Content-Transfer-Encoding: 8bit
Message-ID: <5.1.0.14.0.20020723153951.00af7b98@gemini.herts.ac.uk>
Date: Tue, 23 Jul 2002 17:08:48 +0100
To: rsgb_lf_group@blacksheep.org
From: "James Moritz" <j.r.moritz@herts.ac.uk>
Subject: Re: LF: Sound Cards
In-reply-to: <5.1.1.6.0.20020722224253.009efa70@pop.erols.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Sender: <majordom@post.thorcom.com>

Dear Sandy, LF group,

As I recall, the average sound card ADC claims about 70 or 80dB dynamic 
range - while I have no doubt some have much better performance, the RX 
IF/detector/AF sections are almost never going to be good enough to see any 
benefit from an improved ADC.

This is understandable - as far as the designers are concerned, in a normal 
HF/LF communications receiver design, all the channel filtering is 
performed by the IF filter, and whatever signal is present at the filter 
output is the "wanted" signal. So all the subsequent stages of the receiver 
are expected to do is maintain an adequate signal to noise/distortion ratio 
- for communications purposes, 40dB would be more than enough.

With LF narrow-band modes the situation is different - the selectivity and 
unwanted signal rejection is defined entirely by the software, so all 
circuits up to and including the sound card ADC are effectively the 
front-end of the receiver, and have to handle large amounts of unwanted 
signals and noise without distorting a weak wanted signal - not something 
they were ever designed to do. This is the same as the argument against 
add-on CW audio filters. However, in practice it usually seems to be 
possible to set the gain so that neither receiver nor sound card are 
overloaded, and the receiver noise level is only a tiny fraction of the 
band noise, so it probably does not matter much most of the time.

Where it is a problem is when there are large unwanted signals in the IF 
passband - this sometimes happens here when you are trying to copy a weak 
QRSS signal, and there is a local station operating at the same time. 
During the transatlantic tests I found that with the right adjustment of 
receiver and sound card gain controls, along with the display parameters, 
it was possible to copy a weak signal 50 - 60 dB below a strong local 
signal, when 30s dots were being used.

It would certainly be possible to design an RX with high dynamic range, low 
noise IF/detector/Audio stages with optimised gain distribution to 
complement a high performance sound card. A more integrated approach is a 
receiver with IF DSP, with the ADC digitising the IF signal instead of the 
audio. But with current LF usage, especially in the US, it would probably 
be a benefit only 1% of the time.

Cheers, Jim Moritz
73 de M0BMU

PS - even for the top-notch sound cards on the web site you mention, it 
still says nothing about the sample rate accuracy!


At 23:11 22/07/2002 -0400, you wrote:
>   I have a question? How important is the sound card to the operation of 
> DSP software (Argo etc). I think we have proved that almost any card will 
> work but would we gain any thing by buying a card with a better signal to 
> noise ratio or better dynamic range?
>  This web page has test results showing many DB difference between cards 
> (click on the card name for details on each card).
>
>  http://www.pcavtech.com/soundcards/summary/index.htm
>
>                                         Sandy
>                                         WB5MMB
>
>