To avoid reverse leaking it is necessary to have at minimum 80dB
reverse isolation between LO output and mixer input.
This is usually realized? with at minimum 2 20dB amps, a 20dB
attenuator at the input, to decouple the LO output, followed by a
amp-isolator-amp stage. All very carefully shielded to each other
to have a good reverse isolation. Done this for my Tek 492 and
Anritsu MS710? SAs. Also as a test setup on the bench for my
portable Metronix SA. Because I didn't want to occupy the LO
output with my TG I decided to use a directional coupler with
-20dB side arm to get a probe of the LO signal. This is a proved
concept. My Tek 492 TG was linear within 2dB up to 1.7GHz without
leveler. Only by using carefully matched stages and setting the
compression points of amps and mixer to the right level. The
Metronix SA don't have a LO output and because I didn't want to
make the level calibration worse I decided not to put a DC in the
LO path, I have build the DC by putting a small wire over a trace
with the LO signal to have a linear probe of the LO. It took me a
day to get the wire followed by a 20dB decoupling amp formed to be
a within 2dB level range at full LO span. So now the Metronix SA
has a decoupled LO output.
On all tree SAs the TGs only working in corase span settings,
because to get it also working for lower span settings the TG
needs to have two mixers, two LO chains with filters and so on,
like the original Tek 492 TG was build. But I don't see any need
for this much more complex design. The TG is for me a tool to have
a source for cable and filter tests on the fly. For sharp filters
and so on a NanoVNA is the much better tool. And if you plan to
test frequency translation devices like doublers or mixers then a
SNA is the simplest and fastest solution, or don't sweep, use a
generator and a SA and do step by step scans.
I love this TinySA and also the Ultra. This are very amazing
tools with a very high value for the money. For me the next
version after the ultra can kill the major brands portable SAs in
a second. The advantages of the big player portables are
attenuators with wider frequency and power range, maybe mixers
with broader range and better harmonic driving capability.
But if I look to the big improvements between the TinySA and the
Ultra I can imagine that some thing like a Ultra 2 or Super Ultra
could be here in 2 years? with all this improvements build in.
And I have to say this. Erik, your work and your responsiveness
are amazing. You set a good example how to make things better for
other people. Thank you!
Please continue the improvement.
All the Best,
Isidro
Am 14.12.2022 um 20:58 schrieb Erik
Kaashoek:
toggle quoted message
Show quoted text
To reach the required resolution both the LO and the IF are
"tuned" where the LO does the course steps.
So you never can scan a small span wit a TG using a fixed IF
frequency source
I do not recommend the add a TG unless for large spans and at
an offset tuning to avoid reverse leakage. Main purpose is to
test frequency conversion devices as shown in the wiki
--
For more info on the tinySA go to
John Cunliffe W7ZQ