Didier wrote:

Didier

Network analysers like the Agilent E5071C can be limited by the fixture
used, to around S22 (fixture) ~ -120dB@ 10MHz.

In principle one just feeds a signal into the output and measures the
resultant signal that appears at the input.
However when attempting to measure reverse isolation of 120dB or more
cable leakage, leakage from connectors (non screw mount connectors like
BNC can be quite leaky) need to be considered. Also if one isn't careful
with the grounding system this can limit the measured attenuation.
Leakage from the test source also needs to be considered.
The HP journal article on achieving 120dB attenuation with an attenuator
is informative on some of the issues involved.

The other major consideration is that unless one is making the
measurements within a shielded room the ambient RF signals may make such
measurements difficult.

Bruce

At 06:14 PM 3/8/2009, Didier wrote...

Now I'm curious. What would "forward isolation" be in an amplifier?

At 06:14 PM 3/8/2009, Didier wrote:

No, not really. How else would you measure it? Reverse isolation is
basically the reverse gain - S12 - of the amplifier. How does a VNA
measure S12? Essentially inject a  signal at port 2 and see how much
comes out of port 1.

If you want to get so picky as to determine S12 with a signal in the
forward direction then you have a problem. Because the forward gain -
S21 - is going to effectively swamp the signal going the other way.
So this is almost impossible to measure if the signals are at the
same frequency. In this case measure the S-parameters of the
amplifier in its 'normal ' configuration, then de-embed the
S-parameters of the device. Assuming the device is the only
non-linear element you are dealing with then from the S12 & S22
values you can also figure out how the device reacts in the reverse direction.

Then if you don't mind solving a large matrix you can figure out how
the device might react to passing a signal in both directions. The
reality is however that if the device is within its linear range,
which it is likely to be if you want to distribute a reference or
some such, the reverse behaviour of the well terminated amplifier
will approach the nominal S12 value. The difficult part is to
determine what happens when the amplifier is not nicely terminated!

John

Didier skrev:

If you are concerned about that issue, then you can apply a signal of a
different frequency on the input and then use your network analyser to
sweep the range. To be able to sweep the full range apply two different
frequencies so you can make partial sweeps of the other range.

Otherwise the buffer amplifier should see the 50 Ohm termination of the
network analysers as you run the buffer amplifier in reverse.

I suspect that isolational amplifiers often is analysed in sections.
If you have -120 dB reverse gain it can be a bit hard to measure.

Notice that a network analyser is automatically synchronous to the
source, which makes detection of modulated frequency more sensitive.

Cheers,
Magnus

I've been learning a little bit about this lately from my friend who is
an Agilent network analyzer apps engineer.

He tells me that "hot S12" tests can be done with modern network
analyzers.  The key is being able to offset the frequencies enough so
the front end of the analyzer doesn't get creamed.  For example, you put
+10dBm at 11 MHz on the input, and +10dBm at 10 MHz on the output.  Then
measure S12 using a narrow bandwidth in the analyzer.  Hopefully the DUT
frequency response is flat enough for that to be meaningful.

In any event, it's a challenging measurement.

John

Didier said the following on 03/08/2009 06:14 PM:

Probably does not mean a lot for an amplifier, but it would for a mixer :)

Generic terms need sometimes to be put in the proper context :)

In fact, it's a negative number, now that I think of it :)

Didier KO4BB

Hi Bruce,

Here are the results. First, the circuit draws 33.3 mA at 12 vols. And
it's gain is 1:1 as expected with emitter follower - checked that with
RF generator as an input just to make sure that I built it correct.

With HP10544A the output level is below +8 dBm:
http://www.amigazone.fi/files/gpsdo/544-10.png
(with 2 meters of RG174 and SMA-connectors between the circuit and SA)

Now the harmonics look like this:
http://www.amigazone.fi/files/gpsdo/544-11.png

And 2nd harmonic level:
http://www.amigazone.fi/files/gpsdo/544-12.png

So the spec told that 2nd harmonic should be more than 25 dB down, it is
just in spec...

So the final result looks that the 10544A is OK but it really has this
kind of output spectrum?

--
73s!
Esa
OH4KJU

Esa Heikkinen wrote:

Esa

The distortion is about what one would expect given the buffer circuitry
within the 10544A.
Since retrofitting an improved oscillator circuit isn't really an option
you will need to filter the output to reduce the harmonic content.
Try a bandpass filter driven by the buffer and terminated in 50 ohms.

Should you retrofit an improved oscillator circuit you may as well
replace the oven controller to eliminate the oven switching frequency
related sidebands.

Bruce

Esa Heikkinen wrote:

Esa

Try increasing C2 and C8 in the white emitter follower circuit schematic
to 100nF.

Bruce