[said Larry]
>
> [log in to unmask] wrote:
>
> >This is of interest for Amtrak fans - the new "Acela Express" trains
> >have TGV-based locomotives, but traditional separately-coupled
> >coaches of Canadian design.  Current Amtrak plan is to start operating
> >these trains (at 150 mph max) in November between Boston and Washington.
>
> Speaking as someone who lives at the opposite end of that run from Mike,
> I would like to point out that in 1999 this service was also supposed to
> start in November (and several months in between).  In November 2001 I
> do have to go to Washington, so maybe by then...
>
Yeah, Amtrak is having some teething problems with the new trains.
I think they are serious about the November startup, though - see

http://www.amtrak.com/news/archive/atk00125.html

Delays like this are very common with this kind of new RR technology -
happens in Europe all the time. In the US it is so rare to introduce
new RR technology that we tend to overreact to break-in problems. We also
live in very impatient and politically poisonous times that any
glitch in anything Uncle Sam touches is cause for terrible invective.

AFAIK the teething problems have not been software-related, but more
with mechanical type things like wheel and brake wear.

From the safety-critical point of view, though, they did have one really
interesting computer+mechanical design flaw.

The coaches are about 4" (~10 cm) too wide. Normally this would not
be a problem, but these trains have an active-tilt mechanism for
going faster around curves. GIven the fixed distance between the 2
tracks in opposite directions, one must do a worst-case analysis
to be sure the whole system is safe.

The on-board computers control the tilt mechanism. Suppose that, in
a worst-case scenario, on a certain curve the train on the _inside_
track is tilted (erroneously) toward the _outside_ track (instead of
toward the center of the curve). And suppose a train is going by,
simultaneously, on the outside track and tilts properly. In this
unlikely, but theoretically possible, scenario, the trains would
not touch IF they were the right width. But because of the 4"
extra width, they could collide. As I recall, this analysis was
based on theory and simulation.

This analysis caused a HUGE flap between Amtrak and the contractor.
Needless to say, each blamed the other and the situation remained
fuzzy.

In the end, it was decided to simply adjust the mechanism so the trains
tilt a bit less. But this will also slow them down a bit, which is why
the NY-Boston timings will be just a few minutes slower. From what I
read, once the trains are running regularly they'll have empirical
data on the basis of which they can decide whether to notch up the speed.

My big problem with the new service is the _price_. It looks like they
are setting it quite high to test what the market will bear. They can
always lower it later, I guess.

Mike