HF Noise Reduction Techniques: DSP, Blankers, and Hardware Solutions

HF noise is the thing that'll drive you away from the hobby faster than anything else. You build a nice antenna, fire up the radio, tune to 40 meters, and hear... a wall of garbage. Buzzing, clicking, hissing, and that weird rhythmic pulsing that turns out to be your neighbor's solar charge controller. Sound familiar?

This guide covers what actually works for taming HF noise, from the DSP tricks built into your radio to external hardware and the nuclear option of tracking down the source.

Understanding What You're Fighting

Before throwing solutions at the problem, you need to identify the type of noise. Different sources require different approaches.

Broadband noise (white noise, hiss) comes from thermal sources, galactic background, and the cumulative effect of millions of distant noise sources. Below about 20 MHz, external noise dominates over your receiver's internal noise floor in most locations. You can't eliminate it. You can only filter it.

Impulse noise (clicks, pops, repetitive crackling) comes from arcing contacts, electric fences, ignition systems, and power line hardware. These produce sharp spikes across wide frequency ranges. Noise blankers were designed specifically for this.

Man-made interference (RFI) is the big one in 2026. Switch-mode power supplies, LED drivers, plasma TVs, Ethernet-over-powerline adapters (they're the worst), solar inverters, and cheap USB chargers. This produces steady-state buzz or whine, often with harmonics spaced at regular intervals.

I spent three months tracking down an S7 noise floor on 80m that turned out to be a $12 LED floodlight on my own garage. Three months. The thing was 20 feet from my antenna feedpoint.

DSP Noise Reduction: Your First Line of Defense

Every modern HF transceiver includes DSP noise reduction. The implementations vary wildly in effectiveness.

How DSP NR Works

Most radios use one of two approaches:

• Spectral subtraction: The DSP estimates the noise spectrum (during pauses in speech or signal) and subtracts it from the audio. Works well for steady-state noise. Less effective for noise that changes rapidly.
• Adaptive filtering (LMS algorithm): Correlates the current signal with its own history to separate predictable (signal) from unpredictable (noise) components. This is what Icom's NR does in the IC-7300 and IC-7610.

Kenwood, Yaesu, and Icom all implement DSP NR differently. The IC-7300's NR is considered among the best in its price class. The Yaesu FT-DX10 uses a more aggressive algorithm that can introduce artifacts at higher NR settings.

Practical DSP NR Settings

Here's what I've found works across different radios:

• Start low. NR level at 2-3 out of 10 (or whatever your scale is). The goal is to reduce noise without distorting the signal you want.
• Adjust per band. 80m with heavy QRN might need NR at 5-6. 20m during quiet conditions might need zero.
• Combine with bandwidth reduction. Narrowing your IF filter from 2.7 kHz to 2.1 kHz for SSB drops noise by about 1 dB. Not much, but combined with NR it adds up.
• Don't max it out. High NR settings create a "watery" or bubbly sound that fatigues your ears and can actually mask weak signals.

For CW operators, DSP NR combined with a tight 250-500 Hz IF filter is incredibly effective. A weak CW signal buried in noise on a wide filter can jump right out when you narrow to 400 Hz and add moderate NR.

Noise Blankers: Handling Impulse Noise

Noise blankers (NB) predate DSP. They work by detecting fast-rise impulse spikes and literally blanking (muting) the receiver for the microsecond duration of each spike. Your brain doesn't notice the gaps.

NB1 vs. NB2

Many radios offer two noise blanker modes:

• NB1: Shorter blanking window, less aggressive. Good for spark-type interference and ignition noise.
• NB2: Wider blanking window, catches broader pulses. Better for power line noise and some types of switching supply interference.

Nope, noise blankers are not a magic bullet. They fail badly on dense, continuous interference because the blanking becomes so frequent that it destroys the signal too. If your noise blanker is on and the received audio sounds chopped or distorted, the noise is too dense for blanking to work. Turn it off and try a different approach.

When NB Actually Works

Noise blankers are most effective against:

• Electric fence controllers (that distinctive tick-tick-tick)
• Automotive ignition (passing cars)
• Arcing power line hardware
• Some types of thermostats and relay contacts

They do NOT work well against:

• Switch-mode power supply noise
• Continuous broadband hash
• Powerline networking adapters
• LED driver interference

I run NB1 at about 30-40% on my IC-7610 almost all the time. There's always some background impulse noise in my suburban location, and that light touch of NB1 cleans it up without side effects. NB2 stays off unless I specifically hear spark-type noise.

External DSP Units: The bhi and Timewave Approach

If your radio's built-in NR isn't cutting it, external DSP speakers and filter units offer additional processing.

bhi ParaPro EQ20 DSP Noise Canceling Speaker

The bhi lineup (made in the UK) uses adaptive DSP that processes audio after it leaves your radio. The EQ20 model (~$250) includes parametric EQ and adjustable noise cancellation.

Strengths:

• Works with ANY radio (including vintage rigs with no DSP)
• Adaptive algorithm handles changing noise well
• Parametric EQ lets you notch specific tones

Weaknesses:

• Introduces slight audio delay (10-15ms, not noticeable for SSB, but CW operators may notice)
• Processing happens at audio frequencies only (can't help with noise in the passband that overlaps with signal spectrum)

Timewave DSP-599zx

The 599zx (~$400, discontinued but available used) was the gold standard for years. Multiple algorithm modes, hardware bypass, and genuinely impressive noise reduction. If you find one used for under $200, grab it.

W6LVP Loop Antenna (Different Approach)

Not a DSP box, but worth mentioning: a small receive-only loop antenna can reduce noise more than any DSP trick. The W6LVP amplified receive loop (~$200) mounted away from your house often delivers a quieter receive signal than heavy DSP processing on your main antenna. The loop's small aperture and ability to null directionally make it remarkably effective.

Look, here's my honest take: DSP can only do so much. If your noise floor is S9, no amount of digital filtering will pull a weak S3 signal out of that. DSP works best when noise is moderate (S3-S5) and you need to gain a few dB of effective signal-to-noise ratio.

Hardware Solutions: Attacking Noise at the Source

The most effective noise reduction isn't at your receiver at all. It's at the source.

Common Culprits and Fixes

Switch-mode power supplies: Every wall wart, USB charger, and laptop adapter is a potential noise source. Replace noisy units with linear supplies for devices near your shack. Alternatively, snap-on ferrite chokes (Fair-Rite 31 material, multiple turns) on the DC output cable can reduce emissions by 15-20 dB.

LED lighting: Modern LED bulbs with cheap drivers are the #1 RFI complaint to the ARRL Lab as of 2025. Cree and Philips brand LEDs tend to be cleaner than no-name imports. Some operators have had success with ferrite chokes on the AC feed to problem fixtures.

Solar inverters: Grid-tie inverters, especially older SMA and Enphase micro-inverters, can produce harmonics through HF. The Enphase IQ8 series is better but not silent. Contact the manufacturer for firmware updates or RF filter accessories.

Powerline networking (HomePlug/G.hn): These devices intentionally radiate RF across HF frequencies. They are fundamentally incompatible with amateur radio. Period. If your neighbor has them, you may need to politely ask them to switch to MoCA (coax-based) or mesh WiFi. The ARRL has template letters for this conversation.

Ferrite Chokes: Your Best Friend

Ferrite chokes on cables entering and leaving your shack reduce both conducted and radiated noise. Key materials:

• Type 31 (MnZn): Best for 1-30 MHz. The go-to for HF RFI.
• Type 43 (NiZn): Good from 25-300 MHz. Better for VHF issues.
• Type 75: Optimized for 500 kHz - 30 MHz. Some hams prefer these on lower HF bands.

Snap-on clamps are convenient but less effective than wound toroids. For serious suppression, wind 5-7 turns of the offending cable through a 2.4" FT-240-31 toroid. Each turn adds impedance. At 7 MHz, 5 turns on an FT-240-31 gives you roughly 3,000-4,000 ohms of common-mode impedance. That's serious attenuation.

Choke Placement Strategy

Don't just randomly clip ferrites on things. Be methodical:

1. Identify the noise source (more on this below)
2. Choke the cable closest to the source first
3. If the source device has multiple cables, choke all of them
4. Your antenna feedline gets a common-mode choke at the shack entry point regardless

Finding the Noise Source

This is where most people give up. Don't.

The Breaker Panel Method

1. Turn off every breaker except the one feeding your radio
2. Listen. Is the noise gone? If yes, turn breakers on one at a time until it comes back.
3. Once you find the circuit, start unplugging devices on that circuit.

Seriously, this simple approach solves 60% of RFI cases.

Portable AM Radio Technique

A cheap AM broadcast radio (the older the better, with a ferrite bar antenna) tuned between stations makes a decent RFI sniffer. Walk around your house and neighborhood. The signal will get louder as you approach the source. The null off the sides of the ferrite antenna helps pinpoint direction.

Advanced: SDR-Based RFI Hunting

An RTL-SDR dongle ($30) with an HF upconverter ($40) and a small loop antenna gives you a portable wideband spectrum display. You can see the noise signature (spacing of harmonics, bandwidth) which helps identify the device type before you even locate it physically.

The ARRL's RFI page (arrl.org/radio-frequency-interference-rfi) has extensive resources, including how to file a complaint with your utility company if the source is on their equipment.

Combining Techniques: A Layered Approach

The operators with the quietest stations use multiple techniques together:

1. Eliminate sources in and around the house (ferrites, device replacement)
2. Antenna selection that minimizes noise pickup (loops, beverages, directional antennas)
3. Common-mode choking on the feedline
4. Moderate DSP NR on the radio (not maxed out)
5. Narrow filtering appropriate to the mode

Each layer might only give you 3-6 dB improvement. But stack five layers and you're looking at 15-25 dB total noise reduction. That's the difference between an S9 noise floor and S4-S5, which is the difference between "I can't hear anything" and "I'm working DX."

Next Steps

Start with the breaker panel test. It costs nothing and tells you immediately whether your noise problem is internal or external. From there, ferrite chokes on identified sources give the most bang for your buck. DSP and blankers help with what's left.

And if you've tried everything and your noise floor is still S9+? Consider a portable operation from a quiet location. Even a local park can be dramatically quieter than your home QTH. Sometimes the best noise reduction is distance.