Generate coherent signals using two MG3692A signal generators

In high-frequency system measurement and application, it is often necessary to use multi-channel coherent signals. Generally, we can use devices such as power dividers, phase shifters, amplifiers and attenuators to divide the signal generated by a signal source into multiple channels, and adjust each signal to achieve the phase, amplitude, pulse delay, etc. of each signal to meet the requirements. However, due to the limitations of device performance, such as the frequency response of the phase shifter, we cannot arbitrarily adjust the phase and pulse delay of the signal within the required frequency band.

1. Generation of Coherent Signals

The two signal sources are named signal source A and signal source B. Use a BNC cable to connect the 10 MHz REF OUT terminal on the rear panel of signal source A and the 10 MHz REF IN terminal on the rear panel of signal source B, as shown in Figure 1. At this time, signal source A is the main source and signal source B is the auxiliary source. The so-called main/auxiliary is just a different reference. The auxiliary source will automatically detect the input of the external reference signal and lock it. Manually set the carrier frequency of the two signal sources to 100 MHz.

Figure 1 Schematic diagram of signal source connection

Then set the signal source B as follows:

Press the "Frequency" key (as shown in Figure 2), select the "Phase Offset" item, and make its soft switch display as pressed. At this time, three options appear on the screen: 1. "Phase Offset", 2. "Edit Offset", 3. "Zero Display". "Phase Offset" determines whether to offset according to the set angle. "Edit Offset" edits the size of the offset angle, with a minimum resolution of 0.1o. "Zero Display" can reset the current offset angle display to zero, and then the offset angle can be set based on this.

Figure 2 How to set the carrier frequency phase of the signal

After setting, turn on the RF output and use dual channel

Figure 3: Two signals with a phase difference of 0° and a carrier frequency of 100MHz

Figure 4: Two signals with a phase difference of 90° and a carrier frequency of 100MHz

2. Setting of pulse arrival time difference

Press the "Modulation" key and select the "Pulse" option to display the soft switch in the pressed state. Press the "Edit Period" key to set the repetition period to 200us, press the "Wdth/Dly List..." key to set the pulse width W1 to 50us, use the up, down, left, and right keys to select W1 and press the "Edit Selected" key to set the parameters. See Figure 5.

Figure 5 Setting of signal pulse width parameters

In addition, it is worth mentioning that MG3690A can select pulse train modulation in its pulse modulation, that is, each modulation cycle can select 1 to 4 pulse trains, and the pulse width and interval of each pulse in the pulse train can be adjusted independently, and the pulse widths are W1, W2, W3, W4, and the intervals are D1, D2, D3, D4 (as shown in Figure 6). This can simulate actual working conditions, such as multi-target echoes of radar systems. If the pulse width and delay of the pulse train of the main source and the auxiliary source are flexibly adjusted, more complex related signals can be generated.

Figure 6 Multi-pulse sequence parameter definition

Since we are only discussing conventional pulse signals, we only need to set W1=50us, and other parameter settings are invalid. The setting of signal source B is more complicated. Signal source B needs to be set to an external trigger mode where the pulse width can be changed independently. Press the "Modulation" key and select the "Pulse" option to make its soft switch display as pressed state. Press the "More" key below and select "Trigger...", press the up and down keys to select "Triggered w/delay", and then press "Select" to confirm. As shown in Figure 7:

Figure 7 Pulse trigger mode settings

When "Triggered w/delay" is selected, the signal source is set to external trigger mode. The signal repetition period is determined by the external reference signal, but the pulse width W1 and pulse delay D1 can be set. Using the method described above, W1 is set to 50us and D1 to 0us. This generates a radar signal with two pulses arriving at the same time. The real domain waveform collected is shown in Figure 8:

Figure 8. Two conventional pulse signals with the same pulse arrival time

When D1 = 20 us, the real domain waveform is shown in Figure 9. Oscilloscope channel 1 shows the output waveform of signal source B. It can be seen from the figure that its pulse arrival time is delayed by 20 us compared with the pulse signal of channel 2:

Figure 9. Two conventional pulse signals with a pulse arrival time difference of 20 us

It can be seen from this that the pulse arrival time of the conventional pulse signals generated by the two signal sources can be adjusted at will.

3. Problems and Improvements

1. When setting the phase offset of the carrier signal, we mentioned the "Zero Display" option. Its setting can make the phase difference display of signal source B (sub-source) relative to signal source A return to zero. The question here is: How can the user know when the phase difference of the two signals is truly zero? When does the user perform the "Zero Display" operation? Below we provide two feasible methods. For low-frequency signals, you can use an oscilloscope to directly acquire real-domain waveform signals, use the oscilloscope's measurement function as an indicator, and adjust the phase offset "Phase Offset" of signal source B. When the phases of the two signals are consistent, press "Zero Display" to return the display to zero. For high-frequency signals, you need to use a power divider and a spectrum analyzer of this frequency band. Connect the two signals to the branch end of the power divider respectively, and connect the output of the combined end of the power divider to the spectrum analyzer. The amplitude of the signal received on the spectrum analyzer will change when the phase offset "Phase Offset" of signal source B is adjusted. We conducted an experiment according to the connection above. When the two signals are in phase, the signal power received by the spectrum analyzer is 6dB greater than the signal power of a single signal, that is, the output signal power after the related signals are superimposed is 4 times greater than the output power of a single signal. When the two signals differ in phase by 180°, the signal power received by the spectrum analyzer is about 20dB less than the signal power of a single signal, that is, most of the signal output is offset. Using this property, the signal power received by the spectrum analyzer can be minimized by slightly adjusting the phase offset of signal source B, and then the phase offset is set to increase by 180 degrees on this basis. At this time, the phase display of signal source B can be reset to zero. Through the above two methods, using auxiliary instruments, the operation of adjusting the phase difference of two coherent signals to zero is realized, and on this basis, the phase difference of two coherent signals can be set at will.

2. In the setting experiment of pulse arrival time difference, when the oscilloscope scanning speed is increased, the time base is set to 50ns/grid, as shown in Figure 10. It can be seen that although D1 = 0 us (refer to Figure 8), there is still a slight difference in the pulse arrival time of the two signals. The pulse arrival time of signal source B lags behind the pulse arrival time of signal source A by 225ns. The reason for this lag is determined by the signal source itself, which we do not know, but signal B is triggered by signal A, so it is understandable that signal B lags behind signal A. At the same time, since the delay parameter D1 of signal B cannot be set to a negative number and can only be zero at the minimum, we cannot obtain two pulse signals with exactly the same arrival time through this method.

Figure 10: Two signals with the same pulse arrival time after the time base is reduced

Below, a setting method is provided to solve this problem. We do not change all the above settings, but only set signal source B to fall delay trigger, that is, triggering starts when the fall delay of the video signal of signal source A is received. At the same time, set the pulse delay D1 = 200-50-0.225 = 149.775us, and two regular pulse signals with strictly the same arrival time can be generated. The setting method is: press the "Modulation" key, select the "Pulse" option, and make its soft switch display as pressed state. Press the "More" key below and select "Trigger...", then select "Trigger↓". The setting method of D1 has been mentioned before, so it will not be repeated here. The acquired real domain waveform is shown in Figure 11.

Figure 11. Setting the falling delay trigger, two pulse signals with D1=149.775us, and observing their pulse delay

As can be seen from Figure 11, since the phase difference of the two signal RFs is 0 degrees and the arrival time of the pulses is strictly equal, the leading edges of the pulses are almost exactly the same. On this basis, we can also arbitrarily change the phase difference and pulse arrival time difference to meet the needs of different tests and measurements.