I didn't know the 40673 was gone...
Most of commercial production has shifted away from leaded parts, especially the
expensive metal cans, unless they are required for military applications. They are
still available, though: Dan's Small Parts and Kits sells several types of NOS ("new,
old stock", meaning unused but not recent manufacture.) And I happen to have
a couple hundred in a jar somewhere...
The 40673 was a particular RCA part that the ARRL chose as a "standard" dual-gate
MOSFET because it was rated to 500 MHz, but especially below 100 MHz just about
any of the other similar parts will work, and they are often cheaper.
The BF998 is probably the easiest dual-gate MOSFET to find commercially these days,
and is capable of a lower noise figure than the older parts. If you are building a
printed circuit board it easily solders to 4 adjacent pads: I use a Dremel tool to cut
my boards and just put it at the junction of two cuts across the board at right angles.
There are still some leaded parts being made, including the J310 and 2N4416, which
are good general purpose JFETs. The SMD version of the J310 would be good for
common-gate amplifiers in place of the old leaded U310 (same chip as the J310 but
with the gate tied to a metal can that could be soldered to ground.)
In the frequency range you are looking at, there are lots of different parts that
will work. For ultimate sensitivity, once you get the local noise down, the critical
parameter is the noise figure of your preamp. The MAR-6 is rated at 3dB, which
isn't bad for a broadband device. If you happen to come across an old Plessey
SL-560 IC they are rated at 2dB or better, and good to at least 75 or 100 MHz.
(Some of us have rather eclectic junk boxes...)
A Norton amplifier can provide a good combination of wide-band performance,
low noise figure, and resistance to overload, but you might have to experiment
to find one optimized for your desired frequency range. Further information
here (and other locations around the web):http://www.qrp.pops.net/preamp.asp
I hope to stay below 88 MHz and away from the FM broadcast band, but constant signals are just horizontal lines on a spectrograph that you can ignore.
The problem isn't with them showing up on the spectrograph, but with them overloading
the preamplifier, even if you aren't intentionally listening in that range. When I put a
spectrum analyzer on an antenna, the FM broadcast band shows as a clear splotch of
RF, much stronger than anything else. You'll probably want a low pass filter to reduce
the FM BC signal strength ahead of any wide-band preamp. The same goes for other
signals in the passband: very strong signals can overload the preamp and reduce the
sensitivity to desired signals.
Preamps: I tried a cable TV amp and I could tell immediately that it was dead below 50 MHz...
Preamps for over-the-air TV may cut off at 50 MHz, but those for cable TV service often
work down to 5 MHz, as even the HF bands are used to pass signals.
I spent much of the day modeling a fat dipole and thinking about how to build one cheaply. I've got a 1/4 mile spool of galvanized steel fence wire, also some used network wire I can separate into individual wires. And a stack of useless "coaster" dead CDs. If I can melt or drill holes in them near the edge, hang them from the steel wire by 1 hole then put network wires running parallel in 7 other holes it looks good. I modeled one 14 meters long and it looks pretty good, but not perfect.
While the 14m antenna would have a reasonably low SWR from 30 to 60 MHz when used
with a 4 : 1 balun and 75 ohm coax, the radiation pattern isn't going to be suitable. It
will have multiple lobes and, on some frequencies, a null broadside to the wire. A length
of 14 FEET seems to work much better. If I model an ideal 8" diameter dipole, the SWR
is under 4 : 1 across the frequency range (well, actually 15' worked better) using a
300 impedance when calculating the SWR. Angling the wires out from the feedpoint to
a single wide spreader, then back together, likely will give better results than a fixed
diameter cage: that is the basis of the true fan dipole that I proposed previously, with
the wires in a single plane to make construction easier.