The quartz oscillator serves to provide a stable reference signal present even in the absence of a "phenomenon" which, if present, will cause an increase in the signal. TR2 (original 8Mhz scheme) is the mixer at the base of which the oscillator signal and that of the phenomenon from the antenna arrive. TR2, TR3 and TR4 are the high gain amplifier, TR5 takes the signal (rectified by diodes) with high impedance and  on the emitter there is the signal with low impedance. R15, R16, C17 and C18 are a low pass filter that cancels signal residues at 20 Mhz. I repost schematic and PCB, unfortunately the 20Mz oscillator has 2 transistors and various additional components, therefore it will have to be made on a separate pcb (I don't have the drawing available, but it's very simple). However, it is not a disadvantage because in this way you can find the position of the oscillator with respect to the sensor stage, which must not be too close, to avoid unwanted coupling. It's better to use a double face PCB, on one side we sold the components and the other side there is the shield connected to ground, to avoid self oscillations. It's preferable to use transistors type BC...C because the high beta (gain), I use BC183C because I have a lot but work well also BC549C, BC109C, BC239C and others. All sensor stage is powered by 12V stabilized. For the first test do not connect quartz but connect the antenna, also a piece of wire about 40 cm long, then look at out point, the DC voltage must be 0V. If not the stage is self oscillating. It must be considered that the rectifying diodes have a threshold of about 1V, or a little more, so even with 0 V at the output there could still be a signal of lower amplitude. I recommend using the oscilloscope only to control the operation of the oscillator, but not use it to measure the amplitude of the signal on the collector of TR3 and TR4, as this will radically change the output signal (usually the signal disappears). To vary the gain I have provided a 2.2K trimmer instead of R10 and I changed R12 from 1K to 220 ohm. The gain of the sensor stage depends on the beta of the transistors and this can vary a lot even if they are all of the "C" type. Adjusting the gain is important, as to be sure it is the maximum possible you must first adjust it so that you have self oscillation and then adjust until the oscillation disappears. Since it is not possible to be sure that there is not yet an oscillation lower than 1V (approximately) due to the diodes, if later, after the connection with the oscillator, the DC signal at the output is unstable, it is necessary to further decrease the gain. As I have already said there are different types of quartz and not all oscillate with the same value of the components, this is especially true for the frequency of 20Mhz or higher, therefore I suggest to temporarily put a 4.7K trimmer in place of R3 and adjust it to have a signal of about 1 or 2 V,  with C1 = 15pF, the absorption must be between 5 and 10mA. If value is lower or if it does not oscillate, try C1 = 39pF and possibly other higher ones, until the desired amplitude is obtained. The measuring point with the oscilloscope is on the emitter of TR1. Once the required amplitude has been obtained, replace the trimmer with a resistance whose value is closer to that measured or you can leave it. Then put a 47K trimmer in place of R4 and check the waveform on the collector of TR2, it must be about 1V peak to peak wide and distorted. Now adjust P1 to have about 2 - 4V at the sensor stage DC output, if not, increase C3, then set C10 for the maximum output voltage of the sensor stage. If the voltage exceeds 6V, decrease it by varying P1 and bring it to a value between 2 and 5V. Now check if touching the antenna the signal at the output decreases (even slightly). At this point you are ready for an outdoor test.