Dear Stefan, LF Group,
Today i did some new sensitivity measurements on my homemade RX (137
kHz -125 kHz => SpecLab).
I took an analog homemade VFO to generate a sine wave at 137 kHz, 20 mVp
at 50 Ohm, measured with an oscilloscope. Then i added a 43 dB attenuator
between RX (which has a 50 Ohm input) and VFO.
So, assuming the oscilloscope also had a 50ohm input termination, the input
to the receiver would be (20mV - 43dB) = 142uV, actually 3dB more than the
100uV figure you mention.
In SpecLab (connected to L1) my background noise was -107 dB in 1 Hz and
the achieved signal level was -30 dB.
So, an input signal of 100 uV rms (within the passband) causes 77 dB S/N.
So the noise floor at the RX input in 1Hz bandwidth is (20mV - 43dB - 77dB)
= 20nV, i.e. a voltage noise density of 20nV/sqrt(Hz). This is 6.6dB below
my estimate of 43nV/sqrt(Hz) at the preamp output, in which case the RX
noise would increase the overall noise floor by only about 0.9dB, so yes,
the receiver sensitivity should be OK, if the other objectives are met.
No idea if there is a reduction of the performance of the RX when the full
LF spectrum is applied, instead of that single signal. If no non-linear
effects (IM, etc...) occur, there should be no difference. And we are
talking about a directional narrow band antenna so i do not expect to much
trouble, beeing an optimist ;-)
The antenna bandwidth will only be about 500Hz, so it should be a fairly
effective preselector...
That minimum input signal level is about 20 dB below your stated 0.75 uV
but i am referring to a 1 Hz bandwidth (width of one FFT bin). You were
talking about a 300 Hz CW bandwidth. But i am configuring SpecLab to
convert the signal (talking about CW) to 700 Hz and apply a 300 Hz filter.
This works very well, as presented in the HB9ASB.wav file.
So do i misunderstand some things about the bandwidth
settings/calculations?
The "voltage noise density" in Volts per sqrt(Hz) is the noise level
referred to a 1Hz bandwidth, which makes it easy to compare systems with
different bandwidths. To get the noise voltage in e.g. 300Hz, multiply by
sqrt(300). So your RX noise floor would be 0.35uV in a 300Hz CW bandwidth.
Which is not brilliant, but better than many amateur HF rig receivers at
137kHz!
Cheers, Jim Moritz
73 de M0BMU
|
