The current article refers to direct conversion receivers with applied regeneration. Before reading the current article, you should consider this very well written article by Charles Kitchin N1TEV about regenerative receivers and their advantages. In this page, I will focus on two main things:
A simple method that could be used for measuring the actual received frequency in a simple regenerative receiver, such as the one presented by N1TEV.
An operational process that can be used to select the wanted signal and manually distinguish it from the image one, in such a receiver.
These methods could apply directly into non-regenerative direct conversion receivers too.
To begin, let's consider a basic regenerative receiver shown below. Many radio amateurs do not really know how a regenerative detector is really tuned in different modes, so I will start from this, based on information from Wikipedia.
For AM reception, the gain of the regeneration loop is adjusted so it is just below the level required for oscillation. The result of this, is to increase the gain of the amplifier by a large factor at the bandpass frequency (resonant frequency), while not increasing it at other frequencies. So the incoming radio signal is amplified by a large amount, increasing the receiver's sensitivity to weak signals. The high gain also has the effect of sharpening the circuit's bandwidth (increasing the Q factor) by an equal factor, increasing the selectivity of the receiver, its ability to reject interfering signals at frequencies near the desired station's frequency. For AM signals the tube also functions as a detector, rectifying the RF signal to recover the audio modulation.
For the reception of CW radiotelegraphy (Morse code) signals, the feedback is increased above the level of oscillation, so that the amplifier functions as an oscillator (BFO) as well as an amplifier, generating a steady sine wave signal at the resonant frequency, as well as amplifying the incoming signal. The tuned circuit is adjusted so that the oscillator frequency is a little to one side of the signal frequency. The two frequencies mix in the amplifier, generating a beat frequency signal at the difference between the two frequencies. This frequency is in the audio range, so it is heard as a steady tone in the receiver's speaker whenever the station's carrier is present. Morse code is transmitted by keying the transmitter on and off, producing different length pulses of carrier ("dots" and "dashes"). The audio tone makes the carrier pulses audible, and they are heard as "beeps" in the speaker.
For the reception of single-sideband signals, the circuit is also set to oscillate. The BFO signal is adjusted to one side of the incoming signal, and functions as the replacement carrier needed to demodulate the signal
But, can regenerative detectors filter out the image signal?
Considering simple regenerative
detectors such as the one presented by N1TEV, when receiving AM, these circuits
can distinguish the desired AM signal from the unwanted ones within the band of
interest as they can reject interfering signals at frequencies near the desired
AM station's frequency, but they will pick up both sidebands of the desired AM
signal. When oscillating, to receive SSB or CW, they will receive both the
desired signal and the unwanted image (in SSB, the opposite sideband). This is
because, despite the fact that they achieve good selectivity due to
regeneration, they do not have anything like enough selectivity to reject the
nearby
image (opposite unwanted sideband). Note, this is the general case, but
there are some nice designs which have very smooth regeneration and can
effectively set the unwanted sideband out of puff and attenuate it, so
that they can effectively receive only the wanted one. However, in this
article I am talking about regenerative receivers in general.
Whereas in a single conversion
superheterodyne receiver (assuming 455KHz IF and no front-end RF filtering), the
image signal is 455KHz apart from the first local oscillator, in a direct
conversion or regenerative receiver the image signal is only a few KHz apart,
nominally 1-3KHz, depended on the mode of operation (CW/SSB).
Usually, in a single-conversion superheterodyne receiver the image is filtered
out by front-end RF filtering that "follows up" the BFO oscillator, so it does
not contribute to the noise, apart from internally generated image noise within
the mixer. This internal noise is still present in an AM detector, so there is
no 3dB advantage. Some crude superheterodyne receivers do not filter or
otherwise remove the image, so they suffer a 3dB disadvantage when compared to a
better superhet.
Now that you have a good understanding of the above, we can proceed to a simple method that could be used for measuring the actual received frequency in a simple regenerative receiver. As you go through the article, you will also find out an operational process that can be used to select the wanted signal and manually distinguish it from the image one.
Measuring the actual received frequency
Measuring the exact actual received frequency has a real value only on CW and other ON/OFF keying modes, or other modes that use single RF frequency for transmission. On SSB voice mode, the transmitted sideband is a complex function of the audio waveform that has modulated it, the human voice. There is no single frequency to measure, as there is no single frequency human voice. Apart from switching the carrier on and off, CW and CDW could also be transmitted by modulating an ssb signal with a single frequency audio tone. In this case, the RF frequency that will be produced during each audio tone switching will be of a single frequency as well and measuring the actual frequency has a real value too. An AM signal is composed of the carrier and the two modulated sidebands, so measuring the actual frequency has not a real value.
Before starting describing the method that can be used to measure the actual received frequency an important thing has to be mentioned. There is a confusion between displaying the frequency of a receiver and measuring the actual received frequency. Many of the direct conversion receivers just display the receiver's local oscillator frequency. Some modern superheterodyne receivers are "clever" and they compensate the local oscillator frequeny by the IF frequency and display a more accurate value. Nevertheless, all these methods actually fail to measure the actual received frequency accurately.
Even in the "clever" receiver case described above, the actual received frequency cannot be determined accurately. This is mainly due to the wider bandwidth of the IF filters used. For example, consider a CW received signal on frequency 1455KHz, the local oscillator on 1000KHz and the IF filter on 455KHz with 1KHz of bandwidth, with the peak of the band pass of the filter at 455KHz (ignoring the image for this example). If the frequency of the incoming RF signal changes a few 10s or even 100s of Hz, the display frequency of the receiver will still show 1455KHz. The only thing that changes in this case is the audio tone drift that will be listened, which denotes a change in the incoming signal frequency, eventhough your receiver display frequency has not been changed. If the IF filter is narrower the accuracy will be better, but never good enough. Furthermore, filter responce and local oscillator may be subject to frequency drifting as temperature changes and this can make the things even worse.
The only method I can think of, which can be used in regenerative and direct conversion receivers to measure the actual received frequency accuratelly down to the Hz, is continuously measuring the local oscillator signal frequency as well as the audio tone signal frequency and compensate for their difference to determine the received RF frequency.
This method is not affected by frequency drifting due to temperature changes because since there is only one RF mixer stage, any drifting in the local oscillator will present a drift in the audio frequency. These drifts will be analogous so their difference will always be the same. It is this difference the one that is needed to determine the actual received RF frequency accuratelly.
Using this method, an ultra stable receiver local oscillator is not needed, to determine the received frequency. Only accurate frequency counters are needed and these can be cheaply made with accuracy down to the Hz, using microcontrollers.
This method has a disadvantage though. The drifts of the local oscillator frequency and the audio frequency are known to be analogous, thus their difference can be calculated and it will be always the same regardless of the drift. Their difference is known but what is not known, is where the local oscillator signal lies. It can be below or above the received signal. If the local oscillator is below the received signal, the difference has to be added to the local oscillator signal to determine the actual received signal. If the local oscillator is above the received signal, the difference has to be subtracted to the local oscillator signal to determine the actual received signal.
There is a method that can be used to determine wheather the local oscillator signal is below or above the received signal. This method monitors the audio frequency change to determine the above. To illustrate this, let's consider the next example:
Let's assume that the local oscillator signal is at 100KHz and the RF signal at 101KHz. The audio signal will then be 1KHz and this will include the wanted 101KHz signal as well as the 999KHz image signal too. For the time being let's assume that there is no signal present in the image signal frequency, to make the example easier to illustrate.
Back in the receiver, the only things known to you, are the local oscillator frequency, which is 100KHz and the audio frequency signal (i.e. the difference), which is 1KHz. You do not know if the acual RF signal is at 101KHz or at 999KHz. To determine this, move your local oscillator frequency a bit on the "left" (lower frequency), say at 99.5KHz. In our example, the local oscillator will be moved 1.5KHz appart from the RF signal, producing an audio tone of 1.5KHz.
Now move your local oscillator frequency a bit on the "right" (higher frequency), say at 100.5KHz. In our example, the local oscillator will be moved 0.5KHz appart from the RF signal, producing an audio tone of 0.5KHz.
The lowering of the tone frequency when increasing the local oscillator frequency, can give you the indication that the RF signal is above the local oscillator signal. Thus, to determine the actual received frequency of the RF signal you have to add the difference (i.e. the audio tone frequency) to the local oscillator frequency.
In the inverse case, the lowering of the tone frequency when decreasing the local oscillator frequency, can give you the indication that the RF signal is below the local oscillator signal. Thus, to determine the actual received frequency of the RF signal in that case you would have to subtract the difference (i.e. the audio tone frequency) from the local oscillator frequency.