RE: [linrad] Quietening computer noise

[Leif Åsbrink <leif.asbrink@xxxxxxxxxxxxxxxx>]

Hi Richard,

> I have managed to recieve signals on HF using two SBL-1 mixers driven by 
> a quadrature LO (AD9854) and a couple of low noise opamps (NE5534). 
> Linrad runs in direct conversion I/Q mode. The main limiting factor at 
> present is noise from both computer and monitor.
> The sound card is bog standard and the machine is a PII 400 MHz running 
> RH9, but with only 64MB of RAM
> 
> Questions
> The max filter bandwidth appears to be about 2 KHz on SSB, and narrower 
> on AM  - if I try to make it wider nothing happens.
> Is this due to RAM and processor limitations?
The bandwidth is limited by two mechanisms. 
First of all you can not get more bandwidth than acceptable for
the output sampling rate that you have choosen. 
(If you force a larger bandwidth by setting it while the output
rate is high, then set the output rate low, you may be caught
in an endless loop. I just discovered and will fix it for the
next version.)

For SSB you might want frequencies from 300 to 3000 Hz so the
output rate has to be something like 8 kHz. For AM you need even
more. Perhaps 11025 Hz.

The other factor is the baseband sampling rate. It is a power of two
lower than the input sampling rate. The filter bandwidth is limited
to 50 % of the bandwidth permitted by the sampling rate. The maximum
sampling rate is half the sound card rate (First mixer bandwidth
reduction....[1] )

All of this means that if you want a 10 kHz filter to receive AM
you have to send 40 kHz of bandwidth into Linrad. To do it you must
set the soundcard sampling speed to 40 kHz or more.
Then set the minimum mixer N.


> How to get rid of that computer noise? ( I have made no attempt whatever 
> so far)
Depends on the frequency. On HF I do not think it is terribly difficult.
The way wires are routed is very important. The computer itself (screen and
other parts included) is too small to radiate, but local magnetic fields
may drive currents into wires that send interference into lagre enough
structures to couple to the antenna.

73

Leif  /  SM5BSZ

LINRADDARNIL