[Lowfer] LF Receivers?

[Clint Turner]

Hi all,

I have been using the SDR-14 for LF reception for while now.  This 
device is a bit less-convenient than the SDR-IQ (it needs a separate 
power supply) and its inputs are different:  The SDR-IQ has 
bandpass-filtered inputs that range from a few hundred Hz to 30 MHz 
whereas the SDR-14 has two separate inputs - one that is filtered from 
100 kHz to 30 kHz (no bandpass filtering) and another with no 
amplification/filtering at all that allows inputs from a few hundred Hz 
to several hundred MHz (using undersampling techniques.)  The SDR-14 is 
a bit more expensive and has been discontinued.

The unit works quite well and is fairly stable - although that's not too 
hard to accomplish at LF frequencies:  At HF/VHF where sub-hertz 
stability is required, an external reference (not built-in) would 
probably be required - either that, or one would use a program like 
Spectrum Lab along with an injected, stable reference signal within the 
passband to provide a frequency reference.

Two things that I've observed with the SDR-14 - and is probably true 
with similar units like the SDR-IQ and Perseus:

- The USB line can carry tons of noise.  Since the PC itself has 
switching supplies and is connected to the power mains in the shack, it 
tends to carry quite a bit of grunge.  To keep this out of everything I 
wound a long (6'-8') USB cable - leaving a few inches of pigtails - onto 
a discarded flyback transformer core yielding about 1.5mH of 
inductance.  This effectively stops the conduction of such noises - even 
down to 10's of kHz - when combined with proper earth grounding of the 
LF receive antenna's coax - which, in my case, is also wound on a 
flyback ferrite to further-block "shack noise" into the coax and to the 
antenna.

- The noise blanker on the software can work phenomenally well, taking 
light dimmer noise pretty much completely out - about like my 
synchronous noise blanker except that I don't have to twiddle knobs nor 
worry about adding other channels when other noises sources appear.  
This is true ONLY if I'm running a lower the sampling bandwidth (the 
"50kHz" mode - which is really 48 kHz using an internal sampling rate 
closer to 24 kHz).  In fact, the noise blanker is "worse than useless" 
if one attempts to use it in the "190kHz" mode and becomes increasingly 
effective with the narrow bandwidth.  (50kHz is as narrow as it 
gets...)  I suspect that this has something to do with the fact that at 
24 kHz the lowpass (anti-alias filtering) is far more effective than 
that at the wider frequencies - or it's just not being "desensed" by 
other signals/noise over the wide bandwidth - but these are only guesses.

The programs that I've been using are the SpectraVue and the "SDR 
Console" program.  The former comes with the SDR-14/SDR-IQ (and others) 
while the latter works with a number of devices - including sound 
cards.  The noise blankers in both of these programs work about the same 
(although the one in "SpectraVue" may work slightly better - it's hard 
to say...)  The "SDR Console" program is somewhat buggy, but these are 
being ironed out as time goes on.  (I have also tried the "PowerSDR" 
program - another that will with with a wide variety of devices 
including sound cards) but the versions that I've tried have been too 
flaky on my system - a 32 bit XP box - to be very usable.)

Another program is Spectrum Lab:  It's a bit kludgy when interfacing 
with these USB-type devices, but it works very well with processing 
sound card data - although the learning curve is steep.

Interestingly, I have three sound cards in my computer - plus some 
"virtual" sound cards:

- The built-in (motherboard) sound card is used for the PC speakers and 
microphone and general-purpose uses.  This is the operating system 
default device.

- An old Soundblaster 16.  This is connected to the sound card interface 
of the HF rig for various modes.

- An Asus Xonar D-1.  This is a pretty high-performance (24 bits/192 kHz 
according to the ratings, but it seems to only do 16 bits/192 kHz under 
XP) but relatively cheap (<$100) card that is used for the "sound card" 
radios like the "SoftRock" and similar.

Having these extra cards is nice in that I don't have to constantly move 
cables around, plus with the default Windows card being the built-in one 
(which has only "ok" performance) NOT being connected to any radio means 
that "system noises" (dings, bloops, etc.) will never find its way on 
the air:  I don't know how many times I've heard startup sounds, dings, 
music and things like "You've Got Mail!" in the middle of the PSK bands!

I also have "Virtual Audio Cable" which allows routing from one program 
to the other without involving a sound card without using a cable - say, 
routing audio from SpectraVue to the WSPR program:  I've even used 
SpectraVue to output a wideband (24 kHz) USB signal and feed it to 
several instances of the "SDRadio" program (by I2PHD) using one virtual 
cable where I could demodulate multiple signals across that 24 kHz 
passband and feed them individually - using even more virtual audio 
cables - even more programs for digital "sound card mode" demodulation.

This seems to work pretty well - although I've found a few programs that 
don't play nice:

- Spectran doesn't allow specification of different input/output devices 
which sometimes causes odd feedback effects.  (Argo doesn't seem to, 
either, but this is less important.)

- The WSPR program (at least 2.11) just seems flaky in general when 
trying to change audio sources:  Usually, it does what it's supposed to, 
but sometimes several "exit and restart" things are required.  Also, 
I've not been able to get more than one instance of that version of WSPR 
to run - which would have been handy to decode both 500 kHz WSPR beacons 
and simultaneously run WSPR on HF.  (I've seen this comment from others, 
so it isn't just me...)

One interesting problem that shows up is trying to decouple the 
"wideband" audio - say, the 0-96kHz audio - from an SDR 
downconverter/upconverter rig (e.g. "SoftRock") to the computer to 
prevent ground loops (hum), to keep TX RF from getting into the 
computer, and to keep computer/power system RF from being conducted onto 
the coax and appearing at the RF input somehow:  It's a bit risky to use 
transformers for this wide frequency range if you want to maintain the 
best phase/amplitude response so you are left with either 
differential-input amplifiers (op-amp-types) that are carefully designed 
to be both RF-proof and have equal phase/amplitude response (to maintain 
I/Q "image" rejection) and/or use multifilar chokes on the audio lines.  
So far I've gotten away with the multifilar choke approach...

Yes, I know that this response probably borders on "TMI" with respect to 
the question, but there it is...

73,

Clint
KA7OEI/CT

Bill & Becky wrote:
> Hi Group  I'm in the process of replacing my TT 320D receiver.  I'm looking for opinions on which receiver for LW QRSS, NDB and below 6MHz.
>
> I do not fancy getting a Transceiver because of the added cost and I'm not interested in SSB/CW HF QSO's
>
> I'm looking for time tested experience by you.........where the rubber meets the road.
>
> You may send your thoughts off list to  wmarvin at hickorytech.net
>
> Best Holidays to All
>
> Bill  KB9IV