I agree that a random length dipole fed with ladder line makes a great
antenna. I've used one for years with a traditional tuner. My
question: If I put a balun between the ladder line and the tuner, can I
use one of the automatic tuners built into modern rigs? If this works,
it provides the advantages of one simple antenna for multiple bands
without the hassle of having to retune when changing bands/frequencies.

Depends a lot on the radio and on your individual installation.

The ATUs built into a lot of modern rigs are of the "line flattener"
persuasion. They're intended to be used with an antenna which isn't
too awfully far from a resonant 50-ohm load. As a rule of thumb, I'd
say that most of them can match a 3:1 load, some of them will cope
with most loads of up to 5:1, and few of them can handle 10:1 loads
at all well.

For what it's worth, the ATU in my Kenwood TS-2000 won't even attempt
to match anything above 10:1. It'll struggle with a lot of loads
between 5:1 and 10:1, depending on whether they're low-Z, high-Z,
and/or substantially reactive. I suspect that this ATU is probably
fairly typical of modern rigs.

External tuners often have a substantially wider matching range than
an internal line-matcher, and probably have significantly lower losses
when handling difficult loads.

An unbalanced tuner plus a robust balun is probably going to work
better than a rig's ATU plus a balun. The balun can be a problem in
either case, with difficult (high-Z) loads - it's not easy to build a
balun which has a high enough choking reactance to really balance out
the line currents well if it has to work into, say, 5000 ohms or so.
Link-coupled tuners seem to be a better technical choice for such
difficult loads, although (as per your comment) they aren't the most
convenient beasts in the world.

If your doublet length and feedline length leave you with reasonably
tolerable in-the-shack feedpoint impedances on the bands that you care
about, then you might want to consider a sort of hybrid approach. Use
a robust balun to connect to the feedline, and feed the unbalanced
side of the balun to a (bandswitched) set of L networks. You'd want
one L-network per band, selected to bring the impedance down to
somewhere in the 3:1 SWR range (or so) in the band center. The output
of the L networks would go to the transceiver.

With this approach, the L networks would perform the "gross" tuning of
the antenna feedpoint Z, and bring it down to the point at which the
transceiver's internal ATU could do the rest of the matching across
the full width of the band.

Since the L networks wouldn't need to provide an exact match for a 1:1
SWR, selecting the component values and tuning the networks would be
simplified.

--
Dave Platt AE6EO
Friends of Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!