[Date Prev][Date Next][Thread Prev][Thread Next] - [Date Index][Thread Index][Author Index]

RE: (SOUND CARD TESTING RESULTS) AO-40 K-Band and FEC telemetry



This is a generalization that needs comment.  I hope to
publish my results of how using the wrong cards with a software
defined radio kills its performance. Given the SDR-1000,
the sound card is the single most important element in the
entire processing chain.  While the dynamic range of
almost all modern sound cards is indeed gross overkill
for the most typical amateur radio applications (given the
dynamic range of the audio coming from most receivers), this
does not taking into account other measures: noisy front ends,
the sensitivity of the cards to common mode problems, horrible
phase and amplitude distortions, which can and do drastically
impede a lot of what hams are doing every day with their
cards.  I measured two cards where after digitization, the
two tone mixing products where remarkable.  I still don't
know how this got out of the engineers CAD programs!

I recently took every SB Live in my shack and through them
into the trash after I discovered just how noisy and poorly
designed the analog front end is on the record side. These
cards also regularly showed a 1% DIFFERENCE between the
measured gazinta and gazouta sample rates at 11025Hz, a
very popular sample rate in amateur radio applications. I also
determined that many drivers are doing polyphase resampling
or the cards are doing polyphase resampling and the design
of the resampler is horrible! Many users of other people's
.dll files to do signal processing of the types (modes) that
are now popular, are finding that the assumption of equal A/D
and D/A sample rates is a real killer in terms of performance.
The designer of one of the most popular dll's told me this
himself in a private communication.

I think designers of software should measure the receive sample
rate using the typical WWV method and then loop back to determine
transmit sample rates.  Phil's FEC application is receive only
so this does not apply of course.

As the things we ask our sound cards become more complex
(than digitize straight audio coming from the speaker jack),
these performance measures will become increasingly important.
I am pointing out that one company seems to have gotten their
act together and introduced an inexpensive card that does a
better job I thought possible at the price.

Bob
N4HY



-----Original Message-----
From: owner-AMSAT-BB@AMSAT.Org [mailto:owner-AMSAT-BB@AMSAT.Org]On
Behalf Of Phil Karn
Sent: Monday, August 18, 2003 01:29
To: Robert McGwier
Cc: Amsat Bulletin Board
Subject: Re: [amsat-bb] (SOUND CARD TESTING RESULTS) AO-40 K-Band and
FEC telemetry


Robert McGwier wrote:
> For those of you doing really critical performance
> DSP using the sound card in your machines, I have
> a recommendation or two to make.  I have now done extensive
> testing of all sorts of sound cards and I have picked
> a $750 card for 24 bits but a $50 card using 16 bits
> (18 bit A/D where high order 16 is used).

These premium cards certainly have their place in exacting hi fi audio 
applications, but they're incredible overkill for ham DSP applications.

There is one aspect of sound card quality that is relevant to ham DSP, 
and it's not signal-to-noise ratio or the number of A/D bits. It's the 
accuracy and stability of the sample clock. I've run into a few 
inaccurate sampling clocks in my AO-40 FEC work, and it can be a real 
nuisance.

Phil
----
Sent via amsat-bb@amsat.org. Opinions expressed are those of the author.
Not an AMSAT member? Join now to support the amateur satellite program!
To unsubscribe, send "unsubscribe amsat-bb" to Majordomo@amsat.org
----
Sent via amsat-bb@amsat.org. Opinions expressed are those of the author.
Not an AMSAT member? Join now to support the amateur satellite program!
To unsubscribe, send "unsubscribe amsat-bb" to Majordomo@amsat.org



AMSAT Top AMSAT Home