Post subject: Oscillation
Unread postPosted: Wed Apr
06, 2005 5:44 am
Wed Jan 05, 2005 2:20 am
Is there any guys will share some documents about the Oscillation due
to circuit mismatch. When we did some phone board testing ,from PA output,
the VSWR around 5:1, I suspected there would be a high oscillation power
which might blow up the PA due to mismatch between PA output and Antenna.
I am not very clear with the Oscillation, need some material for explanation
Thank you! :idea:
Unread postPosted: Wed Apr 06, 2005 6:35 am
In general oscillation might occure when your
PA has a reflection coefficient greater than 1 at its output or input
(Gamma in, Gamma out) , this leads to a situation of return gain instead
of return loss.
You can measure your S11, S22 paramaeters (be
very careful when measuring S22 with Netowkr Analyzer, use pad and add
its impact on the retuen loss to the result), then calculate the Rollett
factor, if this factor is less than 1 you are bound to oscillation.
There are several ways to prevent oscillation:
1) To reduce
the gain of the amplifier, you can do that by using a shunt or series
resistor eithr in the input or output. If this is PA it is better to
put the resistor at the input to avoid the use of high power resistor.
2) Uniliterlization - to make your amplifier unilateral. Meaning to
decreas the value of S12. If your application is broadband use a resistor
between the base and collector of the transistor to reduce the reactance
of the junction capacitance.
3) Matching network at the output that
will make your PA become stable. However, there are trade-offs in the
case of a PA between your design parameters as linearity and efficiency
to stability, so you have to be careful with the design of such matching
network that will satisfy all these trade-offs.
Hope this helps,
postPosted: Thu Apr 07, 2005 12:56 am
High VSWR can
be predicted causing oscillation, due to the mismatch. Therefore, RF
PA design always try to avoid having VSWR >2.
I agree with
another fellow about improving the VSWR.
First, you may try putting
resistive load transistor, in the gate side of the device. It will help
to reduce the oscillation, and sometimes also improved VSWR. But some
trade-off maybe caused by putting higher resistor values, the gain and
bandwidth might be worse.
the other popular technique is negative
feedback technique. By using this technique it'll reduce the return
loss as well as S12, therefore it'll improve the PA stability.
I hope it helps :)