> Lower current 24V rail, but with a lot of caps to store up charge
> This seems a bit dangerous, and by my calculations it would
> require quite a bit of capacitance to supply that much current over
> any reasonable timeframe

What might probably work is a cap to provide the pull-in current
and a PSU for the hold current, which can be a fair bit lower than
the pull-in current. I've not tried that with solenoids but it does
work with relays, for example nominally rated at 12V but will hold
at a lower voltage. Like this

http://www.piclist.com/techref/io/relays.htm

As for the power supply, I've just prototyped something similarish.
It's a switcher that takes 8-30V on the i/p and puts out 5V/1A using
an MC34063, and 3.4V/4A using an LM2574-Adj. By using the
appropriate topology a switcher can be made to do pretty much
what ever you want with any i/p voltage range. And you won't lose
a heap of energy as heat. I'd find out what the minimum hold current
is for the solenoid and work from there

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist