Owen wrote:

I took the figures published by Belden for their 8262 cable at
frequencies from 1MHx to 1Gz (9 points) and did a polynomial
regression to MLL=k1*f**0.5+k2*f, then substituted k2 into the
expression described in my original post.

I didn't measure the losses, and I recognise that Belden might have
smoothed their results by an intermediate regression, but I figure
that they are not going to exaggerate the effect of k2 unnecessarily.

I didn't realize that was the source of your loss figures. Reg is right,
you have to measure it. Another thing I found long ago is that the
published specs don't match reality. I recall finding Belden to be
conservative with loss, but that sure isn't the case with some others. I
just had occasion to very carefully measure the loss of some Davis RF
"Bury-Flex" which they specify as having 2.9 dB/100' at 400 MHz. It was
4.1 dB/100' at 400 MHz and extrapolated very nicely to at least 3.5 MHz
with a square root rule. (It has a solid aluminum shield under the
copper.) So there's some real specsmanship going on out there.

Roy Lewallen, W7EL