Newsgroups: rec.radio.amateur.digital.misc
From: gary@ke4zv.atl.ga.us (Gary Coffman)
Subject: Re: 9600 help
Reply-To: gary@ke4zv.atl.ga.us (Gary Coffman)
Organization: Destructive Testing Systems
Date: Tue, 18 Apr 1995 17:10:49 GMT

In article  richard@waveg.demon.co.uk (Richard Sterry) writes:
>
>You may need to play around with equalisation to get two 9600 stations to
>'talk' to each other satisfactorily. Once you've done that, the problem
>starts all over again when a third station joins in the fun and games! 9600
>is 8 times faster than 1200, and sometimes I think it's also 8 times more
>difficult. Setting up both ends of a dedicated link isn't too bad; you can
>do it all 'on the bench' before taking the equipment out into the field.
>Tweaking several different sets of base station equipment by trial and error
>isn't nearly so straightforward!
>
>Keep the deviation modest, and well within the bandwidth of the receivers.

According to some lab tests written up in the May QST, better BER is almost
always achieved with Yaesu-Kenwood-Icom FM radios by using as *much* deviation 
as is compatible with maintaining the spectrum inside the allowable channel
bandwidth (In fact, deviations modestly in *excess* of 5 kHz are best). Most 
amateur FM receivers, if the EQ is bypassed, have acceptable characteristics, 
but most amateur FM *transmitters* truly suck for 9600 baud packet. That's 
because most use a synthesis loop design that limits low frequency amplitude 
and phase response of the audio channel that is critical to successful 9600 
baud operation. There are exceptions. The IC-820H, the Tekk-900, and the new
Azden PCS-9600D do their modulations differently. They modulate an 
intermediate oscillator *outside* the synthesis loop, and then hetrodyne
that with the synthesizer to achieve channel frequency. That gives near
ideal response, and netting several such radios becomes a cinch. Of the
three, the Tekk is by far the most cost effective (it avoids the synthesis
loop problem differently, it's straight crystal controlled). 

When you start raising the baud, using an RF system *designed* for the
mode becomes more and more important. Kludging FM voice radios into
data service often becomes an exercise in futility. That's why the GRAPES
design eschews such kludges and directly generates RF for 56 kb operation.
Since the responses are ideal for the baud, they interoperate beautifully
without any need to tinker. With synthesized voice rigs, even radios of
the same model often need considerable tweaking to get them to work together.
Normal component tolerances disturb the low frequency phase response 
sufficiently to make netting several a frustrating experience.

Gary
-- 
Gary Coffman KE4ZV          |    You make it,     | gatech!wa4mei!ke4zv!gary
Destructive Testing Systems |    we break it.     | emory!kd4nc!ke4zv!gary 
534 Shannon Way             |    Guaranteed!      | gary@ke4zv.atl.ga.us
Lawrenceville, GA 30244     |                     |