FAQ: Lab #2 Narrowband LNA Design
- General Hints:
- Make the input signal path (from BNC to transistor)
as short as possible.
- In implementing bypass caps, you had better use ~2 capacitors
in parallel that the total width of the transmission line (50 Ohm line)
is the same as the width of total bypass caps such that you can avoid
the discontinuity in the line width which could cause parasitic inductance.
- Make sure that most of the data from the 2SC3302 data sheet are obtained
under 'certain' bias condition, i.e. they can change for different bias
conditions.
- You may be aware that the contribution from the input path to the niose figure
is significant. Since the bias path includes two thermal noise sources (R1, R2),
it might be helpful to use RFC (big inductor) between the emitter and R1/R2.
- Make sure to check the DC bias points before you make any other measurements
(s-parameters, noise figure) !! You can easily check these by measuring DC
voltages at several nodes using a volt-meter.
- Q:
What is the general consideration for choosing R1 and R2?
A:
(from HO #12 p.13)
As a final note on this bias method, the resistors have to be chosen small enough
so that the current flowing through them is large compared with variations
in transistor base current, if the bias point is to remain roughly insensitive
to base current. This requirement is somewhat at odds with the desire to keep
the resistors large to minimize their contribution to thermal noise.
Fortunately, it is usually not difficult to find an acceptable compromise,
and net degradations in noise figure can be kept to the level of tenths of
a dB or less.
- Q:
How should I choose the width (W) of the lambda/4 resonant line?
A:
Generally the width of the line controls the conductance and inductance of the line,
which also determines Q factor of the line.
If you are shooting for a high gain, you will want to maximize Q*L of the line,
and if you are aiming for a better slectivity, you might want to maximize Q.
In either case, the exact equation for the Q and L in terms of W is not known,
so you will need to find the optimum point by trying various line widths.
Or, if you come up with a nice equation, that'll be very nice !
- Q:
How should I find the tapping point of the lambda/4 line?
A:
Again, you will need to try several points and find out the place that gives you
the best matching condition (s22).