[NLRS] RF Signal generator desired

[Doug Reed]

I also would comment that the biggest problem is getting enough
attenuation to do a useful sensitivity test on a modern radio. The
current NBFM radios I own measure around 0.2uv for 20dB quieting, that
is about -120dBm into 50 ohms. Minimum discernable signal is less than
that and for SSB you can go even lower. That $80 sig-gen I linked to
*says* it goes to -130dBm. My experience measuring radios is why I
question if the sig-gen actually performs that well. Even if the
connector output goes that low, there may be much higher leakage from
the enclosure.

It is extremely difficult to get 140dB or more of shielding and
attenuation in a homebrew electronics package. You will not do it with
a simple metal Bud box and a few screws. You will also not do it with
BNC connectors and RG58 coax. Single shielded coax tends to leak and
has about 80dB isolation. BNC connectors have around 60-70dB leakage
when badly worn. Stick with threaded connectors and double shielded or
solid shield coax like UT141. And even then you want multiple
enclosures, one inside another, like Russian dolls.

The best shielded little RF project I ever made put a 20mw oscillator
inside a small Hammond die-cast box with some attenuation and an
external antenna connector. Those Hammond boxes are nice but had poor
shielding until I smoothed the box and lid to get nice flat mating
surfaces. I sanded the box top flat with sand paper on a flat surface.
I cleaned up the box lid surfaces with the long sharp corner edge of a
good quality metal file. I stopped when I got nice flat surfaces with
some mating surface area. And I had a nice little threaded power
filter feed-thru that was rated to 10GHz for DC power. I ended up with
no detectable leakage from the oscillator inside the box. But that is
still a long way from 140dB of isolation.

One option I could mention is to look for some old commercial
equipment that you can modify. Perhaps you can find an old tube-type
commercial sig-gen with a real adjustable attenuator you can
repurpose. Or buy a commercial 70dB step attenuator from Ebay and add
an electronic 30dB attenuator ahead of it to get half dB steps. The
trick will be to operate within the frequency range of the original
parts. Those ADF4351 oscillators go to 4GHz which is well beyond the
limits of most old equipment. For the low tech approach I suppose you
could just stack up a series of 20dB SMA attenuators but that isn't a
good long-term approach.

The easiest option is to buy a decent sig-gen that covers the range
you need. But I didn't see anything commercial that was "inexpensive"
on Ebay. The bottom line is what do you need it to do and how do you
use it? If you are checking antennas or tuning filters with a power
meter or O-scope, then you don't need low RF output. If you want to
test a receiver, then you need very weak signals and even moderately
strong signals for tuning a helical front end.....

73, Doug Reed, N0NAS.
-- 

The things they do not tell you are usually the clue to solving the problem.