On Jan 26 23:26 96, Deon Attard of 3:640/771 wrote:

DA> Hello Dave!

DA> Wednesday January 24 1996 18:35, Dave Hatch wrote to Deon Attard:

DH>> On my Hayes v34, it's in S108, labeled Signal Quality Select, range of
DH>> 0 - 3, with 0 putting up with anything, and 3 demanding perfection on
DH>> pain of instant train down to the next lower speed.   That's what the
DH>> manual sez.  The practice is a bit gentler - under some circumstances
DH>> you can't see the effect at all.  You need a noisy line, under some
DH>> pretty precise ranges of "noisy" before it is an
effective tool, IMHO.

DA> I'll check it out and post my results..  So 0 would wanna train up all the
DA> time, and 3 train down?  Like 0 is for speed, 3 is for no errors or 
DA> something?

According to the manual, it's a "noise tolerance" sort of thing. 
0 will put up with almost anything, 3 will insist on a retrain on the
slightest sign of trouble, and usually train down at that - and the rest
come in order in between.  Except that it's no where near that savage in
effect in my observations and experiments.  It's hard to find a line with
"just the right amount" of noise for the parameter to produce
visible differences, according to me, or at least according to the
particular units I have.

Regards,
Dave Hatch.

--- Msgedsq/2 3.20