On Mar 28, 3:36=A0pm, "Phil Allison" <philalli...@[EMAIL PROTECTED]
> wrote:
> "billcalley"
>
>
>
> > This has been really bugging me: =A0When I design and simulate an RF
> >amplifier(LNA) for a class project, it shows points of huge
> > instability (below the desired passband 'K' is *way* under 1 ).
>
> ** K is a measure of the =A0*potential* for an RFamplifierto oscillate.
>
> A K value less than one means that oscillation possible, with the right
> combination of input Z and load Z.
>
> =A0> But
>
> > when I add a bandpassfilterto the input of the unstable simulated
> >amplifier, the instability is "gone" (the K is now way above 1).
>
> =A0** The BPfiltereliminates a whole *range of values* for input Z from
th=
e
> game =A0- =A0so eliminates the main cause of possible oscillation.
>
> This link might be of interest:
>
> http://www.daycounter.com/Calculators/Complete-RF-Amplifier-Design-An...
>
> ...... =A0 Phil
Thanks Phil. That makes sense to me. So I guess if a small-signal
amplifier is simulated to be unstable out-of-band, and you add the
desired bandpass filter in front of it, and K now looks good (K<1
across a wide bandwidth), then the entire circuit will be
unconditionally stable when built. (I was just worried that the
filter was simply "masking" the instability of the circuit by
preventing the S-parameters that determine 'K' from being seen by the
simulator's input ****t, and therefore giving me a false K>1
indication).
Thanks again,
-Bill
|
