Hi,
Assume amplifier.
"ID/VGS" dominates input nonllinearity.
"ID/VDS" dominates output nonllinearity.
This is the first time I hear the term input and output nonlinearity (I have taken many advanced circuits and RF courses which all covered nonlinearity).
Can you elaborate what you mean by that? IIP vs OIP?
Assuming an amplifier, I input an input signal, the amplifier distorts it (due to VGS and VDS dependency) and I get distorted output waveform. I can measure the distortion with HD, IIP, .... I do not see input/output difference.
For example, chapter-2 of the following refers only input nonlinearity.
**broken link removed**
I have this book - and read chapter 2. I do not see where in the chapter even 2D Taylor series would be described.
Example 2.4 only covers "ID/VGS" nonlinearity as everywhere else.
Can you give a more specific pointer?
Thanks, I'll play with this but I'm not sure yet how I'll fit the actual transistor into it.
But how would you design such a simple diff pair if linearity is a concern and the VDS nonlinearity is so dominant?
If VDS is low, output nonlinearity is large.
This is true for case where load is high or output amplitude is large.
You mean large signal VDS (bias point) - right?
Yes, I would think so.
But this wouldn't this imply that lower GM/ID (=higher VDsat) has better linearity?
I also set the CM to change the output common mode of the diffpair and sweeping IIP3 vs VOC for different GM/ID (or other parameters) has completely "arbitrary" shapes.
Can you understand what is imcomplete or lacking in the following ?
Yes. In the differential version the voltage swing is just half and hence the third order coefficient is 1/4 smaller.