|
|
|
|
Balanced Amplifier - RF Cafe Forums
|
gregoryb
Post subject: Balanced Amplifier Posted: Mon Jul 17, 2006
6:45 am
Captain
Joined: Sun Mar 06, 2005 8:02 am
Posts:
6
Hi,
I have a question regarding Balanced Amplifiers with
Quadrature Hybrid couplers.
How does the output coupler help
regarding stability? Is it simply equivalent to a 3dB attenuator at
each amplifier's output?
Also, is there any protection against
reflected power from the load? As I see it, if both amplifier's output
are in quadrature, then all the reflected power from the load returns
towards the amplifiers and not to the quad-hybrid's termination. So,
the Balanced topology does not give protection against reflected power
from the load?
Thanks,
Gregory
Top
languer
Post subject: Posted: Mon Jul 17, 2006 4:46 pm
Captain
Joined: Sun May 14, 2006 8:53 pm
Posts: 17
Location: Earth
Stability:
You still have to account for this (the hybrid only
makes it easier since the operating impedance is built-in), although
there is some loss that the hybrid provides.
Reflected power:
The advantage of the configuration is on the use of the hybrids. On
the input the incident power goes to the amplifiers, the reflected power
goes to the isolated port. On the output the same thing happens, so
the amplifiers are always seeing a very nice load. Also the input and
output ports are always seeing a nice load (whatever is present at the
isolated port).
Read more here (or google it):
http://www.microwaves101.com/encycloped
... cedamp.cfm
http://www.minicircuits.com/appnote/Zrl.pdf
http://www.microwaves101.com/encycloped
... uplers.cfm
http://www.minicircuits.com/appnote/an30003.pdf
Top
gregoryb
Post subject: Posted: Tue Jul 18, 2006
6:30 am
Captain
Joined: Sun Mar 06, 2005 8:02 am
Posts:
6
Hi,
Can you give me a detailed explanation of why do the
amplifiers see a "nice" output load? According to my analysis it turns
out the the amplifiers see the same magnitude of reflection coefficient.
Thanks,
Gregory
Top
Jeanalmira
Post subject: Posted: Wed Jul 19, 2006 1:51 am
General
Joined: Tue Mar 15, 2005 11:43 pm
Posts: 65
Location:
Singapore
Hi there,
I'm currently also dealing with balanced
power amplifer. And it's true by using quadrature hybrid coupler, the
overall input and output return loss becomes very good, due to the explanation
above that the reflected power will be terminated by isolated port.
however, I have some doubts about this configuration. If, let's
say, single-ended configuration has bad output return loss, so that
the output power is not quite there, will it be improved by balanced
configuration (by using hybrid coupler)as it will improve the overall
output return loss?
My intuition said that it does not help
much, as the single-ended has bad output return loss, so the power will
not be transferred optimally. But I'm not sure about this.
Hope
kind souls out there enlighten me out of this doubt.
Thanks!
Best Regards,
Jean
Top
yendori
Post
subject: Posted: Wed Jul 19, 2006 3:32 am
General
Joined: Thu Sep 25, 2003 1:19 am
Posts: 50
Location: texarcana
90° couplers improve return loss of the balanced
amplifier by
reflection cancelation.
Imagine a signal entering the input(or
output).
It takes 2 paths: (one to each amplifier)
90° path (Thru
port)
180° path (coupled port)
If both amplifier are matched
they will have the
same reflection siginiture.
So the reflected
signals with arrive back at the input port(or output).:
180° path
(Thru port)
360°path (coupled port)
These signal cancel and
therefore improve return loss.
The couplers have a 3dB split.
Each amplifier will be driven 3dB less the input power.
But the
power adds at the ouput.
So a balanced amplifier with a 20dBm output
power
will have the linearity
of the amplifier driven at 17 dBm.
Rod K
Top
gregoryb
Post subject: Posted:
Wed Jul 19, 2006 4:17 am
Captain
Joined: Sun Mar 06,
2005 8:02 am
Posts: 6
Let me clarify what I am talking about.
I agree that the output and input impedance of the balanced amplifier
are well matched. However, it seems to me that each amplifier branch
*inside* the balanced amplifier (between the couplers) does not see
a nice load.
Here is the corrected link to my analysis:
http://www.esnips.com/web/rfmwsPhotos
(click on balanced.jpg)
I would be glad to hear any comments.
Thanks,
Gregory[/url]
Top
languer
Post
subject: Posted: Wed Jul 19, 2006 4:52 pm
Captain
Joined: Sun May 14, 2006 8:53 pm
Posts: 17
Location: Earth
Gregory,
I agree with you, the balanced amplifier does not
protect the single amp against load mismatches in the same degree it
does present a nice match to the output load. So it is left to the designer
to make sure the output load is well behaved. However, the hybrids do
"buffer" the single amps to some degree. Right now, simulation agrees
with this but I do not have the analysis to back it up. Long time ago
I did the analysis so I'll see if I can look up some notes.
Top
Jeanalmira
Post subject: Posted: Wed Jul 19, 2006
9:27 pm
General
Joined: Tue Mar 15, 2005 11:43 pm
Posts: 65
Location: Singapore
Hi all :
Thanks so much
for your advice and explanation.
So am I right to say that the output
return loss of balanced amplifier is mainly determined by the each of
single-ended output matching network? and not by the overall configuration
with quadrature coupler?
Once again, thanks so much for your
advice.
Have a nice day!
Warm Regards,
Jean
Top
gregoryb
Post subject: Posted: Thu Jul 20, 2006
3:46 am
Captain
Joined: Sun Mar 06, 2005 8:02 am
Posts:
6
Hi,
Quote:
However, the hybrids do "buffer" the single
amps to some degree
I think that the buffering the hybrids
do is only regarding the "secondary" reflection. i.e. after the reflected
waves from the load return through the hybrid and to the single amplifiers,
they are reflected again with the amplifier output reflection coefficient.
After this reflection they are in quadrature, but opposite (the branch
that had 90 degrees phase in relation to the other now has -90 degrees
in relation to it), so both waves will sum at the hybrid termination
and cancel each other at the load.
This has no effect, however,
if the single amplifiers are well-matched at their outputs, and in this
case it seems that the single amplifiers see the exact same output VSWR
as they would see if they were connected directly to the load (assuming
loss-less hybrid).
Quote:
So am I right to say that the output
return loss of balanced amplifier is mainly determined by the each of
single-ended output matching network? and not by the overall configuration
with quadrature coupler?
It *is* the configuration with the
quadrature hybrid that determins the output return loss of the whole
balanced amplifier, as long as both single-ended amplifiers reflect
the same impedance at their output (which they should).
Regards,
Gregory
Top
yendori
Post subject: Posted: Thu
Jul 20, 2006 4:08 am
General
Joined: Thu Sep 25,
2003 1:19 am
Posts: 50
Location: texarcana
After all that
hoopla the answer is yes. Yes, you are right.
It was a non-intuitive
question, which is better than a lot questions on the forum lately.
I've spent some time thinking about it while my wife told me a story
about something.
I guess the fact that power arrives at the inputs
of the amps unabaitated, which is good. Means that reflected power will
arrive at the outputs of the amplifier which is bad.
Rod
Top
Jeanalmira
Post subject: Posted: Thu Jul 20,
2006 4:24 am
General
Joined: Tue Mar 15, 2005 11:43
pm
Posts: 65
Location: Singapore
Thank you All!
Now
I have much clearer picture about this issue.
Once again, thanks
so much for your advice. It's appreciated!
Any discussion is
welcome.
Warm Regards,
Jean
Top
gregoryb
Post subject: Posted: Thu Jul 20, 2006 7:25 am
Captain
Joined: Sun Mar 06, 2005 8:02 am
Posts: 6
Now that we agree
that each single amplifier sees the same VSWR as without the hybrid
couplers, does anyone know why Balanced Amplifiers are considered more
stable than single-ended ones?
Is the balanced configuration,
for stability considerations, equivalent to an amplifier with a 3dB
attenuator at its output?
Thanks,
Gregory
Top
yendori
Post subject: Posted: Thu Jul 20, 2006 12:06 pm
General
Joined: Thu Sep 25, 2003 1:19 am
Posts:
50
Location: texarcana
I think that could be explained by the
fact that the balanced amplifier are driven with 3dB less power to achieve
the same output power of a single ended amplifier.
I've been
twice today though and I haven't even had my coffee.
Top
yendori
Post subject: Posted: Thu Jul 20, 2006 12:47 pm
General
Joined: Thu Sep 25, 2003 1:19 am
Posts:
50
Location: texarcana
Quote:
I've been twice today though
and I haven't even had my coffee.
meant to say "I've been
wrong twice today though and I haven't even had my coffee yet."
how ironic
Top
Joe
Post subject: balanced
amplifierPosted: Thu Jul 20, 2006 5:57 pm
Captain
Joined:
Fri Feb 17, 2006 6:25 pm
Posts: 6
Location: US
gregoryb wrote:
Now that we agree that each single amplifier sees the same VSWR as without
the hybrid couplers, does anyone know why Balanced Amplifiers are considered
more stable than single-ended ones?
Is the balanced configuration,
for stability considerations, equivalent to an amplifier with a 3dB
attenuator at its output?
Thanks,
Gregory
Yes,
the 3dB attenuator would be a good analogy for stability considerations.
An unconditionally stable amplifier will be stable for any real impedance.
If the amplifier is conditionally stable, and the region of instability
is near the outer edge of the smith chart, adding the 3 dB attenuator
or the hybrid coupler will make it impossible to present a real load
which falls into the region of instability.
Top
IR
Post subject: Posted: Fri Jul 21, 2006 5:17 am
Site Admin
Joined: Mon Jun 27, 2005 2:02 pm
Posts: 373
Location:
Germany
Hello,
Hybrid coupler have high isolation between
the output ports, therefore it adds to the isolation between the 2 amplifiers
and by this improves the overall stability.
Correction for yendori
- In the analysis you gave the 180° phase difference is caused by the
fact that the signal flows in the 90 deg port incidents (90°) and reflected
(Another 90°) to form total phase difference of 180° between the ports.
This makes the reflections cancellation.
_________________
Best regards,
- IR
Top
languer
Post subject:
Posted: Fri Jul 21, 2006 5:55 am
Captain
Joined:
Sun May 14, 2006 8:53 pm
Posts: 17
Location: Earth
Quote:
Now that we agree that each single amplifier sees the same VSWR as without
the hybrid couplers, does anyone know why Balanced Amplifiers are considered
more stable than single-ended ones?
Actually the single amps
do not see the same VSWR with and w/o hybrids. The worse return loss
the amp will see with the hybrid is twice the hybrid loss (i.e. somewhere
between 6-7dBs), not bad at all. The analogy to the attenuator is right
on, the travelling wave gets attenuated by this loss as well as the
reflected wave (thus twice the loss). As IR mentioned, the isolation
between quadrature ports is high so the main driver on reflections is
the single amp and the output load "buffered" by the hybrid (thus the
term "buffer" before).
In reality one of the more critical aspects
on the design is to get as good isolation between the quadrature paths
(i.e. between amplifiers) in the PCB. Usually the gain in the amps is
high and careful layout and shielding is necessary.
Top
Peter Raynald
Post subject: Posted: Fri Jul 21, 2006 11:28
am
Captain
Joined: Tue Sep 07, 2004 3:09 pm
Posts:
11
From my understanding, i think it is important to take the full
propagation path into consideration when figuring out the reflection
coefficients presented to the devices.
The test signal is coming
from both devices with 90 phase differences. Assuming full ouput reflection
at the output of the hybrid coupler, and scaling down the levels to
the output of one single stage.
The power from one stage (Po)
gets combined with the power of the other stage at the sum port (2Po)
then gets fully reflected. The power of the two stages is spitted back
and returned to each device (Po). Then this power is reflected to the
difference load.
Therefore 100% of the power comes back to each
device and the output VSWR presented to the devices is the same as the
hybrid output one, neglecting a few linear lossless transformations.
Top
yendori
Post subject: Posted: Fri Jul 21, 2006
12:44 pm
General
Joined: Thu Sep 25, 2003 1:19 am
Posts: 50
Location: texarcana
Quote:
Correction for yendori
- In the analysis you gave the 180° phase difference is caused by the
fact that the signal flows in the 90 deg port incidents (90°) and reflected
(Another 90°) to form total phase difference of 180° between the ports.
This makes the reflections cancellation.
Yes. from the
persspective of the port(s) looking into the coupler(s).
But
from the output of the ampliers perspective looking into the load, reflections
from the load will arrive back at the outputs with no cancellation.
The same way the source signal arrives at the inputs with no cancellation.
Posted 11/12/2012
|
|
|
|