SHORTWAVE RECEIVER
10 kHz - 30 MHz
(2002)

KLIK HIER VOOR DE NEDERLANDSE VERSIE


The simple shortwave receiver for CW and SSB.

SSB and CW receiver 10 kHz to 30 MHz
This receiver was made with use of the experiences of my first receiver. This new one has some improvements and modifications: Enough ideas to justify the construction of a new receiver!

No AM reception
The receiver does not have AM reception. But if you want, you can receive AM in SSB mode. The advantage is that there is no distortion due to selective fading.
Speech sounds acceptable, just like SSB, music sounds like an old 78 rpm gramophone...


Block diagram

A single super, a ladder filter, no Mid Frequency and regular parts please!
Those are actually the basic principles for this receiver! A double super increases the risk of spurious signals, problems with mirror frequencies, extra filters and oscillators, all of which also need to be adjusted. We therefore choose a single superhet. When the MF frequency is chosen above 30 MHz, the mirror frequency is far from the receiving frequency. This is therefore easy to suppress with a low-pass filter and a simple selective input circuit. 36 MHz has been chosen here, but 48 MHz also works fine. We can then use cheap overtone crystals for the ladder filter. In the low-frequency section we apply some extra audio filters to improve the selectivity, especially for CW.
The VFO runs from 36 to 66 MHz, a special stabilization circuit ensures excellent stability. Due to the large range, the VFO is not exactly low-noise, but it is good enough for a very decent receiver.
The 36 MHz ladder filter is immediately followed by the BFO. No MF with possible oscillation problems and so on. All amplification and also the AVR (Automatic Gain Control) takes place at a low-frequency level. This is less critical in terms of design and construction and is also easy to test and check with simple means!

Lower Side Band and the Ladder Filter
What a shame that those amateurs still use LSB below 10 MHz. And they can't keep it up... Unfortunately, the sideband suppression of a ladder filter is much worse when the BFO frequency is at the bottom of the filter, as is necessary here for LSB reception. But we also have a solution for that. We start the VFO at 26 MHz instead of 36 MHz. For LSB reception we then tune this VFO below 36 MHz instead of above and we also have good sideband suppression for LSB. So there is no USB/LSB switch!


A ladder filter with 48 MHz crystals in USB mode, the BFO frequency is at the top of the filter. The upper trace is the filter curve, the lower the suppressed sideband.
The same ladder filter but now with the BFO frequency at the bottom of the filter as required for LSB reception. The sideband suppression is now much worse!

A closer look at the details
Do not attempt to copy this receiver exactly. Of course that is possible, but do not hesitate to realize your own ideas and improvements and use parts that you still have or that better suit your working method.


HF circuit

HF amplifier and preselector
Simple transistors, a tuned circuit, a low-pass filter for better suppression of the mirror frequency. There is also an extra notch filter for suppressing 36 MHz signals. There's not much to say about it really, it just works. Of course there is also the important input attenuator. Use this at 40 and 30 meters in the evening to prevent overload. The sensitivity is still excellent on these bands, even with the maximum 18 dB attenuation.

Mixer with transistors
My first receiver had an SL6440 IC as a mixer. Incredibly good! But that was not the intention here. I was curious about the behavior of an ordinary old-fashioned mixer with transistors, I also wanted to use standard parts and so this mixer was chosen. It certainly works well, but if there is anything to improve, I would start here. Replace it with an NE612 or better, an SL6440. You then have a double balanced mixer instead of a single balanced mixer. The intermodulation behavior, the suppression of 36 MHz signals via the antenna input and also the buffering of the VCO are better.
The 820 ohm resistor between HF amplifier and mixer may be omitted. But without this resistor, a signal of 10 dBm on the input of the receiver caused so much disturbance towards the VCO that it changed frequency.... These signal strengths only occur when you switch on your own transmitter on a second antenna.


The 36 MHz ladder filter adjusted for maximum sideband suppression, 40 dB at 600 Hz. The upper trace is the filter curve, the lower the suppressed sideband.
The 36 MHz ladder filter adjusted for optimal flatness. The sideband suppression is now less but still 25 dB at 600 Hz.

Ladder filter
The ladder filter is made up of 5 crystals of 36 MHz that are connected to each other via 40 pF trimmers. You probably need to have a bit of luck with the accuracy and quality, just like with everything else in life. With the trimmers you can adjust the bandpass for maximum sideband suppression or optimal flatness of the filter. The easiest way to adjust is with a noise source and an audio spectrum analyzer program via the PC sound card. The ladder filter is followed by two anti-parallel connected diodes as a limiter for strong interference peaks.

BFO and low frequency preamplifier
The BFO is very simple: two anti-parallel connected diodes. This mixer is also known as the Polyakov mixer or the RA3AAE mixer. It was chosen because the injection frequency is half of the 36 MHz signal or 18 MHz. A cheap crystal can now be used for the BFO. An oscillator with a fundamental crystal is much easier to realize and can be set to the correct frequency much easier than an oscillator with an overtone crystal. The proper functioning of such a mixer depends on the (sinusoidal!) oscillator signal. Adjust the 1k resistor to maximum audio level when receiving a CW signal.
After the BFO, an LF amplifier follows to amplify the signal a bit before the audio filters start adding some noise. By changing the value of the 47 ohm emitter resistor, the gain can be adjusted to your own wishes.


The audio part

Audio filters
We start with a low-pass filter of 2250 Hz to weaken the high tones and the noise from the preamplifier. At the output we then have an excellent signal for SSB reception.
For CW reception I have often used a fairly narrow filter in other designs. Works great and is definitely recommended if you want good selectivity. But this receiver is more intended for "easy listening": Set it to approximately the QRP frequency and listen to the various signals. That is why a somewhat wider filter has been chosen here, consisting of a 600 Hz high pass filter and an 800 Hz low pass filter. The -3 dB points are approximately at 500 and 1000 Hz, 2000 Hz is already 24 dB weaker, quite a nice filter for pleasant listening.

AGC and low frequency amplifier
The BF256 fet takes care of the AGC (automatic gain control) and also mutes the audio during transmission. The receiver contains mute circuits here and there and a side tone circuit to use it together with a transmitter. Feel free to leave it out, I haven't even tried it myself yet and can therefore not guarantee it will work properly.
After the FET follows the first audio stage IC5B, the volume potentiometer with two anti-parallel limiting diodes and the power amplifier LM386. IC5A is the amplifier stage for the AGC signal with the audio detector behind it. The AGC does not work with weak signals. I think an advantage is that the receiver sounds much quieter and does not fill the room with noise when no signal is received. You can modify the behaviour by changing the gain of IC5A (changing the 330k resistor between pin 1 and 2). If the AGC "pops" very annoyingly, a resistor of 470 - 680k from the input (470 ohm, 470k ohm, fet, 1nF node) to mass can improve the performance. This will change the setpoint.
The IF potentiometer is located in the low-frequency section (after all, there is no IF amplifier stage), but its operation can be compared to that of the IF gain controller in a "normal" receiver.


VFO

The most difficult part, the VFO
Indeed, I have now built three of them, but I always have to think hard to understand "how it worked again". The principle is derived from a design of an all band QRP transmitter from PA0KSB that has once was published in a newsletter. You will never find it in a professional design, it is only suitable for amateur use!
First a note: If it doesn't work properly, try installing a 4k7 - 2k2 resistor between pin 9 of IC1D and the +5V, especially if you use a TL072 instead of the TLC272. The set point is somewhat close to zero volts then.

How does the VFO work?
Many small tuning ranges (23.6 kHz at 10 kHz and 52 kHz at 30 MHz) can be tuned with the 10 turn potentiometer. So you have an excellent band spread of 2.36 to 5.2 kHz per revolution! Press the up/down switches to go to the next small tuning range. For large frequency changes you must set S1 to coarse (coarse) and you can fine tune with the 10 turn potentiometer. Without this feature you would have to press the up key 1000 times to change the reception frequency from 10 kHz to 30 MHz!!

The system is based on a frequency lock circuit with a sampler, a VCO (Voltage Controlled Oscillator) and a VXO (Variable Xtal Oscillator). Harmonics from the (VCO/16384) are locked to the VXO. The frequency of the VXO is 8867 kHz but you can choose a different frequency. The frequency variation of the VXO is approximately 7 kHz. If you want to calculate a tuning range you can use the following formula:

(VXO frequency variation) x (VCO frequency) / (VXO frequency)

So you see that the size of the frequency band that can be tuned with the 10 turn potentiometer depends on the VCO frequency.
The maximum output frequency of IC4 (here 4028 Hz) must be significantly less than the 7 kHz VXO variation for proper operation.
You do not need measuring instruments to adjust the control loop resistance (the 100 ohm potentiometer), this can even be done better by ear rather than with measuring instruments! Tune the receiver to a strong and clean CW signal at a low frequency somewhere below 3 MHz. Adjust the 100 ohm resistor so that you hear a completely distortion-free signal without "noise". Very small noises were more audible than visible on the oscilloscope!

The Up/Down switches of the VFO
Just an important note: The UP switch only works if the VXO is tuned to a high frequency, i.e. in the upper part of the 7 kHz range. The reverse applies to the DOWN switch, which only works if the VXO is tuned to a low frequency.
When you press a switch, the control loop cannot follow the rapid frequency change and will lock onto the next harmonic relationship of the shared VCO signal and the VXO. When you release the switch, the VXO slowly returns to its frequency and now the control loop can follow this slow change.
Sometimes it happens that the VCO locks (not very stable) when there is a ripple voltage on the control loop. You can solve this with some DC offset by turning the 10k potentiometer a little.


The frequency counter

Frequency reading
Any frequency counter that you can program with an offset of 36 MHz can be used. Even the simple circuit with only one 7 segment display... How does that work? We take the frequency 14062.3 kHz as an example. In MHz mode it will then display:

"1" is displayed for 250 milliseconds.
" " display is off for 80 milliseconds.
"4" is displayed for 250 milliseconds.
" " display is off for 80 milliseconds.
"0" is displayed for 250 milliseconds.
And after a measurement time of 500 ms the "140" is shown again.

In kHz mode it will then display:

"6" is displayed for 250 milliseconds.
" " display is off for 80 milliseconds.
"2" is displayed for 250 milliseconds.
" " display is off for 80 milliseconds.
"3" is displayed for 250 milliseconds.
And after a measurement time of 500 ms the "623" is shown again.

The circuit with the 74HC4066 has been installed to also read the frequency of a transmitter. The transmitter is intended to apply a DC voltage to the TX input together with an HF signal, which switches the frequency counter. However, that transmitter is still a future project, the housing has been ready in the cupboard here for years....

Some more adjustment tips
Adjust the trimmers of the 36 MHz reject filter for maximum attenuation of a 36 MHz signal when the receiver is tuned in the 10 meter band.
Adjust the 100 ohm adjustment potentiometer of the mixer for minimum noise at a reception frequency below 100 kHz. Adjusting the ladder filter is a bit more difficult. Start by adjusting the trimmers at the input and output for maximum reception and set the four other trimmers to half their value. Turn off the AGC.
Then connect the speaker or telephone via a 10k resistor to the sound card of the PC on which you have installed an audio spectrum analyzer program.
By detuning a CW signal (for example the 10 MHz clock frequency of the frequency counter) you can see how the level varies at the various audio pitches. You can also use a noise source, but then you do not know how the suppression of the unwanted sideband is.
Now the trick is to adjust the trimmers in such a way that you have a good filter curve and good sideband suppression. Don't forget to always adjust the 18 MHz BFO oscillator on the edge of the filter.
Well, after about 2 hours of tinkering I had a filter that is a bit optimized for CW and has a pretty good sideband suppression.


The SSB filter (green) and
the suppressed sideband (orange).
The CW filter (green) and
the suppressed sideband (orange).

Results
The receiver really works! Many enjoyable hours have been spent listening to this receiver. Although it is not as good as a real expensive commercial receiver, a lot of DX has been heard on all amateur bands. The selectivity and sensitivity are good, tuning is very comfortable. The CW and SSB filters work fine, the CW filter does not sound "narrow". The AGC also works well, although the control range is somewhat limited. With strong signals the HF attenuator must be used to get the signal within the AVRGC range.
Because the mixer is not very good, the HF generator must be used in the evening on 30 and 40 meters. But even with the maximum attenuation of 18 dB, the sensitivity on these bands is still sufficient. And the advantage is that your third intercept point also improves by 18 dB!

Of course there was also a mistake: the 56k and 10k resistors of the under-voltage circuit of the frequency counter were mixed up (the diagram is correct, but they are wrong in the photo). The result was that the EEprom values sometimes changed when the receiver was turned off.

Frequency3525 kHz10125 kHz28125 kHz
Sensitivity (SSB)-130 dBm-131 dBm-123 dBm
3rd IP (-50 / -100 kHz)-4.5 dBm-10.5 dBm+1.5 dBm

Supply current 100 to 150 mA

SOFTWARE FOR THE FREQUENCY COUNTER

"FREQRX02.ZIP" WITH "FREQRX02.ASM" TO PROGRAM THE FREQUENCY COUNTER

PHOTOS


Top view of the inside.


Bottom view of the inside.


HF preamplifier, mixer with trifilar wound coil on a plastic shaft (No ferrite!)
ladder filter, BFO with Polyakov (RA3AAE) mixer, LF preamplifier


The frequency counter with SMD chips.
Before the chips were mounted, thin wires were soldered to the pins.
While soldering such a thin wire, the pin was separated from the adjacent pin with aluminum foil.
The SMD chips were glued to a piece of wood (from a match) that was glued to the PCB.
Pieces of a glue stick from a glue gun were used to glue on the resistors, etc.
(Melt the pieces with the soldering iron).


Rear view of the receiver

Index PA2OHH