EZNEC says that a 66 ft dipole used on 3.8 MHz, fed with 450 ohm
ladder-line, will have an SWR of greater than 100:1. This can
lead to all sorts of undesirable effects including an almost
impossible to match impedance at the tuner. A practical rule of
thumb might be in order, e.g. mine = no more than 20:1 SWR on the
ladder-line.

Fact is Cecil I never pay any attention to VSWR, just complex numbers.
Anyway your comments made me re-analyze the problem, and I realize I made an
error in the transmission line and antenna tuner analysis. I thought the
transmission line loss I came up with was a bit low. The final numbers are
shown below. See if you agree with me; then explain where I went wrong, and
why the match will not work.

66 ft dipole, 30 ft high, #14 AWG: Input Z = 11.3 - j961 ohms.
50 ft of 600 ohm line: Input Z = 5.48 + j189.85 ohms, loss = 1.95 dB
Matching network: Shunt C = 296 pF, Series L = 22.8 uH.
(Obviously half the L for each side of a balanced.line.)
Max line voltage: 1.5 kW in = 3 kV.
Tuner loss: 0.44 dB.

Incidentally have you ever looked at a typical airport NDB site? For
example a 45 ft monopole on 350 kHz has an input impedance of 0.2 - j7054,
which is a VSWR of 4.9 million : 1 -- whatever that means -- virtually
touching the edge of the Smith Chart. I have even seen 5 mile approach NDBs
with 30 ft monopoles. Marine installations for 400 to 500 kHz operation
frequently had electrically very small inverted "L" antennas. Of course 5
S/m sure helped, but the losses in the tuners must have been significant.

73,

Frank