Null Detector

I designed and built a null detector circuit using cheap ICs. First I built a breadboard to validate operation and then I used KiCAD to build a PCB using JLCPCB from China. In addition to the board I used another soldered breadboard to implement a DAC that had to be connected to the null detector board. The schematic of the circuit is shown in this figure.

It consists of a 555 timer and the following ICs: CD4017, CD4053, 74HC04, 74HC00, CD4053, 40193, some op amps, two transistors, LM393 comparator, and AD6333JNZ. The DAC board is made of CD4067 with fifteen 10k ohm resistors. A small perturbation (+Δ) is temporarily added to the phase control voltage, and the isolated- port detector output is sampled to obtain an error measurement E+. The process is then repeated using a negative perturbation (−Δ) to obtain a second measurement E−. The control algorithm compares the two measurements: If 𝐸+ < 𝐸-, the control voltage is incremented. If 𝐸+ > 𝐸-, the control voltage is decremented. This process effectively estimates the local gradient of the isolated-port power function and drives the system toward the minimum-power operating point. The PCB layout is shown in this figure.

This PCB is a two layer board that I designed using KiCAD and sent the Gerber files to JLCPCB in China for fabrcation. The board came to me after about two weeks and cost me less than $50. I soldered the parts myself and connected the DAC soldered board to it. Testing went smoothly and with minor troubleshooting, it worked perfectly. I used my Siglent digital power supply and my Rigol digital oscilloscope. for the testing, both from China. The test set up is shown in the picture.

The picture shows the controller in operation. The oscilloscope shows the three colored lines: the top blue line is the comparator voltage which is either 0 or 5V that causes the voltage (yellow line) to increase or decrease by Δ. At convergence the error line (violet) becomes close to zero. The timing of the circuit was intentionally made long (about 1 second) to monitor the convergence process. When convergence is reached the comparator keeps alternating between 0 and 5 V at each period, and the varactor bias accordingly varies around the optimum bias that gives minimum error by ±Δ/2. I also used an Arduino microcontroller to do the same function, and in that case once the optimum bias was reached it remained fixed at that value. This null detector was used in my paper that was accepted for publication at the International Microwave and Antennas Symposium (IMAS 2026) to be held in Jeddah in October 19-22, 2026. The paper title is "Adaptive Scalar Null-Seeking Phase Alignment for Microwave Power Amplifier Combining". Here is a video showing the controller in action.

https://youtu.be/bHKavMMv7eM