosfet-amplifier

In the simulation results shown, you can see the increased bandwidth of the cascode amplifier circuit in both the time domain (with 5 MHz input to both circuits) and the AC or frequency domain. The transistor models are all identical and have reverse transfer capacitance Crss set to 10pF.

In the case of the single transistor amplifier, the Miller effect increases the effective input capacitance by a factor of (1.0 + gfs*Rload), or 16 times its nominal value. This 160pF capacitance, along with the small residual Ciss, combine with the 500 Ohm input source resistance to create an RC low-pass filter that rolls off the amplifier gain with a pole at just under 2 MHz.

The cascode circuit avoids this Miller effect because the lower transistor m2 has almost no voltage gain, so the effective input capacitance is just Ciss, 20pF in this example. You can see the correspondingly greater amplifier bandwidth.

Note that the circuit probes can be moved to look at other time or frequency domain signals, and component parameter values can be observed by double-clicking on any part. The circuit can also be copied and saved, so the user can make any desired changes and re-simulate to see the effects of those changes.