[Moon-net] Linrad and SDR-IQ for weak CW

[Leif Asbrink]

Hi Lance,

> I got a second computer last 
> summer with the intent of using it with my SDR-IQ, LINRAD and possibly other 
> programs that can captitalize on the wide bandwidth of the SDR-IQ receiver.
Good:-)

> I do finally have my second computer working (after sending the MOBO back twice), 
> and it really seems to fly with the dual AMD 64 processor :-)  I have installed 
> Vista 64 on one hard disk and 32 bit Windows XP on the other, and I have pretty 
> much resigned myself to just using the Windows XP O/S for the time being because I 
> can get it to work with various drivers.
I have no idea to what extent Vista can be used with Linrad. I have had only
one preliminary report saying one has to set Vista to support old style
drivers or something similar. It would not work with 32 bit drivers
however so I had to make a couple of changes. I have not been able to get
any feedback whether the changes were sucessful because my source of Vista
information decided he did not want the new OS for various reasons.

Once you have Linrad running under XP I would be interested in getting feedback from
you whether it works on Vista also.

> I also finally now have an M-Audio Delta 
> 44 sound card in it, with the idea of being able to have 96 kHz of bandwidth one 
> way or another. 
???????????
You plan to not use the SDR-IQ? Or do you want to run two instances of Linrad
simultaneously? With SDR-IQ as input there is no need for another output than
the low performance built-in soundcard. The S/N of the signals you are interested
in is well below what the worst soundcard possibly can provide:-)

> I realize the SDR-IQ has twice that bandwidth, but I understand 
> that Linrad is limited to 96 kHz.
NO NO. Linrad has no restrictions. You can set whatever decimation
you want, but the USB drivers of your system will set a limit.
Under Linux I can run at something like 220 kHz on my Pentium IV.
Under Windows 2000 the speed is limited to something like 180 kHz
on the same computer. I do not know what your high performance 
machine might be capable of under XP (or Vista).

> I have heard that LINRAD can work very well for copying weak CW
YES.

> but also that my 
> "filters need to be calibrated" - whatever that means.  Has anyone with more 
> technical savvy and equipment than me who is using an SDR-IQ for input been able 
> to "calibrate their filters"?  Is there some setup file or something that you 
> could share with me?
Calibration in Linrad serves the purpose of knowing the exact frequency 
response of your system. It is needed for the clever noise blanker, but
if your interference situation is uncomplicated there is no need. The dumb
blanker in Linrad is already very efficient and it will work on an uncalibrated
system.

Once you have Linrad running and have decided what decimation ratios you want
to set on the different decimation steps of the AD6620, someone else can
provide you with the calibration functions for your SDR-IQ. (It only contains
digital filters with known responses) Any contribution from whatever filters
your transverter contains would then not be included, but if those filters are
reasonably wide they would not make any difference.

> By the way, I have been using this new faster computer with the Delta 44 sound 
> card as a backup to monitor JT65A skeds on 6m EME, and I have been very pleasantly 
> surprised by the results!  I had been very pleased with the performance of my 
> Turtle Beach Santa Cruz sound card on my primary computer (a single AMD 64 
> processor).  However I now find that the dual processor computer with the Delta 44 
> sound card almost always decodes the JT65A signals 1 dB stronger than my main 
> computer....and often it will produce a full decode showing the callsigns even 
> when my main computer only shows a "*" or *#" and no decode of the signal trace. 
This is a clear indication that something is wrong. 

> Of course, I entered the corrections factors for each sound card into the version 
> of WSJT running on each computer. So, although my original goal for getting the 
> Delta 44 was to be able to have more options with LINRAD, it is turning out to be 
> very effective for narrow band JT65A also.
I suggest you get a stable test signal and inject it into the antenna
somehow. A small X-tal oscillator with very loose coupling mounted
on the boom tube for example.

You can then use Linrad to measure S/N for that test signal on the output
of your transceiver when using one or the other soundcard. You can
see how S/N changes with antenna orientation and noise blanker on/off
and what else you can change. You can also measure S/N by use of the SDR-IQ
with and without the Linrad noise blanker. (In case you have frequent 
noise pulses that are too weak to trigger a conventional blanker you
might become surprised.....)

Just set the Linrad bandwidth to i.e. 1 kHz (or if there are few spurs, 5kHz
for a quicker response) and Use the Linrad S-meter. (the RMS, not peak values)

It will be a good idea to verify that the RMS noise floor falls by at least 17
dB when DC is switched off for your preamp. (Loss less than 0.1 dB due to 
subsequent amplifiers.)

> I would appreciate any suggestions anybody might be able to provide regarding the 
> SDR-IQ and filter calbration and/or LINRAD setup files for use on weak CW.  MNI 
> TNX and VY 73, Lance
Linrad is actually very easy to use - but not so easy to describe - since
it allows so many options. I suggest you join the mailing list:
http://groups.google.com/group/linrad
It will not overload your mailbox;-)

It could be a good idea to start from here:
http://www.vhfdx.radiocorner.net/LinradSDRIQ.html

The SDR-IQ is a first generation VHF-sampling amateur receiver. It may
suffer from a problem with spurs due to non-linearities in case it is 
operated at low signal levels. Have a look here:
http://www.sm5bsz.com/digdynam/practical.htm

The next generation, where the Perseus is the first unit I know of
is about 25 dB better (which is very good indeed):
http://www.sm5bsz.com/digdynam/practical.htm

The SDR-IQ should be perfectly adequate provided that you operate it at
a reasonably high signal level and have a pre-selector in front of
it to remove strong out-of-band signals. (A transverter should be fine
although high LO leakage could be a source of spurs.)
There are a couple examples linked to here:
http://www.sm5bsz.com/linuxdsp/usage/examples.htm

At my site and its mirrors there are 300 html documents. More than 50% 
of them are about Linrad but there is unfortunately no pedagogic 
structure. Ask at the Linrad mailing list to get hints on where to look.
Questions and answers would then be archived at the appropriate place:-)

73

Leif / SM5BSZ