[Moon-net] Linrad

[Leif Asbrink]

Hi Lance,

> I don't think a Tektronics pulse generator is in the cards for 
> me right now....I sure hope there is an easier way to get this software to work!
Yes. Calibrate your SDR-IQ at HF with a simple square wave.
http://www.sm5bsz.com/linuxdsp/flat/ampcal.htm

Actually this is quite stupid. The exact filter response of the
AD6620 is known, I have just been too lazy to implement the
equations necessary to have the SDR-IQ pre-calibrated.
It is quite another thing that you might want to calibrate
from the antenna in case you want an extremely flat noise
background when looking at long-time averaged spectra for
finding signals where (S+N)/N is only 0.1 dB or so. 

> Today I was able to get LINRAD running for SSB on Windows Vista 64 bit O/S, and it 
> did not freeze ;-)  So I may just try to keep running it on that platform. 
Maybe the latest updates from Microsoft would make XP
behave better...


> The 
> signal level appears to be VERY strong from the SDR-IQ (no-signal noise level is 
> around 20 dB over S9), so maybe that is part of the problem here.  Anyway, I am 
> looking forward to slowly improving the settings and Linrad capabilities here.

This is because you mis-use the SDR-IQ;-) The USB has 16 bit only
while the AD6620 provides 23 bits. Use the "Set RCF output shift (0 - 7)"
to use the bits that place the noise floor correctly.
(Read the data sheet of the AD6620 for info.)

The SDR-IQ is the first of what will become our "conventional receivers"
in the future. All VHF sampling receivers share some common problems that
the operator has to handle to get optimum dynamic range. For the SDR-IQ
one should do the hardware set-up like this:

1) Set the gain between the antenna and the A/D converter as high
as possible. Occasional AD overflows are harmless (on QRN) but 
the gain must be set low enough for the strongest narrowband signals
that may reach the A/D converter to never cause saturation.
On the SDR-14 there is a red LED that shows when AD saturation
occurs. I do not own an SDR-IQ so I do not know if it also has one.

2) Set decimation numbers as required for the application.

3) Set the RCF output shift for the noise floor to become 
about 20 dB. Verify that the strongest signal you ever 
have in the visible window does not cause overflow in the
USB link. If it does you must place the noise floor 
even lower and accept some quantization noise from the 
USB link that would degrade your system NF a little. 
Ultimately you might find that the USB link is the
dominating noise contribution in your system noise
figure - but then the strong signals that you have to
put up with would make a low NF useless anyway because
of the sideband noise all the amateur transceivers
produce.

Note that most amateurs do not need the optimum dynamic range.
You may use almost anything for the gain at the SDR-IQ 
input and output if you never suffer from really strong
undesired signals. A too high noise floor could then be 
lowered with the "First FFT amplitude" parameter of Linrad.

73

Leif / SM5BSZ