Dear LF friends
Hope you all experienced a "good landing" in new year 2000.
Using the DSP Receiver KWZ30 manufactured by Kneisner & Doering
http://kd-elektronik.com/kwz30_e/index.html
I come acros receiption problems which are related to the
spectrum occupancy and the Receiver Design.
=========================
The frontend of the KWZ30 handles the whole input spectrum
from almost DC up to 30 MHz, independent from the tuned frequency.
This is not a principle problem, as the first mixer has sufficient
large signal handling capabiity.
First Mixer Output is at 75 MHz, which is filtered by a roofing
filter, a 8-pole crystal filter with phase linear characteristic.
The nominal BW of this filter is +/- 7.5 kHz (15kHz total) with
moderate shape factor in order to keep phase deviations to a
minimum. It does not (cannot) have the sharp shape as is common
with SSB filters of the 8-pole design!
Next in the IF chain comes the first IF amplifier, the second mixer,
which outputs 456 kHz as the second IF unfiltered to the second IF
amplifier. Up to this point the IF chain is in traditional analog
design.
All signals falling in the passband of the roofing filter
i.e. +/- 7.5 kHz are processed in the same way as is the tuned
frequency. The total level at the second IF amplifier output is
limited by AGC action to a fixed level which is compatible with
the following A/D converter (in order to not overload it).
The strongest signal in the analog IF chain mainly determines the
effective gain set by the AGC.
After digitizing the IF on 456 kHz level, a DSP accomplishes
final IF filtering with 9 selectable banwidths from 9kHz down to
50Hz, accomplishes demodulation, notching and noise reduction.
After all a DA converter retrieves the audio signal which is then
fed to a constant level recorder output and via audio amplifier
to headphone or speaker.
A more comprehensive description and test results can be obtained
from http://www.rnw.nl/realradio/kwz30_a4.pdf
=========================
Now lets have a look on the spectrum in vincinity of 137 kHz at my
place, captured by a (broadband) electrical vertically polarized
active antenna, 1m above GND. I concentrate on "big" signals only.
(1) DCF42 with it's pilot tone on 122.5 kHz is -55dBm into 50 Ohms
(2) DCF49 on mark frequency of 128.930 kHz is -31dBm
(3) DCF39 on mark frequency of 138.830 kHz is -52dBm
(4) DLF on carrier frequency of 153 kHz is -29 dBm
When tuned to exactly 137 kHz the frequency offsets
of these signals are:
(1) -14.5 kHz
(2) -8.07 kHz
(3) +1.83 kHz
(4) +16.0 kHz
=========================
Signals (1) and (4) fall into the stopband of the roofing
filter and are attenuated sufficiently and do not affect the
AGC action of the receiver (do not reduce the gain of the
analog IF chain).
Signals (2) and (3) however fall into the passband and experience
almost no attenuation relative to the tuned frequency of 137 kHz.
This leads to AGC to reduce gain in the analog IF chain in order to
limit the level fed to the ADC some dB below its overload point.
Theoretically (under laboratory conditions) the DSP filtering
and demodualtion IF chain can process signals with a dynamic
range of no more than 90dB (restricted by the ampitude resolution
of the 16 bit ADC).
Practically I can detect a weak signal on 137 kHz which is
-102dBm in level (equal to the noiselevel). With the level of
DCF49 of -31 dBm this makes a useful dynamic range of 71 dB
from largest signal (passing the roofing filter) to noise.
=========================
Practically this situation prevents me from receiving LF
amateur stations at distances beyond 70km !
The only stations heared up to now are:
DJ8WL alias DA0LF (27 km)
DF2PY (27 km) level -91 dBm
DF8ZR (31 km) level -102 dBm
DL3FDO (72 km) level -98 dBm
=========================
The limitations set by my special Receiver model is simply
due to physical law. Any Receiver of similar frontend design
will experience the same problem. Most Receivers I have operated yet
do not have roofing filters of less than +/- 3 kHz (6 kHz) BW
and shape factors better than 3 at the same time, except for
special CW or RTTY receivers.
To overcome gain alias sensitivity reduction due to AGC being
affected by the nearby (frequency and distance) strong stations,
one shall use Receivers with "narrow" roofing filters.
For people living close to DCF49 a receiver shall have a roofing
filter (assuming reasonable steep roll off, i.e. shape factor of
2 or better) of no more than about +/- 4 kHz (8 kHz) in order to
prevent AGC is affected by DCF49 signal level.
For people living close to DCF39 a receiver shall have a roofing
filter (assuming reasonable steep roll off, i.e. shape factor of
2 or better) of no more than +/- 0.9 kHz (1.8 kHz) in order to
prevent AGC is affected by DCF39 signal level.
To make use of Receivers having broader roofing filters,
a dedicated LF preselector is indispensable. I have contacted
DK1AG Bernd Neubig [[email protected]] who is about to design
such a dedicated LF preselector. He will show up in this forum
as soon as he finished measurements and a short presentation.
Using "selective" antennas is of no great help. I measured
the -3dB BW of my GWS-loop-9 antenna to be 900 Hz @ 137 kHz,
which leads to a Q of 152. This sounds great but signal (2)
DCF49 is attenuated by only 7 dB, signal (3) by 2 dB whwn antenna
tuning was optimized for 137 kHz.
=========================
DK8KW, Geri Kinzel provided me with data on his LF receiver
LWF45 from Teletron. The built in co-tuned preselector
when tuned to exactly 137 kHz provides for the following
attenuation figures:
(1) DCF42 -14.5 kHz -41 dB
(2) DCF49 -8.07 kHz -25 dB
(3) DCF39 +1.83 kHz -4 dB
(4) DLF +16.0 kHz -50 dB
Best 73 de Gamal Soegiono
|
