> It may be.  But I think more likely it's generating some 
> other framing error, which the host is treating the same way.

I'm too lazy to go and find the exact definition but a break is
defined as mark (or space) for n milliseconds (where n is of the
order of a few characters worth).

It's really very likely that the PC *does* generate something
that *is* a break when it powers up. VT220s and VT420s that
I've tried do this too.

On the older processors, you needed Control-P to get into
console mode. (On the VAX 8200 break did something else
instead - I think it set the baud rate or some such!).

This is not normally a problem on any VAXstation or
MicroVAX system - unless you run them in alternate
console mode. In that case, for many of them (possibly
for *all* of them) there is no way to disable this
behaviour. It's wired into the console. It may even
be deeper than that (i.e. wired into the hardware).

> {\rant I really wish more people building systems like this 
> would take the trouble to distinguish between a break 
> condition and other framing errors.}  Not that that would 
> help people with existing systems, which don't make the 
> distinction, any....

This is DEC, remember. I bet it checks for BREAK or something
really close to BREAK. It just so happens that BREAK is not
at all uncommon when power comes (and goes ...).

Antonio
 
-- 

---------------
Antonio Carlini             arcarlini@iee.org