Good comments so far. Thanks. A few points:

1. The backup device MUST produce a log that's acceptable for U.S.
regional and national contests. As I read the rules, a handheld GPS
receiver is acceptable--even for a U.S. nationals--so long as it logs
altitude; a PDA-based system with a GPS card is not.

2. The most likely failure mode may well be a power problem, but I've
already got a backup power supply (that I've used 3 or 4 times over
the past 30+ years for genuine failures).

3. Regarding failure of the flight recorder:
a) the probability may be low but it's already happened once (the
UART comm. chip, I was told); to date, that's at least as frequent as
a power supply failure on a per hour or per flight or per year basis
b) the EFFECT is disastrous; if I'm going to drive halfway across
the country and spend the money to enter a nationals, I don't want to
zero a day because some $2 part fails. That's like tossing a live .45
caliber shell into a bin with 99 blanks, mixing them up, then picking
one at random, loading it into a pistol, aiming it at my head, and
pulling the trigger. The percentages are in my favor but the
consequences of that 1% chance are severe.

4. I can live without certain navigation and flight computer features
(I still carry a current sectional and a cardboard final glide
calculator) but if there are several alternatives and one allows me to
switch the source of GPS info for my LNAV and PocketNAV from my
Cambridge 20 to the backup device quickly and easily, it makes the
choice easier.

5. Gaps of up to 15 minutes are allowed, as are multiple incomplete
flight logs. So there is time to switch from one device to another so
long as (i) the primary flight recorder retains the flight log for the
portion of the flight preceding the failure point (not true,
apparently, when mine failed last year) OR the backup device has been
logging the flight from takeoff.

To date, Paul Remde's suggestion of a Cambridge 302A sounds easiest
(though not cheap). But that raises the same questions about
Cambridge's long-term (or even short-term) survival as were discussed
on this forum last year. What's the latest?

Chip Bearden