4/29 I am confused about reading frequency response from a bode plot
when the curve starts at a frequency other than zero. Any pointer
to reference on the net or book or kind enough to explain here?
\_ Well you should be since measuring zero frequency is physically
impossible.
\_ DC is zero frequency. It's not impossible to measure. The problem
is that Bode plots are log/log plots. What's log(0)?
\_ DC is zero frequency. It's not impossible to measure. The
problem is that Bode plots are log/log plots. What's log(0)?
It's sort of hard to have a scale that goes all the way to
-infinity, isn't it? So you just stop at some other frequency.
So I remember anyway. I apologize if I'm wrong. --PeterM
\_ To measure DC you must have an infinite period.
\_ are you a lawyer? Yes, you're technically right.
In practice, you're an idiot.
\_ Measuring DC is not the problem (you can measure the
voltage of a battery, right?), it's displaying it
that's a problem since log(0) = -infinity. So they
just start it at some small value, for example 0.01Hz.
\_ Ooops, sorry for the confusion. I was really confused. It
should be read as "starts as zero degree..." Help?
\_ Do you mean a phase reponse plot? Plotting phase in degrees
starting from 0 and going to 360 is eminently sensible. How
*else* would you do it? I suppose you could start with 45
and go to 405.
\_ YES! So how does one interpret the phase margin when the
phase response start from 45 up to 405? I know how to
identify the phase margin when starting from 0 dropping
down to -180.
\_ You're talking about a phase out vs. log(frequency in)
plot, right? The phase is always with respect to input.
\_ Yes, for example, starting from 0 with unity gain
at -135. I have 45 degree phase margin. What if I
start from -90 with unity gain at -200? What is
the phase margin in this case? What about starting
from 100 with unity gain at 0? What is the phase
margin in this case? |