80 meter band CW transceiver with direct conversion receiver with sideband suppression.

Side band suppression with a direct conversion receiver
A superhet receiver with a crystal filter is a very good solution for single side band reception.
So why should we make a direct conversion receiver with side band suppression with the phase filter system? In the old days with electronic tubes, such a receiver was very complex due to the number of required tubes, big sizes and power supply. Nowadays it is easy, a few cheap op-amps and mixer IC's and that is it! I was curious about the performance of such a receiver and one of my intentions was to make it once in my life.
It is a nice receiver for home brewing. All processing is done on audio frequencies, making the construction not so critical. There is no complex or expensive crystal filter with its problems, no IF amplifier, no BFO oscillator, no mirror frequency that has to be suppressed. The VFO frequency is the same as the reception frequency. Therefore, no complex frequency counter with IF frequency offset correction is required, any simple counter is usable.
And a very big advantage is that you can use the same VFO for the CW transmitter without any extra mixers and filters!
When the design of a 74HC4066 mixer simplified the mixer problem considerably, it was time for me to start with the experimental construction of an 80 meter CW version of such a receiver.

A success!
And here already the result of the experiment so that you do not have to read the whole story to find that out.
It behaves just like a good superhet receiver without any spurious responses from mirror frequencies! Also the sensitivity is very good. Construction was easy with all good obtainable standard electronic components. Almost nothing is heard on the unwanted side band. Only some noisy signals from very strong commercial telex stations and some strong QRO amateurs, but very weak. It is really a pleasure to listen to this receiver. I am so enthousiast about the performance of the receiver that I also made a four band version. After a few years the mixer, side tone oscillator and audio preamplifiers have been modified to simplify and improve the design.

Side band suppression
A direct conversion receiver has less side band suppression than a superhet receiver with crystal filter. But do we really need the 50 to 60 dB suppression of a good commercial amateur receiver? To investigate that, I did an experiment on the 80 meter band.
An attenuator between antenna and receiver was set to a value that the atmospheric background noise just disappeared in the receiver noise, to eliminate its influence on the experiment. Then the 80 meter band was scanned with extra attenuators inserted and the results were noted down:

The conclusion is that the side band suppression should be better than 20 dB but between 30 to 40 dB is desirable. And the receiver has indeed 40 dB suppression, the first version had 37 dB a few years after the adjustment.

A very successful Barefoot Technology Project!

The phase method used here to suppress a side band

The phase method used here.
big diagram

The phase method of the 80 meter band receiver
The RF signal is split into two signals that are shifted 90 degrees out of phase (one plus 45 and one minus 45 degrees). Both are mixed to audio frequencies. The two audio signals are again shifted 90 degrees out of phase (again one plus 45 and one minus 45 degrees). When we add the two signals, one side band is in phase and added, the signals from the other side band are 180 degrees out of phase and substracted.
The RF phase shift circuits are simple RC combinations, the trimmers are adjusted for maximum suppression at 3550 kHz. The disadvantage of such a simple phase shift network is that the side band suppression is a little frequency dependent.
For the LF phase shift network I did not use the more common but complex polyphase network but a simpler version with operational amplifiers.
Both mixers are a CMOS switch. One 74HC4066 IC contains even four CMOS switches of which only two are used here! So a very simple and cheap solution for the mixer!


Circuit diagram of the receiver
big diagram

The receiver in detail

Preselector, RF preamplifier and RF phase shift networks
At the input, you will find the very useful RF attenuator. It is the main volume control. The preselector is a bandfilter tuned to 3550 kHz. The RF preamplifier has some gain, a high input impedance for the bandfilter and a low output impedance for the phase shift networks. After this preamplifier we have the RF phase shift filters. It are simple RC networks, both tuned with the trimmers for approximately 45 degrees phase shift. They also compensate for amplitude differences. For optimum settings for phase and amplitude, it is possible that one network is plus 55 degrees, the other minus 35 degrees as long as the difference is 90 degrees. Adjust it by ear, try different trimmer settings while adjusting the other at 3550 kHz while listening to a signal on the suppressed side band.
Usually, phase shifting is not done in the RF signal path but in the VFO. In that case, the VFO works at 4 times the reception frequency. The 90 degrees phase shift is obtained by dividers. The advantage of phase shift in the RF part instead of in the VFO circuit is that the VFO can work at the 4 times lower reception frequency and the RF phase shift network can also be used to correct for amplitude inaccuracies.

Mixers, LF preamplifiers and LF phase shift networks
The plus and minus 45 degrees phase shifted signals are mixed to LF frequencies by two mixers. These mixers are CMOS switches of a 74HC4066, very cheap and performance is good. Adjust the 5k potentiometer for minimum audio detection of strong broadcast stations.
To obtain the side band suppression, we also need phase shift networks in the LF part. The shifts have to be 45 degrees over the whole LF spectrum we want to receive. The networks in this receiver do their work with acceptable accuracy from 150 Hz to 5 kHz. However, for this CW transceiver only frequencies between 400 and 1000 Hz are important. I did use components with an accuracy of 5 percent with good results. The phase shift networks are a circuit with two op-amps and a few resistors and capacitors instead of the better but more complex polyphase networks.

CW filter and audio circuits
Both signals from the LF phase shift networks are added via 5k6 ohm resistors. Here the summation and substraction of the wanted and unwanted side bands happens. The wanted side band components are in phase on this point and are added, the unwanted side band are 180 degrees out of phase and substracted.
The audio CW filter has to bandwidths, a wide filter and a narrow filter. The wide filter is more pleasant and less tiring when listening for long periods, the narrow filter is very good when there is interference or for digital mode's like Feld Hell.
LF volume control is done by a switch as there was no place for a potentiometer. The switch lowers the gain of the next LM358 LF amplifier. The advantage is that not only the audio signal is decreased but also the noise of the LF amplifier. The diodes are limiting the peaks of the LF signal to 0.7 volt to prevent overloading of the BF256A mute switch. Take an =A= type for the BF256 as it has the lowest cut off voltage.
The LM386 is the final LF amplifier that has sufficient output power for a loudspeaker.

Modifications for SSB reception
The modifications are simple:
1. Two capacitors of 0.1 uF after the 470 ohm resistors at the output of the mixers IC1a and IC1b have to be replaced by 33 nF.
2. After the junction of the 5k6 ohm resistors at the output of the LF phase shift filters, the LF part is replaced by that described on the page of the shortwave receiver. Check the correct polarity connections of the elco with a voltmeter. If the LF gain is too high, increase both 5k6 resistors and that is it!
For SSB reception, Automatic Volume Control as applied in the LF circuit of the shortwave receiver is a very pleasant feature.


Circuit diagram of the VFO
big diagram

The VFO and RIT

Did I just mention that it is an advantage that the VFO frequency is the same as the reception frequency? Indeed, but as this VFO is also used for the transmitter, it should not oscillate on the working frequency. This will cause frequency instability during transmission. That is why the VFO frequency is twice the working frequency: 7000 - 7200 kHz and is divided by two. It has to be screened!
There is also a temperature compensation circuit with the NTC. It is adjusted with the 10 k potentiometer while tuned on center frequency. I did that by measuring the frequency in the evening when the room temperature was 21 C and in the morning at 15 C. The frequency drift was improved with a factor 3 by this circuit.

The RIT is activated by a CMOS switch of the 74HC4066 IC that also contains the two for the mixers. The RIT on/off switch enables you to determine the zero position of the RIT potentiometer. At zero position, there should be no frequency change when switching the RIT on and off. I also made an input for an external control signal (for example frequency shift for telex or frequency stabilisation purposes), but I never used it.


Circuit diagram of the transmitter
big diagram

The transmitter explained

1st driver stage
With the 1k potentiometer, the output power is adjustable between 0 and 10 watt. After this potentiometer, the signal is amplified by a BC547 transistor. The output circuit (10 uH coil and capacitors) is resonant at 3550 kHz. Adjust the 242 pF capacitor for resonance at that frequency (maximum output power). It is not critical, it has a wide peak. This driver is switched by the morse key via the BC557 transistor. The diode and 1 uF capacitor are added for a correct shape of the CW signal.

2nd driver stage
The second driver is a 2N3553 transistor. The 1k resistor damps possible oscillations. The 2x10 ohm emitter resistors are a kind of limiter to prevent overdrive of the stage.

Final RF amplifier
Of course you should not use such an expensive VHF transistor but for example a 2SC1969. I had a MRF238 transistor unused in the junkbox.
The 2x 12 ohm at the base create a low input impedance, important for a good stability. The 1 nF capacitor has a low impedance for higher frequencies, it prevents HF and VHF oscillations. And finally, the 2x 100 ohm resistors with the 2x 0.1 uF capacitors are a negative feedback circuit that prevent oscillations at frequencies below 1 MHz.
Due to the 6 element antenna filter, the suppression of the 2nd harmonic is 53 dB and that of the higher harmonics are more than 60 dB.

Antenna switch
The 90 pF trimmer is adjusted for resonance at 3550 kHz with the 33 uH inductance. Both anti-parallel diode pairs are limiting the RF signal from the transmitter to the receiver. The first diode pair is connected to a point with a very high impedance, so the current via the diodes at that point is low. The resistor of 12 ohm is a kind of fuse. Smoking if something goes wrong. The second diode pair is not necessarey. It is just there as an extra protection if something else in the circuit fails.

Simple frequency counter
The simple frequency counter is described somewhere else at this website. Six leds are used to display the frequency. One green led at Q7 is the in-band led (frequency between 3500 and 3600 kHz).
Add the value of the leds and you know the frequency. For example: Leds 3 kHz, 6 kHz and 50 kHz are on for 3559 kHz. If you tune a little higher in frequency, the 1.5 kHz led starts burning and you are at 3560.5 kHz, just above the QRP calling frequency. Turn a little backwards and you are exactly on 3560 kHz. A very simple circuit and it absolutely not difficult to read out the frequency.

Built via the ugly method (dead bug method). Parts are soldered at one side of he double sided unetched print.
The VFO has to be placed in a screened enclosure. The frequency counter is also screened with chicken wire.... The advantage is that you can make some modification or adjustments through the holes of the chicken wire.
Inductances are commercially available types looking like big resistors. Lx are wired 6 hole cores.
Do not use a 74HCT type but a 74HC type for the IC's!
I made a seperate transmitter and receiver, both in it's own enclosure. They are connected with plugs and cables. The RF signal from the VFO to the transmitter goes via coaxcable with BNC plugs. The other wiring is with screened audio cable.

Sensitivity: -120 dBm (0.2 uV) signals are readable
AM dynamic range: 85 to 90 dB (good)

Side band suppression:
3500 kHz: 42 dB
3550 kHz: 45 dB
3600 kHz: 41 dB
Note: a few years after the adjustment, the first version had more than 37 dB suppression.

Transmit power: max. 15 W at 14 V
Suppression of harmonics: Better than 50 dB

As said before, I like this receiver, side band suppression is good and I really do not have a need for a commercial superhet transceiver.


Direct conversion receiver with sideband suppression.

VFO of Direct conversion receiver in a screened enclosure!


The very simple but efficient frequency counter with 7 leds, only 3 IC's and it's own 5 volt stabilizer.
Screened with chicken wire...