[UK-CONTEST] Data contests

David G3YYD g3yyd at btinternet.com
Tue Feb 1 05:02:06 PST 2011


Andy

Fair question. Simple answer my own experience.

End of last year I was writing a RTTY decoder for the PC. I noticed that 
while I was playing with the sound card settings that the sample rate 
for 11025 was not 11025.The actual rate was different, which meant that 
what I thought was 1275Hz Mark tone was not. It does not matter if you 
are AFSK on TX and AFSK on RX using the same sound card the different 
tone frequency will apply to both TX and RX so no worries. However using 
FSK there would be be a small difference for the sound cards I am using 
here.

It is possible depending on how well the drivers are written to have a 
different TX and RX sampling rate at 11025 on the same sound card, but 
unlikely.

For receive I use two sound cards one for each rig but only TX on one 
sound card. I do SO2R RTTY. (Actually SO4R as use second RXs on both 
K3s.) What I found was the mother board sound card and the Creative PCI 
sound card do have different 11025 sampling rates. But when I set both 
to 12000Hz sampling rates they are identical as far as I can tell. Any 
difference will be down to the clock crystal in each card.

When I measured the motherboard sound card it would appear the nominal 
11025 sample rate is 11100. This would make 1275 1% higher in frequency 
or almost 1278 but if using 2125Hz then it would be almost 2130 a 5Hz 
offset. Does not sound much but an optimised 45.45 baud tone filter is 
only 46Hz wide at -6dB and if using FSK means RX and TX frequencies will 
appear to be 5Hz apart.

My home made DSP decoder can copy on a single tone unlike MMTTY with 
about a tenth of the processor load. Overall performance is very close 
to optimised MMTTY sometimes better sometimes worse depending on the 
type of QSB. The AFC is better than MMTTY and the filters are already 
optimised so no fiddling around with values. Judging by the available 
variables in MMTTY the "circuit" architecture used is non-optimum. If 
anyone is interested in my decoder's architecture please contact me off 
list. I did it as a learning project in DSP Windows programming. The 
hardest part was understanding how Windows works rather than the DSP 
bit. Really exercised the "little grey cells" to borrow a phrase and was 
great fun. I did have a really excellent tutor, Peter G3PLX.

73 David G3YYD



On 31/01/2011 21:29, Andy Swiffin wrote:
>>>> On 22/01/2011 at 10:36, in message<4D3AB33D.9090105 at btinternet.com>, David
> G3YYD<g3yyd at btinternet.com>  wrote:
>> On the sound card settings do not use 11025Hz sampling rate as modern
>> sound cards will not provide an accurate rate. It is better to use
>> 12000Hz. Modern sound card sample at 48KHz or integer multiples of
>> 48KHz. The driver software then converts the 48KHz to the requested
>> speed. As a result 12KHz is easy to do by sending every 4th sample to
>> the decoding software. Generating 11025 from such a card will result in
>> a compromise rate, which could be significantly different. This does
>> result in TX/RX baud rate being incorrect along with the received tone.
>> If using FSK this can mean you call people off frequency another reason
>> for using AFSK rather than FSK.
>>
>> The baud rate setting should be 45.45 and not 45.
>
> Hi
> I've been doing a little investigation on this and submit my findings in case other people might be concerned.
>
> I use the default 11025hz sampling rate on mmtty (+n1mm) on my laptop.  I was a bit concerned about the above even though I've generally noticed when running in contests that people do seem to call me back dead on frequency.
>
> I did a few tests on the bands, calling CQ and seeing where people called me,   they all seemed to be right on my frequency (including one nice chat with well known rttyer Mike K4GMH who certainly knows what hes doing).
>
> So, certainly in my case on my laptop,  11025 doesn't seem to present any problem with a frequency offset between using FSK for transmit and the audio coming in for receive.
>
> I'm posting this in case anyone else is thinking they might have to change settings to get this right,  I certainly don't have to.
>
> David - do you have any references to the information you post above,   I did some google searches, but couldn't come up with anything?
>
> 73
> Andy
> gm8oeg
>
>
>
>
>
>
> ************************************************************
> Please consider the environment.  Do you really need to print this email?
>
>
>
> The University of Dundee is a registered Scottish charity, No: SC015096
> _______________________________________________
> UK-Contest mailing list
> UK-Contest at contesting.com
> http://lists.contesting.com/mailman/listinfo/uk-contest


More information about the UK-Contest mailing list