Basic questions on RHP zero cancellation using nulling resistor approach

[diarmuid]

Hello All,

Im putting together a 2 stage opamp. Cc and Cl are close enough that the RHP zero is affecting my stability.
Im opting to use the nulling resistor method to transform the RHP zero into a LHP zero. Never done this
before so have the following questions:

1.) Method uses a FET for Rz and not an actual resistor to generate Rz >> 1/gm? Is this because gm will vary less than R?
Is it also because a FET will use less area?

2.) Say Im using a FET. Should it be NMOS or PMOS? At a guess I would say PMOS due to its lower mobility and hence higher
impedance. Is my thinking correct here?

3.) When using a resistor for Rz it doesnt matter which side of Cc I place the resistor. Is this the same when using a FET?
I have only seen the FET placed at the LHS of the Cc. Is this to de-couple it from output voltage swings or something
like that?

4.) This method gives me a LHP zero which is great for stabilty. It does this without degrading my internal slew rate. Whats
the catch? Are there any downsides to this method i.e. generates thermal noise in the heart of the OTA or something?

If you have any general tips also about this method that would be great aswell.

Thanks a bunch,

Diarmuid

 

[Muhammad Khurram Khalil]

Hi Diarmuid ,

1. A FET is used as a voltage controlled resistor(VCR) in a linear region (Active region), so yes u r rite the gm is more stable than R, as the FET is an active device whose output is controlled by external biasing.

2. You can equally use nMOS or pMOS. Generally, pMOS is preferred because of its high impedance as per u said...!!

3. You cannot use VCR after Compensation capacitor (Cc) because it will effect (reduce) the output bandwidth.

4. This method will not degrade your slew rate.
if u use a passive resistor, then, yes, the thermal noise will be a factor in the performance of the OTA, but this thermal noise become negligible if the VCR (FET) is used in the circuit.

i hope it will help...!!

 

[diarmuid]

Thanks Muhammad,

Good answers.

1. A FET is used as a voltage controlled resistor(VCR) in a linear region (Active region), so yes u r rite the gm is more stable than R, as the FET is an active device whose output is controlled by external biasing.

Yes I have seen for the process I am using the gm to be slightly less variable than resistance, becomes less variable the further in triode the FET is biased.

3. You cannot use VCR after Compensation capacitor (Cc) because it will effect (reduce) the output bandwidth.

This is interesting. Can you please elaborate. Placing it after Cc increases output impedance or capacitance?

Thanks again,

Diarmuid