Hello, I read that for a source follower/buffer, the gate capacitance is not an issue. What would this mean ? Is it because it is unity gain ?
Thank you.
Thank you.
Yes, for the gate to source capacitance. Because (ideally) the source follows the gate, there is no voltage change between gate and source. The gate-source capacitance is still there but there is no charge/discharge because there is no voltage difference, so for the gate signal the gate-source capacitance is 'invisible'.
The story is different for the gate-drain capacitance of course.
Jan
The story is different for the gate-drain capacitance of course.
Jan
Thank you. As invisible, do you mean, not effective ? I mean if I use a 100k in front of the gate, will it affect frequency response, relative to this capacitance ?
Last edited:
I believe my description was clear. If you have any capacitance, but the voltage across that capacitance does not change, it has no effect on the circuit.Thank you. As invisible, do you mean, not effective ? I mean if I use a 100k in front of the gate, will it affect frequency response, relative to this capacitance ?
And don't forget the gate - drain capacitance! There you can have just the opposite. Say you have 1V at the gate, but you have 10V at the drain, in opposite phase (simple gain stage). Then, you have 11V across the gate-drain capacitance! So the gate-drain capacitance 'looks like' it is 11 x what it really is, because it requires 11x the charge/discharge current than what you'd expect from the 1V gate signal.
This is called the Miller capacitance, and it can wreck your freq. response.
Jan
Last edited:
And don't forget the gate - drain capacitance! There you can have just the opposite. Say you have 1V at the gate, but you have 10V at the drain, in opposite phase (simple gain stage). Then, you have 11V across the gate-drain capacitance!
That would apply in common-source circuits. However here its common-drain. The Cgd is hugely non-linear but its not magnified by circuit action in the common-drain connection.
Long ago Nelson wrote this article. In the section about disadvantages of mosfets he made the observation that when used as a follower the gate sees only a small amount of gate source voltage swing so the input capacitance may look about 5 to 10 x less than it's ciss. https://www.firstwatt.com/pdf/art_mos_citation.pdf
In a follower situation, the g-s capacitance is bootstrapped, and sees its effect reduced, but a deeper analysis shows that it may promote a negative resistance behaviour (=instability) in some circuits, some circumstances.
That is the case with all bootstrapped topologies, which can lead to peaking or even oscillation in extreme cases
That is the case with all bootstrapped topologies, which can lead to peaking or even oscillation in extreme cases
The reason for that is that the source impedance increases with frequency, and an impedance that increases with frequency looks uncanny like an inductance!
And an inductance with a capacitance can resonate thus giving rise to instability.
All electrical engineering ;-).
Jan
And an inductance with a capacitance can resonate thus giving rise to instability.
All electrical engineering ;-).
Jan
- Home
- Amplifiers
- Solid State
- Source follower Ciss