TenTec
[Top] [All Lists]

Re: [TenTec] Orion vs SDR 1K

To: tentec@contesting.com
Subject: Re: [TenTec] Orion vs SDR 1K
From: "Dr. Gerald N. Johnson" <geraldj@storm.weather.net>
Reply-to: geraldj@storm.weather.net,Discussion of Ten-Tec Equipment <tentec@contesting.com>
Date: Sun, 30 Apr 2006 10:38:36 -0500
List-post: <mailto:tentec@contesting.com>
On Sun, 2006-04-30 at 09:28 -0400, Toby Pennington wrote:
>      Lee, how about another update on your SDR?
> I'm wondering if the ultimate solution may be
> an SDR receiver but a companion analog TX (N.B.
> Orion does not have an analog transmitter, hence
> its problems with QSK, time jitter, etc).
> 
>                                          73,  Bill  W4ZV
> 
> There seems to be some improvement in the latency area with the SDR 1K 
> running a high powered computer and used with a foot switch for keying up the 
> rig, and then turning off the internal sidetone and using the sidetone from 
> an external keyer.  This may not be considered an improvement but more like a 
> work around. 
> 
> However,  there is still the problem of latency when working contests because 
> the SDR will always be slow to crank up and by the time it does the contact 
> is lost. 
> 
> The latency issue will never be solved by the current hardware,  and a new 
> rig will have to be produced which will be one to two years down the road.  
> There are a lot of smart people working on the SDR code and I believe in time 
> they will produce a really top performer.  The receiver already takes the 
> prize for the best specs of any rig on the market,  and I believe it will be 
> just a matter of time before the latency issue is resolved.  It's going to 
> take a new rig however,  the present hardware will not cut it.  
> 
> Toby  W4CAK
> 
Its possible that the fastest Pentium computer is slower than a
dedicated DSP chip and that using floating point rather than long word
integer is partly responsible for the latency. However there is a
fundamental rule of DSP that the more coefficients needed in a
computation for a fancier function, the longer it takes for signal to
come out.

On top of that fundamental rule, there are many ways to configure
filters that give different approximation qualities and each has its
advantages and latency to trade off.

While using a PC for the DSP is handy and cost effective compared to a
custom DSP board, the processor in the PC is not so well equipped for
doing DSP and the OS keeps taking clock cycles away from the DSP
functions to run I/O, the files system, and the user display. And the
most common of mudsoft operating systems is really a PC cycle hog
cutting way back on the power available for DSP operations.

One may help the DSP latency by going along the bottom bar of the
windoze screen and unloading every task taking up memory (and
occasionally demanding I/O service just see if its needed). One
definitely helps a stumbling OS by removing all those preloaded tasks.
Even windoze 95 runs reliably if there's nothing in the preload
background trying to catch some attention.

Where the SDR 1K gains in experiences of many open source programmers,
it looses in close coordination and in the few of those who work at DSP
programming full time.
-- 
73, Jerry, K0CQ,
All content copyright Dr. Gerald N. Johnson, electrical engineer

_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec

<Prev in Thread] Current Thread [Next in Thread>