On Jun 21 09:33 96, Poe Lim of 3:623/630 wrote:

PL> Hello Dave,

PL> On Jun 20 17:38 in a msg to All, Dave Hatch of 3:711/808 wrote:

DH>> Further to my last - bad conditions continue.  The Microcom still
DH>> holds. 

PL> As a matter of interest, try reducing the symbol rate, perhaps down to 
PL> 3200, then 3000 (S54). Perhaps not using the full bandwidth may help. If 
PL> it's someone you call, set a custom dialstring for that node.

So far, not a positive result.  The problem has packed up and left for the
nonce, but while trying to identify the situation I tried all kinds of odd
rubbish, symbol rate included.  The signal/noise of the statistics during
such tests is pretty poor, but I don't THINK it was significant.  (Which
goes contrary to every other thing I've seen.  Arrrgggghhhh!!!)

Current "maybe it's so" observations not fully disproven:   

VFC seems to hold, at least in the Microcom implementation.  It does seem
to hold -through- the problem, not avoid it, as I previously thought.  This
pretty well invalidates most of the theories I had previously come up with.

I have seen the E5 at 9600 sailing along during a horror case of the
problem.  Next time I both see it, and have time and inclination to play
with it, I'm going to revisit that.  I may revive the E5 as a fallback unit
for horror lines, as our Telstra "improvments" march on into the
brave new world.

Regards,
Dave Hatch.

--- Msgedsq 3.20