Topband: MFJ 1025

[Tom Rauch]

> I have never been able to find a null when the 2 signals 
> (noise and main)
> are "equal". In fact it always seemed that the noise had 
> to be at least 3 db
> less than the main signal using the receivers S meter.  I 
> know this fly's
> against the manual guidance and the nulling theory, but 
> this is what I, and  other
> have found here using either the MFJ unit or the ANC4.

Noise often has a high ratio of peak to average power, S 
meters don't always follow noise levels well for sharp noise 
spikes with wide spacing. This is especially true when DSP 
systems with wide IF stage filtering are involved since they 
overload (not enough bits or speed to maintain pulse shape).

There was a fellow on the RFI reflector who was puzzled why, 
when he switched in an attenuator, his S meter didn't show 
noise dropping at the same rate as desired signals. His 
noise dropped faster than the signal. This meant he could 
use an attenuator and improve S/N ratio. After some emails 
between us he tried a receiver that used crystal IF filters 
instead of  a DSP for primary selectivity and everything 
behaved normally.

A guesS meter is certainly not the final authority on signal 
levels, especially when some of the signal content is a 
rapidly rising and falling pulse mixed with steady 
signals....and we are trying to null the signal that does 
not move the meter well.

By the way, the preamp in the 1026 is a voltage follower. It 
depends on the input impedance being higher than the output 
impedance to have gain. It has current gain, not voltage 
gain. If the input is not high impedance (case in point the 
internal whip) compared to the stages following the preamp 
(the preamp load) it has no gain. If you need to amplify 
without a high Z source on the noise input my suggestion is 
a good external preamp.

I match the background noise level on all my antennas here 
before my switching system by using preamps and/or 
attenuators. This way I don't get a false feeling of an 
antenna being "louder and better" when checking on moderate 
signals when it really isn't better for S/N ratio weak 
signals.

73 Tom