Ken Hornstein writes:
|> 
|> >When we first put the AMD79c30 on the Sparcstation-1- the data
|> >sheet did indeed refer to ISDN- but the intent was solely
|> >for Audio. I forget why, but the 79c30 isn't up for ISDN.
|> 
|> Um, the 79C30 is the chip in the ISDN phone on my desk (I looked :-) )
|> so it's _definately_ up to doing ISDN.  Whether or not the additional
|> stuff you need is available on your workstation is another question.
|> 

Well it probably uses something like a 68320, which is a 68k core
cpu with lots of hdlc controllers hanging off it, to ride shotgun
the AMD ISDN port...

I do know DEC had a devil of a time with the ISDN port.  When they
arrived, the system had a piece of tape over the ISDN port.  I knew
they were having problems with both the hardware and the software.
I think you could eventually get support from a layered product
on DU, but this was something like 2 years after the 3000/xxx systems
shipped.

I can imagine the problem is what others have speculated - without a
dedicated real-time cpu or a hdlc framer, you probably drop lots of
perfectly valid incoming frames, and eventually the switch or the
sender gets sick of this and hangs up.

BTW most of the ISDN telephone and modem vendors just buy call control
stacks from folks like Trillium or Telenetworks.  ISDN is just too
arcane a speciality to make it worth the effort to ramp up in-house
engineers on it and expect them to get it right in a reasonable amount
of time.

-- 
Dave Cherkus ------- UniMaster, Inc. ------ Contract Software Development
Specialties: UNIX Internals/Kernel TCP/IP Alpha Clusters Performance ISDN
Email: cherkus@UniMaster.COM  When the music's over, turn out the lights!