Cascode. When to cascode and when not. Your experience? Examples?

[lineup]

I am working on a power amplifier.
Cascoding of VAS lower distortion considerably.

When do we cascode and when not?
Have you cascoded? VAS? Output? Examples?
Drawbacks of cascoding?

I'd like to know more of this.
Cascode ... is this all heaven to do?

🙂 Stefan

 

[lineup]

This is an example of Hiraga Le Monstre.
He has cascode of JFET in input.

🙂 Stefan

 

[EdGr]

Cascoding eliminates the Miller effect. It is often used in RF amplifiers. Cascoding also increases the gain.

The downside is that less voltage swing is available.
Ed

 

[MarcelvdG]

I am working on a power amplifier.
Cascoding of VAS lower distortion considerably.

I guess it's because the non-linear junction capacitance of the transistor is no longer in parallel with the Miller compensation capacitor, but that's just a guess.

When do we cascode and when not?

Whenever it has a big enough advantage to do so. For example, to improve high-frequency behaviour, increase gain, protect a low-voltage device against excessive voltages. A disadvantage it that it costs some headroom and some extra transistors.

 

[Alexandre]

Baxandall cascode works very well for cancelling Cob, and will also compensate for the base current of the cascoding device (Q58).
The small bjt can be operated at Vce=0.6V if you choose the right part.
(It's also called Hawksford cascode).

 

[Elvee]

Cascoding also eliminates most of the distortion caused by the Early effect. It is generally the dominant non-linear effect in a BJT VAS

 

[knutn]

But cascoding does not increase the gain.

 

[MarcelvdG]

A common emitter stage with a cascode on top has a higher voltage gain and transimpedance gain than one without, because of the reduction of the influence of the Early effect.

 

[cumbb]

The tonal character of a second part modulates the signal. It will sound a bit more cloudy, more technical, more artificial.

 

[lineup]

Baxandall cascode works very well for cancelling Cob, and will also compensate for the base current of the cascoding device (Q58).
The small bjt can be operated at Vce=0.6V if you choose the right part.
(It's also called Hawksford cascode).


View attachment 1161283

Thanks.

🙂 Stefan

 

[knutn]

A common emitter stage with a cascode on top has a higher voltage gain and transimpedance gain than one without, because of the reduction of the influence of the Early effect.

You are right, but only if the load is comparable to the output impedance of the JFET.

 

[MarcelvdG]

I don't understand where that reference to a JFET comes from, but indeed, for a stage with a load impedance that is small compared to the internal resistance of the amplifying device, cascoding won't increase the gain noticably. It could even reduce it by a very small amount.

 

[Mark Tillotson]

But cascoding does not increase the gain.

The high frequecy open-loop gain is increased - in a typical VAS at high frequency the closed-loop gain is set by the compensation capacitor but the more open-loop gain available the more linear the closed-loop response will be (assuming the capacitor is linear!)

 

[knutn]

I don't understand where that reference to a JFET comes from, but indeed, for a stage with a load impedance that is small compared to the internal resistance of the amplifying device, cascoding won't increase the gain noticably. It could even reduce it by a very small amount.

Sorry, I was referring to the schematic in post #3. And, yes, in a typical VAS, cascoding gives good results, both regarding bandwidth, distortion and gain.

 

[lineup]

Yes, in the circuit in post #3 there is no advantage using cascode for JFET.
This I have discovered by simulations.

JFET can need cascoding when used in high voltage stages.

🙂 Stefan

 

[huggygood]

Thank you to all of you, you allow us to progress again and again.

 

[lineup]

Nelson Pass has written a paper on Cascode in amplifiers:
https://www.passlabs.com/technical_article/cascode-amp-design/

He also has a good example figure.

🙂 Stefan

 

[indra1]

Yes, in the circuit in post #3 there is no advantage using cascode for JFET.

Removing the cascode will affect bandwith, phase shift, produce different harmonic profile and may increase gate leakage with all of it's related issues. It will behave and sound differently but should hopefully result in nothing fatal.

 

[lineup]

Okay. I see.
There is a reason Hiraga put cascode in the input.

🙂 Stefan

 

[Mike Gergen]

Erno Borbely was a big fan also.