Looking at these articles about stacking,
http://kyes.com/antenna/stackluge.html
http://www.signalengineering.com/ult...o_phasing.html
one worries about exact cable lengths to not get the antennas out of phase.

But say for a temporary set up, with two small UHF TV yagis, each on a
pedestal that can be moved around the lawn, it seems one can cut the
all three pieces of coax to any short length, and just move the
antennas about, until one gets the best picture, meaning they are then
also in phase, no?

I believe so. You'll find that there are numerous (in theory,
infinitely many) pairs of positions which will result in an in-phase
relationship. If you assume (without loss of generality) that the two
coaxes from the two antennas to the combining point are of equal
length, then placing the two antennas so that they are equidistant
from the desired transmitter will put them in-phase, and you can put
them anywhere along the equidistant line which is convenient. Putting
them too close together might result in interaction and pattern
disruptions.

Or, you can put one at a distance X from the desired transmitter, and
the other along a line which is (X + lambda) or (X + 2*lambda) or ...
away from the desired transmitter. This will also result in an
in-phase relationship.

If one coax is a bit shorter than the other, then the antenna with the
longer coax would go at distance X from the transmitter, and the other
at any position (X + delta + N*lambda) where "delta" is the
(electrical) difference in coax lengths and N is any integer.

Of course the driven element would be on the same side of each (but by
using opposite sides, the second antenna could be pointed near
directions we wanted to null out?)

No, I don't think that would work well. That would result in the
second antenna having maximum signal pickup for the signals arriving
from the "null" direction. The antenna pointed towards the desired
source would have a much lower signal pickup from the "null"
direction, and even if the signals from the two antennas are out of
phase you'd find that their amplitudes don't match, and they would not
cancel out well.

A better approach is to aim both antennas at (or nearly at) the
desired transmitter. Put them in positions as indicated above, so
that the signals from the desired transmitter are in-phase. Then,
slide one of them sideways along the equidistant line. You should be
able to find a position where the signals from the desired transmitter
arrive at the combining point in-phase, but the signals from the
undesired direction arrive at the combining point with equal intensity
but at 180-degree phase difference, and null out.

Once again, there are numerous combinations of antenna position and coax
length which can make this possible.

If you know the directions to the two transmitters/sources, and their
frequencies, and the velocity figure of your coax, you can
precalculate all of the lengths and positions and be very close to
"spot on" when you set up your antennas.

There was a very good article on this topic (dealing specifically with
FM radio broadcast reception) in Audio Magazine back in the first half
of the 1970s.

--
Dave Platt AE6EO
Hosting the 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!