Production of Sweep Frequency Signal Source

When RF electronics enthusiasts make sweep frequency analyzers, spectrum analyzers and other tools, they will definitely use a sweep frequency signal generator. This article introduces the production of a sweep frequency signal source.

The simplest frequency sweep signal source is to use a selected periodic signal to control the voltage-controlled oscillator so that its frequency changes accordingly according to the period of the control signal. The control signal may be a sine wave, a sawtooth wave, or a triangle wave, and the waveform depends on the needs. There is nothing special about the voltage-controlled oscillator. It just replaces the device in the oscillator's frequency selection circuit with a voltage-controlled one. The current common practice is to replace C in the LC frequency selection circuit with a variable capacitance diode, change the junction voltage of the diode, and the capacitance of the diode changes, so the oscillation frequency will also change accordingly, becoming a voltage-controlled oscillator. The figure below is a very practical circuit.

The circuit parameters in the figure are suitable for working at 200-500MHz. If the frequency is above 500MHz, the 6.8P capacitor can be replaced with 3.3P, and above 800MHz, it should be replaced with 2P. As for the capacitor connected in series with the varactor, it can be selected within a large range according to the properties of the modulation signal. If the modulation signal is video, the figure above becomes a very practical FM TV modulator. If the modulation signal is a sawtooth wave, it becomes a sweep signal generator.

However, the use of such a swept frequency signal generator is very limited because the sweep width is too small. It is difficult to obtain a wider sweep source using such a simple oscillator.

For frequency-adjustable LC resonant circuits, the adjustable range is very limited. At lower frequencies, such as medium wave radios, the highest frequency can be more than 3 times the lowest frequency. At higher frequencies, the influence of distributed capacitance is greater, and this multiple gradually decreases, so the variable frequency made by an oscillator has a very limited range.
The relative range is not easy to increase, but the absolute range can be increased. For example, an oscillator of 1000-2000MHz has a minimum frequency of 1000MHz and a maximum frequency of 2000MHz. It is very easy to achieve an adjustable range of 1000MHz. Therefore, if a fixed 1000MHz oscillator is mixed with a 1000-2000MHz variable frequency oscillator, it is very easy to obtain a difference frequency of 0-1000MHz, as shown in the figure below.

It is easier said than done. If such a device is really needed, the workload of designing and making it is still not small. For product design, this workload is not a big deal, but it is a bit uneconomical for hobbyists to experiment. Fortunately, there are many finished products that can be used, such as the UHF local oscillator of a TV, which can at least guarantee an electrically adjustable range of (470+38)MHz to (860+38)MHz; the absolute adjustable range of the local oscillator frequency of the satellite receiver's tuner demodulator is larger, and most of them can guarantee an electrically adjustable range of (850+480)MHz to (2250+480)MHz. Even for old-fashioned tuners and demodulators, the local oscillator has a large adjustable range. For example, we have some very old-fashioned tuners and demodulators on hand now. The adjustable range of the local oscillator is (920+612) MHz - (1450+612) MHz, which is no less than 500 MHz. When combined with a suitable fixed-frequency oscillator, it should not be difficult to make a 0-500 MHz swept frequency signal source.

The first step is to test and select the tuner demodulator. First test the lowest oscillation frequency, connect the tuning voltage terminal to the ground, and the measured frequency is the highest frequency; at the highest frequency, connect the tuning voltage terminal to about 12V, and the measured frequency is the highest frequency. The reason why it is called "the highest frequency" is that increasing the tuning voltage can further increase the oscillation frequency, but the frequency sweep control voltage is often generated by an operational amplifier, and it is difficult for the commonly used operational amplifier to output a higher voltage. Moreover, the linearity deteriorates sharply when the frequency is further increased, which is also somewhat inappropriate.

I picked up a few of them and measured them. I found that their lowest oscillation frequency was below 1300MHz, and their highest oscillation frequency was above 2000MHz. Most of the adjustable range was above 800MHz. There is definitely no problem in using them to make a 0-500MHz signal source.

There is no crystal resonator with a frequency of more than 1300MHz, so I have to use a 670MHz one instead. The 670MHz is multiplied to 1340MHz, which is quite suitable. Its fundamental wave is 170MHz higher than the highest frequency of the signal source, which can be easily filtered out by a low-pass filter. The experimental circuit is connected as shown below.

At first, the experimental results were not ideal. When the tuning voltage was changed, two spectrum lines moved in a large range on the spectrum analyzer. One was the difference frequency we wanted, and the other moved in the opposite direction to the 0-500MHz line. It was estimated to be the difference frequency between the third harmonic of 670MHz (2010MHz) and the local oscillator of the satellite TV tuner and demodulator. After measurement and calculation, it was indeed the case. It seemed that the fixed frequency signal input to the mixer must suppress its third harmonic. This problem is not difficult to solve. There are still many old satellite TV first intermediate frequency filters on hand. The nominal frequency range is 950-1450MHz, which can just take out the second harmonic of 670MHz. After taking this measure, the 0-500MHz oscillation signal obtained was clean and the experiment was successful.

Looking at the 670MHz component at the output signal end, it is more than 50 decibels lower, so there is no need to consider it at all, not to mention other out-of-band components.

This experiment was done by connecting several unit devices together, which is not very practical. When I have time in the future, I will make it into a practical device. As for what signal to use to control this signal source and what kind of sweep source to make, I am afraid I can't control it.