Optimising weak signal HF reception:
Good weak signal reception is achieved by improving the signal to noise ratio (S/N) of the received signals. This is not necessarily by having a 'higher gain antenna' (more 'S' in the S/N equation), but also by 'reducing noise' (lower 'N' in the S/N). The main ways to reduce noise are:
... suppress or remove local noise sources
... locate the receiving antenna away from local noise sources
... use a directive antenna system
QRN from appliances in the house and shack can often be reduced with clip-on ferrite cores (using HF rated ferrite material) on power cords and data/USB cables. Use a systematic trial process. Common mode chokes on coax feed lines (again made with good HF rated ferrite material) can reduce QRN 'feeding into' the transmission line. Antennas should be located away from the house ... e.g. away from what little man-made noise is left after suppression. Having done those steps here, I can now plug an antenna into a radio and see just 2 or 3 dB increase on the WSJT-X noise graph ... e.g "no noise", as it used to be in the 1970's. Noise from neighbouring properties can be a problem, but locate antennas away from them and use directional antennas to null out their unwanted noise.
The last factor in my quest to improve weak signal reception was to learn about antenna Receive Directivity Factor. 'RDF' quantifies the ability of a directive antenna's radiation pattern to suppress noise from unwanted directions (less "N" in the S/N equation). Search the internet for 'Receive Directivity Factor' and read the notes from K7TJR, W8JI, and others.
I compared many antennas while tracking 'Pico Space' balloons and their 20mW HF telemetry transmitters. At first I thought 'more gain' was needed, but after reading about 'RDF' I decided to try a Beverage antenna 'scaled down' for 20m. It 'worked like magic' ... my reliable pico balloon tracking range extended from 3000km out to 8000 - 9000km.
The Beverage antenna works so well that they are used here almost exclusively for HF receiving. On-air A/B receive tests against dipoles, Vee Beams, Flags, and wire colinears gave results in favour of the Beverage ... sometimes by as much as 15dB better S/N! Even a single wire Beverage that is just 1 wavelength long is better than those antennas when it comes to receive performance.
The Single Wire Beverage Antenna as used at ZL1RS:
Easy to make, easy to erect, great RX performance.
Keep in mind that this Beverage antenna is for receiving weak signals from one general direction on frequencies between 7 and 22 MHz, and has been designed, sized and erected accordingly.
Beverage wire size - 0.5 to 2 mm diameter (not critical, just physically strong enough). I use 0.7mm dia because I have plenty
Beverage wire length - between 1 WL and 4 WL long (not critical). Longer = better S/N but narrower beam width
Beverage wire height - 60 to 90cm (not critical), supported by tread-in electric fence posts (I have also used wooden 'tomato stakes')
Feed point transformer - 1:9 impedance ratio (1:3 turns ratio) on a binocular or toroidal ferrite core (not a powdered iron toroid) *
Common mode choke - e.g. 25 or so HF rated ferrite sleeves slipped over the coax jacket near the feed point **
Termination - 400 to 500 Ohms (not critical). I've used 3 x 1.2K Ohm resistors in parallel or a single 470 Ohm resistor
Preamp (not usually required) - I sometimes use a home made W7IUV preamp (the "13Sep09" version)
Feed line to shack - any 50 Ohm 'radio' coax or 75 Ohm 'TV' coax (not critical)
* 73 ferrite material seems to work best at low HF frequencies, and 43 material on the higher frequency HF bands. Keep the two sides of the transformer separate ... do not connect the coax braid to the earth stake! Just connect the feed line coax across the transformer low impedance winding, and the Beverage wire and ground stake to the high impedance winding. (Galvanic isolation? ... anyway, it reduces the noise and improves the SNR).
** or 12 turns of "Mini178" miniature Teflon coax wound on a Fair-Rite 43 ferrite toroid core p/n 5943004901 which provides a compact common mode choke for receiving with a broad band response of about 2k Ohms from 5 to 20MHz, peaking at over 3k Ohms around 10 MHz. Even better (especially for 40m and 30m) is 9 turns of RG-316 Teflon coax wound on a Laird 28B0735-300 core (very tight fit) ... this provides >5k Ohms of common mode choke impedance from 5.5MHz to 12.3MHz.
You may have noticed a liberal use of "not critical". All the Beverages tried here have been made from materials laying about the shed and from the junk box ... total monetary investment = zero! However, a quiet countryside location with several acres of space to erect long wires helps immensely.
The important thing is to not earth the feed point end of the coax and use a common-mode choke ... both reduce noise.
Alternative hand drawn image
I've used ground radials on the termination resistor earth side, and alternatively used another earth stake as the termination ground (no radials). Both seem to work (good F/B observed).
The feed point end. A galvanised pipe is used as a feed point support and ground stake.
The view from the feed point towards the termination.
The termination end. A single 470 Ohm resistor works just as well.
One of the tread-in electric fence stakes and a view back towards the feed point.
The view across 10km of rolling farmland to some low hills 30km away, and on towards the southern Pacific Ocean & South America.
A single wire Beverage works very well ... but a pair of phased Beverages works even better!
Comparison of a single wire Beverage with a phased pair of Beverages.
The single wire Beverage antenna is a very good antenna, but a pair provides more signal on the 'nose' of the radiation pattern and less noise from other directions thus improving the S/N. I did experiment with a phased Beverage array but no longer use it ... it performed extremely well, but was a lot of work to change direction or to remove/replace when cattle came to graze on the grass.
The following screen shots are of QRPp "Part 15 HiFER transmitters" from the USA (9,000 to 11,000 km away) as received on the phased Beverage array. They are typically 5mW transmitters operating in the ISM band between 13.555 and 13.558 MHz, using a 1/4 WL vertical or 1/2 WL dipole antenna, and operating 24/7. Spectrum Lab screen shots taken on 08 December 2015 at about my sunset.
Transmitter "EH" 5mW QRSS FSK CW from FN31nh East Haven, CT
Transmitter "WM" 5mW(?) QRSS FSK CW from EM68es Lawrenceville, IL
Transmitter "PBJ" 1.8mW QRSS CW from EM78cf Tracy City, TN
Transmitter "RY" ?mW QRSS CW from FN43sv Raymond, ME
(The waterfall scroll interval rate was changed part way through).
Multiple transmitters decoded