[TowerTalk] receiver protection

[Jim Lux]

At 04:36 PM 3/28/2005, Tom Rauch wrote:

>I just was measuring something like that today with a 50 ohm
>source and load and  two tones near 5 MHz.
>
>1N916's TOI -12dB and started to show IM at 0dBm each tone
>(about 1/4 volt !)
>1N4005 TOI -14.6dB and started to show IM at -3dBm
>UF4004 TOI -15dB and started to show IM at -3.5dBm
>1N5408 TOI -18.8dB and started to show IM at -5dBm
>
>Stacking the diode pairs increased things about 6dB, as
>expected.

Interestingly, there's a plot on p156 of the latest (March 2005) Microwave 
journal showing large signal S21 of a Schottky diode at 1GHz.  At -1V, it's 
pretty steady at
-27dB from -20 to +5dBm.  At 0V, it's about -25dB isolation until you get 
to around -5dBm when it starts to climb. Unfortunately, they don't give a 
circuit diagram of the test setup, so it's not real clear how the diode is 
wired.

>  I
> > figure that much voltage at a 50-ohm receiver input
>terminal to be
> > something like S9+80dB.
>
>As if S-meters mean anything. Next thing someone will claim
>they are all 6dB an S unit.  :-)
>
>Actually IM starts at around 1/4 volt RMS, or about .3V
>peak, with the BEST diode I tested (so far). Unless you're
>running a very high impedance system or very high
>frequencies, non-linear junction capacitance isn't an issue.
>I'd back-bias the diodes. Remember the clamp voltage source
>has to have VERY low impedance or the diodes won't limit
>hard.

It might be interesting to try a pair of very low voltage (1.8V) zeners in 
series (back to back). Then you don't need clamp voltage because you're 
clamping to the ground line.  When the signal level is low, the zener will 
be well away from the knee, so you'll just be looking at the reverse leakage.