Support Doc Question

[Grant]

In Support doc question 1, what does it mean by "bode plot of open
loop amplifier loaded with output and feedback capacitance"? do I need
to take out the opamp and put c2 between V- and Vout and cl between
Vout and ground and then do the bode plot of Vout vs V-?

[Grant]

never mind...i got it

[reza]

Could you inform me though as to how you do it?

[mattholland]

if i understand correctly, you take your amp, switches caps and all,
disconnect the inverting input from everything so that you don't have
negative feedback, but leaving all other connections at that node
intact, then apply an ac input (common mode 0v as the prof. said in a
different post) and do a bode plot of at the node that is usually
connected to the inverting input.

the problem that i have is that my phase starts positive (why?),
something like +160, so i'm not quite sure how to interpret phase
margin. anybody know?

mh

[Grant]

On May 15, 10:54 am, Grant <grants...@gmail.com> wrote:

I got the same weird phase...and another thing I don't understand is
that my unity gain magnitude is 64k in the plot, but my gain 16
magnitude in the plot is 8k...why is that?

[prof....@gmail.com]

On May 15, 3:20 pm, mattholland <mattholl...@gmail.com> wrote:
> if i understand correctly, you take your amp, switches caps and all,
> disconnect the inverting input from everything so that you don't have
> negative feedback, but leaving all other connections at that node
> intact, then apply an ac input (common mode 0v as the prof. said in a
> different post) and do a bode plot of at the node that is usually
> connected to the inverting input.

That's exactly right.

>
> the problem that i have is that my phase starts positive (why?),
> something like +160, so i'm not quite sure how to interpret phase
> margin. anybody know?
>
> mh
>

Don't forget that phase wraps around, so -180 is the same as +180, and
+160 is the same as -200. So the inverting input (V-) starts at -180,
and then you are getting another 20 degrees from somewhere else. If
you go to lower frequency you'll probably see it go to exactly +/-180.
For stability, you want the inverting input to stay inverting. When
that inverting input stops looking like negative feedback (180 degrees
of phase) and starts looking like positive feedback (0 phase), that's
when you have to worry about stability. So if you're doing a bode
plot from the inverting input (which is the right thing to do) you
measure the phase margin as the difference from zero phase when the
magnitude of the gain is 1.

ksjp

[mattholland]

presumably your path to ground from vout looks something like this.
vout-rswitch-C2-C1-rswitch-gnd

when you increase your gain, you increase the size of C1. C2,C1 from a
capacitive voltage divider. which works oppsite the way a resistive
divider works, so it would make sense that you would see less gain
there as you increase C1. (disclaimer: as always i may in fact not
know what i'm talking about)

mh