|
| TM
11-5820-917-13
high-speed recirculating memory is D-to-A converted and the resulting analog
waveform is a "time-compressed- frequency expanded" replica of the sampled input
waveform which repeats every memory recirculation time. The D-to-A output is then
fed to a heterodyne mixer whose local oscillator injection is a frequency sweep (from
a linear VCO) covering the entire bandwidth of the frequency-expanded signal
every 60 milliseconds. The output of the mixer is applied to a single, amplitude-
sampled filter that is synchronized to the 60-millisecond VCO frequency sweep and
the 300 microsecond memory recirculation. (The filter is sampled every 300 micro-
seconds.) The output of this filter is a series of voltage samples which represent
the amplitudes of the spectral components of the analyzed waveform. The resolution
of the analyzer is "200 lines", (200 memory recirculation every 60 milliseconds)
which is equivalent to dividing the analyzer input bandwidth (e.g. 500 Hz) in 200
parallel, frequency analysis filters. It is this high resolution spectral analysis which
gives the chirp sounding technique the ability to make clearly defined, highly sensi-
tive soundings with good "mode" or time delay, resolution. Functionally, the spectrum
analyzer consists of an input section, dealing with the digital processing (time com-
pression speedup) of the input signal, and an output section, covering the analog
swept filter analyzer.
is first amplified and then sent to a sample-and-hold circuit whose sampling rate is
selected relative to the analysis range of interest. The analysis frequency range is
programmed by changing TTL programming lines and controlling the selection (U15)
of the sample rate divider. The two analysis ranges normally used by the sounder are
500 Hz and 5000 Hz, which require sample rates of 1500 and 15000 samples per second,
respectively. The sampling rate is always three times the highest input frequency of
interest. The normal analysis range for the receiver (except during auto sync) is
500 Hz. Analog samples of the input waveform taken by the sample and hold circuit
Q1 at the selected sampling frequency, are entered into an A-to-D converter
used to latch the parallel output of the A/D converter before entering the data into
the memory load buffers. The clocking pulse for these flip-flops is the end-of-con-
version pulse (TP5) derived from the sampling circuit. The memory load buffers are
eight-bit shift registers whose digital bits are clocked forward by the buffer control
circuit (figure FO-9/4). The new data is clocked into the load buffers at the sampling
rate. The transfer of data from the buffer to the memory is keyed by the 2-MHz
memory clock but is also dependent on the status of the buffer control anti-coinci-
dence circuits. These circuits prevent coincidence problems between the loading
and unloading of the load buffer registers and insures that the contents of the
buffer are only clocked out following the "tail" of the circulating memory.
4-13. The memory (figure FO-9/5) is a 600 word, 6-bit per word, MOS shift
register loop operating at a two megaword per second rate (300 microseconds per
memory revolution). Two 2-MHz clock signals of different phase, derived from the
12-MHz oscillator (figure FO-9/1) continuously clock each of six parallel storage
registers which collectively circulate data words at a 2-MHz rate. Each of the six
parallel registers consists of three, serially connected, 200-bit long, MOS shift
registers to produce the required 600 word long memory. Each time a new input
sample is available from the load buffer, it is loaded in place of the oldest memory
sample following the "memory tail" (figure FO-9/4). This is achieved using a divide-
by-600 counter (U6, U7, and U8) which interfaces with shift register U29 to move
|
Privacy Statement - Press Release - Copyright Information. - Contact Us |