Tom,
Yes, they both "work" ... that's the problem. By the time the AR has thrown the
turnout (due to detecting the approach of the loco - a short) the keep alive
has run the loco far enough to derail (before the turnout has completed the
realignment). It's the "catch-22" of keep alives ... they aren't meant to keep
the loco running when there is a real short - only when the loco looses
power for just a bit (dirty wheels/track) ... but they do and so you still have
to line your turnouts ahead of your train.
And, specific to running around a reversing loop, there isn't any device
that I know of that allows you to do that in a "fire and forget" method. IF
you always ran the train the same direction and you left a long (1 foot or
more?) enough approach to the turnout and your turnouts completed
changing direction before the keep alive got the loco all the way to the
derailing point ... then it would work. Of course, that would use up some
of the track in the reversing section. Do we need that last 2 car lengths
in the reversing section ... possibly not - most guys would say "yes".
The other gotcha in this method is that if you don't stop the train before
it crosses into the point where it is detected ... you can't have the train
"stored" and run another train into the loop - because it will cause a
short. Many (most?) reversing sections are actually more than one
track in the loop in order to store more than one train at a time.
As I've already said - there just isn't an "automated reversing loop"
that works for all situations.
Some/many/most of the mfgrs of decoders are -finally- including a
setting (CV) that allows you to control how long the keep alive will
keep the loco running.
I, for one, simply don't understand why the keep alives have to
have such large storage capacities. There has to be a better
design for the keep alive itself that makes it work for dirty track
for "any" decoder (without a CV in the decoder).
- Jim