Hi,
Given the info from Scott about the coming Orion, I thought
some of you might appreciate a bit of info about the Ten
Tec HF DSP rcvr, the RX-340. Will be interesting to see
what similarities/differences between these two rigs may
be. What follows is directly from Chap. 3, 340 "General
Theory of Operation", and just the first few paragraphs of
that to show the overall rcvr plan. Some phrases are omitted
for brevity, but all the tech info is included:
"The Ten Tec Model RX-340 combines a high dynamic range
front end with a versatile DSP back end to provide extraordinary
performance/versatility.
"RF signals from the antenna input are preselected by a switchable
1/2 octave bandpass filter network (note: rcvd frequency coverage
is 50kHz to 30 mHz for all specs" (unless otherwise given);
but the rcvr will respond to signals in the VLF range down to
5 kHz at 300 Hz bandwidth, with the preamp OFF, to signals at 5 kHz
as weak as -82 dBm[18 uV] to provide a 16 dB SINAD.
"Balanced amplifiers and a high-level 1st-mixer preserve 2nd and
3rd-order intercept points during conversion to the 1st-IF
frequency of 45.455 mHz. A pair of 2-pole crystal filters
provides a 1st-IF selectivity of 16 kHz to reject 1st-mixer spurious
products and the 2nd-mixer image at 910 kHz offset." Both
mixer outputs are available at BNC connectors on the rear panel.
"After conversion to the 2nd-IF frequency (455 kHz), signals are
filtered for a 16 kHz bandwidth and split off to the 3rd-mixer
stage and "Signal Monitor" output (J3) BNC."
"The 3rd mixer converts signals to a center frequency of
16 2/3rds kHz where they pass through a low pass filter
and are applied to an analog-to-digital converter. The
A/D converter produces a serial data stream at a 66 2/3
kHz sample rate for input to the Digital Signal Processor."
"Serial data from the DSP (at a 133 1/3 kHz sample rate)
is applied to a digital-to-analog converter. The D/A output
samples are de-multiplexed into two or three output
channels, depending upon mode selection. Half of the D/A
output time is devoted to the DSP's IF output, which is first
converted back to 455 kHz by mixing with the third LO, then
filtered for a 16 kHz bandwidth and finally made available
at the IF Output connector (J4)BNC."
The other half of the D/A bandwidth is separated into
USB and LSB audio channels in ISB mode, or into a
single audio channel in all other modes."
At this point in the General Theory section, the discussion
breaks off into a board-by-board operating/interconnect
description, and makes frequent reference to a rather complete
block diagram of the 340.
Some features of the 340 I have learned.
The rcvr noise level, from all internal sources, is below
the HF band noise level at all frequencies and operating
bandwidths I have checked. On the 15 meter band:
Noise floors
a. First with the antenna input terminal shorted to ground.
CW mode, 600 Hz bandwidth: Noise level at -137 dBm.
(Note: all these numbers are read directly from the astonishingly
accurate "S" meter on the 340; see another of my posts about
the S meter, back on 12/30/01 are two posts about it).
b. As above, but 100 Hz bandwidth: Noise level at -140 dBm.
c. With the antenna terminal connected to my C-3 antenna,
600 Hz bandwidth, noise floor at -134 dBm.
d. As above, but 100 Hz bandwidth: noise floor at -138 dBm.
10 Meter Results from Noise Floor and Birdies
a. Noise floor, antenna terminal grounded, CW mode,
600 Hz bandwidth, noise level of -135 dBm seen.
Reduce bandwidth to 100 Hz, noise level of -140 dBm.
b. Connect antenna terminal to the C3 antenna system,
noise floor at 600 Hz BW increased to between -125 and
-130 dBm, meaning the 340 noise floor is below the band
noise at 10 meters, our most quiet HF band do to external,
galactic, etc sources.
At 100 Hz bandwidth connected to the antenna system: -138 dBm
seen.
c. I have been asked about spurious birdies in the 340. There
are essentially none.
c.1 On 15 M: Scanning slowly using both 10 Hz and
1 Hz "tuning rates", I found none.
c.2 10 Meter Birdie Scan Results
Scan with the antenna terminal of the 340 grounded.
Two birdies found on 10 meters:
one at 28139, amplitude is of -138 dBm.
second at 28631 area at -117 dBm.
Birdies at 10 metes are of concern to those using VHF/UHF
transverters; 340 seems an excellent choice for such.
Note that the specs by Ten Tec for the 340 look "bad" for
these parameters. As they sell these to the government,
with the specs and handbook prepared per MIL-HDBK-1221,
per the manual front cover, TT must be extremely conservative,
fessing up to the very "worst" possible numbers, even if they
should only occur one time somewhere within the specified
frequency range at one of the possible operating bandwidth
selections.
Final comment about the 340 I have at this time, is about
the remarkable S-meter -- it's accuracy and behavior.
It is driven by the CPU within the DSP section. It is the
only S meter I have ever seen which operates independent
of the RF gain control level. And, per my checks with a
"precision" step attenuator, the S units are almost
exactly 6 dB steps from S9 all the way down to S2, and
very close on down to S1. There is a parallel dBm scale
to the S unit scale. S1 is about even with an input signal
power level of just under -120 dBm. The meter is calibrated
on down to -140 dBm, and as noted above in the noise
floor checks seems to be accurate down there as well.
Well, that may be more than you really wanted to know, hi.
73, Jim KH7M
|