Maker Pro
Maker Pro

Cascading two 50 ohm RF blocks...

B

billcalley

Jan 1, 1970
0
Something has me really stumped: Since cascading two 50 ohm BP
filters can lead to big trouble*, wouldn't similar problems occur when
connecting a 50 ohm filter to the input of, let's say, a narrowband
LNA? Or to the output of a PA? Would the narrowband LNA (or PA)
disrupt the filter's response, since the filter will now no longer see
a wideband 50 ohm termination at one of its ports? I know I'm missing
something here, but I just can't figure out what...

Thanks,

Bill

* Since the BP filters themselves are expecting to see only a very
wideband 50 ohm termination at both in/out ports in order to function
as designed -- but when two filters are cascaded the passbands of these
50 ohm filters are the *only* frequencies that actually and truly see
50 ohms; the stopbands of each filter will now not see 50 ohms at all,
as these stopbands are now each looking at the very low return losses
inherent each other filter's reflective stopbands.
 
J

jgreimer

Jan 1, 1970
0
billcalley said:
Something has me really stumped: Since cascading two 50 ohm BP
filters can lead to big trouble*, wouldn't similar problems occur when
connecting a 50 ohm filter to the input of, let's say, a narrowband
LNA? Or to the output of a PA? Would the narrowband LNA (or PA)
disrupt the filter's response, since the filter will now no longer see
a wideband 50 ohm termination at one of its ports? I know I'm missing
something here, but I just can't figure out what...

Thanks,

Bill

* Since the BP filters themselves are expecting to see only a very
wideband 50 ohm termination at both in/out ports in order to function
as designed -- but when two filters are cascaded the passbands of these
50 ohm filters are the *only* frequencies that actually and truly see
50 ohms; the stopbands of each filter will now not see 50 ohms at all,
as these stopbands are now each looking at the very low return losses
inherent each other filter's reflective stopbands.

Yes, that's why it's a good idea to isolate each component (filter, LNA, PA)
with a resistive pad of 2 or 3 dB. We once had a problem with a 3 W
exiciter feeding a 40W PA. Both were in spec. individually but oscillated
out of band when connected together. The problem was that the input to the
PA, while 50 ohms in band, was essentially open out of band. It reflected
enough energy to cause the exicter to oscillate. The solution was to pad
the front end of the PA so the exciter would see 50 ohms at all frequencies
and then add enough gain to complensate for the loss of the pad.

The exception would be the output of the PA. Our PAs were required to be
stable into both a short and an open attached to the other side of a
stretchline at all phase angles. We never had a problem putting filters or
cavities on the outputs of the PAs.
 
Top