Direct Conversion Receiver: A Simple Radio That Really Works
If you have ever wondered why so many home-built CW and SSB receivers consist of just a handful of parts, the answer is often the direct conversion receiver. It skips the whole chain of intermediate-frequency amplifiers and filters that a superheterodyne needs, and turns the radio signal into audio in a single step. A beginner can build one in an evening, it works surprisingly well, and the same idea lies at the heart of many modern software defined radios. At the same time it has well-known weaknesses that every builder should understand. This guide explains how it works, where it shines, what its problems are and how to deal with them.
The Idea: A Mixer at the Signal Frequency
A superheterodyne receiver mixes the incoming signal with a local oscillator to produce a fixed intermediate frequency, which is then amplified and filtered. In a direct conversion receiver, the local oscillator is tuned to the same frequency as the signal. The difference between the two is zero, which is why the receiver is also called a zero-IF receiver or a homodyne receiver. What comes out of the mixer is the audio itself. If the signal is a CW carrier 700 Hz above the oscillator, the output is a 700 Hz tone. If it is a single sideband voice signal, the output is the voice at the correct pitch.
The mixer used for this purpose is usually called a product detector, because its output is the product of the two inputs. It is the same circuit that detects SSB and CW in superhet receivers in the last stage, and a direct conversion receiver simply puts it first. You can read the background in the article on the direct-conversion receiver and on the product detector. A complete receiver can be built with a band-pass filter at the input, a mixer, a variable oscillator, an audio amplifier with a low-pass filter and a speaker or headphones.
Why Builders Love It
The appeal lies in simplicity. There is only one conversion, so there are no IF filters, no image problem in the normal sense and no complicated alignment. The tuning is done by one oscillator, and the selectivity comes from the audio filter, which is far easier to make than a crystal filter at a radio frequency. A good audio low-pass filter of a few sections gives a bandwidth for CW of 300 to 500 Hz. The amplifier can be built with common operational amplifiers, and the whole receiver has few parts to buy. Many classic QRP transceivers began with a direct conversion receiver as the first stage of a project, and our guide to the joy of QRP operating shows the kind of station in which such a receiver is at home.
Performance can be good, too. With a well-designed mixer and enough audio gain, a direct conversion receiver can hear very weak signals, because the noise is set by the mixer and the first amplifier. The dynamic range of the best designs, such as those with a Tayloe detector, a switch-based mixer with very low loss, competes with that of a commercial radio at a fraction of the price.
Where the Selectivity Comes From
In a superheterodyne the narrow IF filter does the job of rejecting neighbouring stations. In a direct conversion receiver the same job is done at audio frequencies after the mixer. The bandwidth of the audio filter sets the bandwidth of the receiver. For CW a simple active filter centred on 600 to 800 Hz is enough. For SSB the audio response is limited to about 300 to 2700 Hz. This works well, but there is a catch: the mixer sees all the signals that reach it from the antenna, and the strong ones can overload it, however narrow the audio filter is. So a good band-pass filter at the front, as the first protection, is necessary. Our article on receive filter bandwidth explains why filtering ahead of the main gain matters.
The Sideband Problem
When the oscillator is at the signal frequency, the mixer cannot tell whether a signal is above or below it. A signal 700 Hz above the oscillator produces a 700 Hz tone, and a signal 700 Hz below it produces the same tone. Both sidebands are therefore received at the same time, and a station on the opposite side of the oscillator frequency is heard as well. For CW that means that each signal can be tuned in twice: once on each side of zero beat. For SSB the unwanted sideband appears as a second copy, folded onto the wanted one, so that two conversations on both sides of the carrier are heard together.
The cure is a phasing or quadrature design. The signal is split into two paths, mixed with oscillator signals that are 90 degrees apart, and after audio phase shifting the two outputs are combined so that one sideband adds and the other cancels. The cancellation of the unwanted sideband is typically 30 to 50 dB, depending on how accurately the amplitudes and phases match. This is the same principle as the image-reject mixer found in other receivers, and it is also the basis of modern SDRs, which perform the phase shifts by calculation. Our comparison of direct sampling SDR and superheterodyne receivers shows the relation to modern designs.
Known Weaknesses and How to Handle Them
Four problems appear in almost every direct conversion receiver. The table summarises them with the usual remedies.
| Problem | Cause | Typical remedy |
|---|---|---|
| AM broadcast breakthrough | Strong AM stations are detected by the mixer diodes or transistors | Band-pass or low-pass filter, better mixer, careful grounding |
| Hum and microphonics | High audio gain amplifies mains hum and vibration | Filtering of the supply, shielding, firm mechanical assembly |
| LO leakage | The oscillator radiates from the antenna and mixes with itself | Buffer stage, RF amplifier with good reverse isolation, shielding |
| Both sidebands received | Zero IF has no sideband selection | Phasing design with I and Q channels |
AM breakthrough deserves more explanation. Strong broadcast stations on the medium and short wave can reach the mixer with so much power that the nonlinearity of the circuit demodulates them, and you hear the programme in the background of your amateur signals, whatever the tuning. A good mixer with a high level of linearity reduces the effect, and a band-pass filter for the band you listen to removes the worst offenders. A simple low-pass filter in the audio line has no influence, because the detection takes place in the mixer. A related problem is the local oscillator leaking into the antenna and radiating, which can mix with itself after reflection from nearby objects and produce a DC offset that moves when you wave a hand near the antenna. A buffer or an amplifier between the antenna and the mixer, with good isolation in the reverse direction, solves this.
Audio Gain and the Noise Floor
The input signal of a receiver is very small, often a microvolt or less. After the mixer it appears as an audio voltage of a few microvolts as well, so the audio amplifier has to supply a very high gain, typically 80 to 100 dB in total, to drive headphones. At this level, every small effect matters: hum, noise of the first stage, flicker noise at low frequencies and instabilities from feedback in the amplifier. The first stage after the mixer should be a low-noise amplifier with a good noise figure, and the layout should keep the amplified signal away from the input. The low-frequency 1/f noise of semiconductors is a particular issue, since all the signal sits at audio frequencies where this noise is highest, and it gives the receiver a characteristic hiss at low pitch. Choosing a low-noise device and an audio tone of 600 to 800 Hz for the CW signal puts the wanted signal above the worst of the 1/f noise.
Improving a Direct Conversion Receiver
Experience gives a few reliable rules. First, put a band-pass filter at the input, matched to the band, and add a low-pass filter on the oscillator output to keep harmonics out of the mixer. Second, use a mixer with a high level of intercept point, such as a diode ring or a switching mixer, driven with enough oscillator power. Third, apply the first gain at audio frequencies with a low-noise amplifier, then the filtering, then the remaining gain, so that strong signals cannot overload the later stages. Fourth, power the receiver from a battery or a well-filtered regulated supply to avoid hum. And fifth, build the circuit in a metal box with short leads, with the oscillator in its own shielded compartment.
If a phasing design is your goal, remember that matching is key. The two mixers and the two audio channels should have the same gain, and the audio phase network should give a precise 90 degrees over the speech band. Trimmers help, and many designers adjust until the unwanted sideband disappears when they listen to a steady tone from a signal generator on one side of the oscillator. Our general guide to how single sideband works describes the sidebands that the phasing receiver separates.
A Practical Example: A 40 Metre CW Receiver
To see how the pieces fit, imagine a receiver for the CW part of the 40 metre band. At the input is a band-pass filter of two or three sections for 7.0 to 7.2 MHz, which also protects the mixer from the strong broadcast stations just above the band. After the filter comes a diode ring mixer or a switching mixer of the Tayloe type. The local oscillator is a stable VFO or a synthesizer, tuning 7.0 to 7.2 MHz, with an output buffer that lets only a small amount of power leak back toward the antenna. The mixer output goes to a low-noise audio preamplifier with a gain of 30 to 40 dB, followed by a low-pass filter with a cutoff near 1 kHz and a peak at 700 Hz, and then a power stage for headphones.
With such a receiver you tune the oscillator until a CW signal produces a pleasant tone of about 700 Hz. If you tune lower, the pitch falls to zero, the zero beat, and rises again on the other side, which is the sideband problem seen in practice. To be sure which side you are on, use the habit of always tuning from the same direction. This simple receiver, built on a small board, will hear stations across Europe on a decent antenna, which usually surprises people who think that sensitive reception needs a complicated radio.
Testing is easy as well. Use a signal generator or a known beacon, and measure the audio output voltage as you reduce the input signal. Find the smallest signal that gives an audible tone above the noise. Then touch the circuit gently with a finger or tap the case, and listen for microphonics. Wave your hand near the antenna to check for oscillator leakage. These simple checks identify the most common weaknesses before you decide on improvements.
Where It Fits Today
A direct conversion receiver is no longer the only simple choice. Cheap SDR dongles, as described in our overview of SDRs for ham radio, give far more flexibility for a small sum. But the direct conversion receiver remains the best teaching tool, because every block can be seen, measured and changed. It is also at the heart of the quadrature sampling front ends of many modern SDRs and transceivers, so the lessons carry over. For a home builder who wants to hear the first signals from a circuit made with his own hands, there is hardly anything more rewarding than a direct conversion receiver for 40 metres CW.
The Bottom Line
A direct conversion receiver mixes the signal with an oscillator on the same frequency, so the output is audio and the audio filter provides the selectivity. It is simple, cheap and sensitive, which is why homebrewers love it. Its weak points are AM broadcast breakthrough, hum and microphonics, local oscillator leakage and reception of both sidebands. A good band-pass filter, a linear mixer, a low-noise audio stage, shielding and, if needed, a phasing design deal with most of them. Build one for a single band, learn from it, and you will understand modern receivers much better.
Frequently Asked Questions
What is a direct conversion receiver?
It is a receiver in which a mixer and an oscillator tuned to the signal frequency convert the signal straight to audio, with no intermediate frequency stage.
Why is it called zero-IF?
Because the intermediate frequency is zero: the oscillator and the signal are at the same frequency, and the mixer output is the audio itself.
Why do I hear AM broadcast stations on my direct conversion receiver?
Strong broadcast signals are demodulated by the nonlinearity of the mixer. A band-pass filter at the input and a more linear mixer reduce the effect.
Does a direct conversion receiver receive both sidebands?
A simple one does, so each signal is heard on both sides of zero beat. A phasing design with I and Q paths cancels the unwanted sideband.
Where does its selectivity come from?
From the audio filter after the mixer. A low-pass or band-pass filter at audio frequencies sets the bandwidth, so no IF crystal filter is needed.
Is a direct conversion receiver good enough for the air?
Yes. Well-built designs, especially with a good mixer such as the Tayloe detector, hear weak signals very well, although strong-signal performance needs attention.