Message-ID: <20827.200104191655@gemini>
From: "James Moritz" <j.r.moritz@herts.ac.uk>
Organization: University of Hertfordshire
To: rsgb_lf_group@blacksheep.org
Date: Thu, 19 Apr 2001 17:56:02 +0000
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII; format=flowed
Content-Transfer-Encoding: 8bit
Subject: LF: <Tech> WOLF tests, soundcard trouble, VE2IQ interface
In-reply-to: <003e01c0c800$0d3eab60$0700000a@parissn2>
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Sender: <majordom@post.thorcom.com>

Dear Brian, Stewart, LF group,

Thank you Brian for the report - the 30mW ERP signal was 
generated with abt 120W into my inverted L at 8m height. I will put 
the beacon on again tonight, starting at 2000utc, this time with 
30mW ERP for an hour, then going down to 10mW ERP.

I checked the sample rate errors of my soundcard - on 
8000samples/sec this is about -5600ppm and on 
11025samples/sec it is +2000ppm. So a lot bigger than likely 
errors due to crystal tolerance, for some reason. I have been 
looking at the clock crystal frequencies in all the soundcards I 
could find - the older ones had all sorts of different frequencies, 
most of which had no obvious integer relationship with popular 
sample rates. Quite a few used  14.31818 MHz, 46.61512 was 
popular too. Some had 2 or 3 crystals, one had none at all. The 
newer ones have mostly 24.576MHz crystals - this seems to be 
the norm for the AC '97 standard based soundcards which Alberto 
mentioned recently, which seem to be universal now. I found some 
data on 3 types of "audio codec" A/D and D/A which I saw, which 
conform to this standard:

Sigmatel STAC9721T
http://www.sigmatel.com/technical_docs.htm#9721
Crystal CS4201
http://207.87.22.21/design/products/overview/index.cfm?ProductID
=138
Analog Devices AD1818
http://www.analog.com/techsupt/application_notes/EE_notes/pdf_fil
es/ee_54.pdf

Although 48k seems to be the basic sample rate of the standard, 
they all have hardware support for the popular sampling rates such 
as 11.025k. However they do it in different ways, eg. the Sigmatel 
A/D converter always runs at 48k and converts the sample rate 
digitally, while the AD1818 seems to have a synthesiser that can 
tune between 7k and 48k samples/s in 1 sample/sec steps. This is 
all very interesting, but probably means that frequency oddities will 
continue to be a feature of soundcards.

For VE2IQ's "Coherent" software, I have one of G4JNT's PIC-
based interfaces - this performs the same function as the original 
VE2IQ interface, but is rather simpler, having just the PIC, a 5V 
regulator, and an audio preamp. It certainly seems to work 
perfectly well, and the crystal can easily be tweaked.

Cheers, Jim Moritz
73 de M0BMU