Tuesday, December 8, 2020

The RFI Menace And Reduced Noise Antennas


Having been in this game for 60 years, I can say that RFI and line noise has grown out of control, and especially since the advent of the cheap controller chip + home computers + digital communication + smart TVs + micro-electronics, et-al. And by "out of control", I mean that the crescendo of noise on the bands is becoming virtually impossible to identify and corral. Back in the 1980s when it started getting worse, it was still possible to identify sources and eliminate them using time-worn choke-suppression methods. Now, not so much. The genie is out of the bottle and it ain't going back in.

One of the best tools I have found to identify RFI is a spectrum analyzer. No, a $2000 unit isn't necessary. You already have one if you own an SDR receiver. I have an SDRPlay RSP1a, purchased at $119 U.S. and it's quite easy to take a look at any frequency from 10 KHz on up to see where the problem areas are. Spread a short wire across the floor in the house, connect it up, and you will see all kinds of mysterious RF. A pocket or portable sniffer receiver can work for this too but it's much easier to see the RFI's extent on an SDR receiver's spectrum display. The sniffer receiver is better used to locate the RFI.

I currently use my RadiWow R-108 as a sniffer receiver when walking around the house or property. This is used once the RFI "problem" frequencies are identified on the analyzer. Your Tecsun PL-380, PL-310, or other portable receiver can do the same.

RFI hash @ 500 KHz - the elevated noise floor is -85 dBm! NOT signals!


Let's go over the big RFI offenders to our DXing. The big offenders at my DXing home are:

My Hewlett-Packard 24 inch computer monitor. Huge, wideband, low frequency buzzing in a range across the VLF, longwave and lower mediumwave bands, particularly in the 300-900 KHz segment. The switching power supply creates some of this but the majority comes right off the screen's surface when the display is lit. Efforts to reduce this RFI have only been mildly successful, but luckily its range is only about 15 feet. The downside is the radios need to be within 15 feet of the monitor, particularly the SDR.

My laptop's switching power supply. I have a recent (2020) Acer Nitro 5, 15.6 inch, with AMD Ryzen 5 4600H mobile CPU. Huge, wideband, low frequency hash between 0 and 600 KHz. Virtually all of this disappears when running on battery only. You can't run on battery forever, however.

Old style fluorescent lighting, particularly the old 4 ft. shop lights. Best is to just keep them turned off.

Light dimmers. Don't use them. Keep them off or remove them.

LED light bulbs for house lighting. The bad ones create a high frequency hiss. Luckily the range is only a few feet, but the house is full of them now due to power saving measures. Use good quality LED bulbs. Philips has been highly recommended.

Low voltage lighting used in the kitchen. Lots of wiring through the walls go to a transformer box in the cellar. When the lights are on they inject an additional huge buzz at the lower end of the mediumwave band, peaking at about 550 KHz. The emissions from these range throughout the house. The condition is virtually eliminated by keeping the lights off.

A myriad of switching "chopper" style wall transformers. Some are much worse than others. Try to identify the worst offenders. I try to put all of these on power strips so I can switch them off when not in use.

Unknown sources of frequency spikes. Strong 10 KHz spaced spikes from 9 MHz to 16 MHz, peaking in the 9.5-9.9 MHz and 10.7-12.5 MHz area. This one is intermittent. It can last ten minutes or an hour or more, then disappears. I have not ruled out that this signal may be coming from the mains feed to the house.

**Note: this RFI source just above has been identified. It comes from a $2000 Fisher & Paykel kitchen refrigerator. Fisher & Paykel is a major appliance manufacturer which is a subsidiary of Chinese home appliance manufacturer Haier. It is a multinational corporation based in East Tamaki, New Zealand. In 2012, Haier, a major Chinese appliance manufacturer, purchased over 90% of Fisher & Paykel Appliance shares. Partial solution: wrapping the power line cord through two Workman RFC-1 snap ferrite cores has reduced the problem 50%. More cores have been ordered.

A new 43 inch Toshiba smart TV and DISH satellite box combo. Tremendously strong RFI, a high-pitched squeal in the LW and MW bands coming out of these boxes out to a 6-8 ft. radius, which then couples to lines. It might be possible to put these on a switchable power strip, but then you have the device reboot problem every time you want to use them. Satellite box boot time is often 5 minutes. That's a no-go.

Those are just the biggest offenders. Not mentioned is the RFI coming off the computerized de-humidifier in the cellar, the computerized water conditioning system, and the two computerized heat pumps hanging off the back of the garage.

So you can see the frustration. It's not practical to try to eliminate all of this RFI unless you'd like a lifetime career in RFI removal. I suspect this is the case almost everywhere.


Being a ham as well, I've experimented with just about every wire antenna you can imagine over the last 60 years. My days of winding power line chokes are over. Common-mode chokes, current isolators, et al, are the rage these days - these to reduce RF pickup on the feedline and to lessen the possibility of the feedline from becoming part of the antenna system. They can help, but they are a Band-Aid to the real problem. Why not lessen the noise in a different way? My solution is to build inherently quiet antennas which are resistant to noise, and feed them correctly. 

Three things are important.

1. Get the antenna well away and out of your house.

An end-fed longwire attached to your shack window fed with 15 ft. of coax across the floor isn't going to do it. If possible, on your lot, put the feed point as far away as you can. This, for starters, is one of the most important things you can do. Don't worry about cable feed length. Coax feed at mediumwave or even shortwave frequencies has minimal loss. 100 feet of the old 50 ohm RG-58 on mediumwave presents only about 0.37 dB signal loss, virtually unnoticeable. RG-6A TV coax, 75 ohm, is even less at about 0.28 dB per 100 ft. I use RG-6A here almost exclusively, as it is cheap and readily available through many suppliers.

So, get that feed point as far away from your house as possible.

2. If you can, choose an antenna that is basically a short circuit. What did you just say?

Loop antennas are essentially short circuits to high frequency impulse noise. Long wires, verticals, and dipoles are not. They are RFI magnets, and particularly so if they are not balanced antennas (the dipole is at least balanced). Much of the high frequency noise component of RFI is short circuited in the loop. Small loops are even better for noise suppression, but their drawback is they often need active amplification due to lower signal delivery. Loops work well when placed close to the ground and you don't need high supports for wires.

They can also be laid flat on the ground itself which reduces RFI even more. This is where our Loop-on-Ground antenna will come in.

3. Use an isolating transformer at the antenna feedpoint. Very important. Feed any antenna with a transformer-balun isolating device, even if it is naturally a 1:1 match. There must be no common ground connection between the coax feedline and the antenna, i.e., between the primary and secondary of the transformer-balun. The antenna should remain floating and the coax remain floating. This isolating-matching device does three things which help abate noise:

     1) Matching the antenna greatly increases received signal strength. Increasing signal strength often will raise the signal above the noise floor. Remember when receivers had preselectors to peak the antenna, which made the difference of hearing a signal or not? This is what a broadband matching transformer is actually doing - matching the antenna to the receiver across a wide range of frequencies.

     2) The transformer, at least the one we will use, totally isolates the antenna from the receiver, eliminating the direct wire connection and lessening RFI picked up by the antenna from transferring to the coax. Much of the RFI will be consumed in what I call the secondary, or load side (antenna side) of the balun, as it appears as a direct short to the high frequency component of noise.

     3) The transformer/balun reduces antenna loading because it presents a proper load impedance to the antenna. Loading down the antenna destroys bandwidth and lowers signal strength. Take a longwire for example. A longwire antenna has an inherently high feed impedance, generally 450 ohms, nothing near the usual 50 ohms of a receiver. With no matching device, the input signal delivered to the receiver is a simple resistance ratio. The signal is delivered through a 450 + 50 ohm series divider. The receiver gets 50/500ths of the available signal without the proper transformation. That's 1/10 of the signal being picked up by the antenna! No wonder my receiver can't hear!

The Balun One Nine by NooElec, a 9:1 balun

Balun One Nine on Amazon. NooElec makes a cool little 9:1 ratio balun transformer for about $15.


The Quarterwave Folded Monopole antenna. Everybody starts out in radio trying a longwire or dipole. These are huge noise magnets in RFI-prone locations. If you are an old timer you remember the folded dipole. It traditionally was a halfwave length antenna, like the dipole. It too is essentially a short-circuited antenna as it loops back on itself at the mirrored low impedance node, opposite the feed point. Another version of the folded dipole is the quarterwave folded monopole, a vertical, though it can be configured in other positions. It is half of a folded dipole. The quarterwave folded monopole is also short-circuited and is easily grounded as well. It's inherent impedance is 150 ohms at resonance (468/f-MHz), half that of the 300 ohm halfwave folded dipole, so if possible use a 3:1 matching balun to get to 50 ohms. If you don't have a balun, don't worry too much about using one on this antenna as the 3:1 matching discrepancy isn't that far off. If the antenna can't be erected as a vertical due to height restrictions it can be run as an elevated end-fed antenna of any length. Possible configurations are an end-fed inverted-V (feed end starts at ground, high in the middle) or an end-fed slanted wire (feed end starts at ground).

This antenna is essentially a transmission line antenna. Keep the wires parallel and anywhere from a quarter inch to an inch apart. Erected as a vertical, it has great low angle response for that extremely distant DX.

The LOG antenna, or Loop On Ground is another variation of the close-circuited loop only it lays flat on the ground. It is also best fed with a balun. A spool of 100 ft. of 18 gauge wire on Amazon will only cost you about $9. Lay it out in a square, 25 ft. to each side, and feed it at a corner. It is an excellent low noise performer, though with shorter lengths of wire the signal pickup is quite reduced. My 100 ft. length lying on the ground shows close to 15 db less noise than the 6x12 ft. flag antenna in the tropical band (60 meters), with about equal signal strengths. The difference is in the substantially better signal-to-noise ratio. A 15 dB reduction in noise while holding the same signal strength as the flag antenna is a 15 dB SNR improvement!

I've written an extensive article on the Loop-on-Ground antenna which might be of interest:

The Loop-on-Ground Antenna For The Noise-Challenged

The LOG antenna is somewhat directional, having a fattened hourglass pattern, with slight nulls at the feed corner and the corner opposite the feed. Both high and low angle reception are good, within its range. Best results are when the overall loop length is about 15% of a full wave for the frequency of interest. A 60 ft. total length works well for the 2-8 MHz range.

KK5JY has an excellent article on the Loop On Ground antenna, with illustrations. Be sure to check it out.

The Flag Antenna is a smallish but very efficient antenna especially for mediumwave work. It is usually configured in the shape of a rectangle and is easily ground-mounted if outside. I have a 6 ft. tall by 12 ft. long flag antenna erected indoors on the second floor, running east-west. The lower wire runs along the floor. Two 6 ft. fiberglass rods form the uprights for the ends. Although my house is very noisy with RFI, the noise pickup on this antenna is very low. The rectangle is broken at one corner on the floor nearest the radio, a vintage tabletop Allied A-2515. A 9:1 balun is used to match the antenna to a short 9 ft. length of coax feeding the receiver. Even un-amplified, this broadband flag has wonderful sensitivity from the AM broadcast band through about 6 MHz. On the mediumwave band, it is about the equivalent of a 4 ft. passive loop which is usually tuned.

The BOG antenna, or Beverage On Ground is a good choice if you have the room on your property. It is basically a very long wire laid on the ground (100 ft. or more) and may be terminated through a resistor to ground at the far end. Termination to ground gives it directional characteristics off the end. It is a variation of the classic Beverage antenna, which is usually a few feet off the ground.


For the AM broadcast band, feeding any of these low noise antennas to a pocket or portable radio is easy. I find inductive coupling best. Salvage a short ferrite rod or bar from an old pocket radio. Three inches in length is about right. Remove all the magnet wire from it. Using some solid, insulated telephone wire of about 24-26 gauge, wind about 15-20 turns close-wound around the ferrite rod. Solder or clip the two ends of wire from this coil to the coax feeder coming from the antenna, one to the center and one to the shield. Hold the ferrite close to the radio's internal ferrite which will inductively-couple the signal to the radio. The advantage over a passive loop here is you have a broadband antenna which does not have to be tuned.

Inductive pickup loop

Shortwave antenna coupling to the pocket or portable is more difficult. A simple clipped wire to the telescoping antenna can greatly increase the noise pickup. If your radio does not have an external antenna jack to safely connect the coax feeder with adaptors then you might have to perform surgery on the radio. Be sure to ground the coax shield to the radio's ground. In any event, be extremely careful if directly connecting outside wire or coax to these modern DSP radios. I cannot stress this enough. You can easily fry the inputs to them. I destroyed a $200 Sangean ATS-909X this way two years ago. Luckily it was still in warranty and I was able to get it repaired and reprogrammed.


Antennas that are not essentially short-circuited can work but be aware they will capture more noisy RFI. Above ground dipoles or end-fed longwires are two such types. Be sure to use a matching device in any case, which will help.

I hope this has been helpful to you. Please experiment!

No comments: