This new 20 Mhz version of the sensor stage is stable and more sensitive than the 8 Mhz one, as the resonant L / C circuit is tuned on a harmonic of the 20 Mhz oscillator and this allows a more efficient operation of the Mixer (TR2). I built an oscillator range 95 - 135 Mhz and approaching the antenna of the lrl to the coil of the oscillator I noticed an increase in the DC signal at the output of the sensor stage at the values of 100 Mhz (fifth harmonic) and 120 Mhz (sixth harmonic ). I added the capacitor C19 to obtain a greater amplitude of the oscillations, in the 8 Mhz version it was not necessary.



                     Components and helps

All resistors are 1 / 4W, capacitors are ceramic, only C5 and C7 are electrolytic, all transistors are BC183C or equivalent (BC547C, BC549C and similar), TR2-TR3-TR4 must have a beta (gain) greater than 500 , you have to measure with the tester, in fact I found BC183C with beta of 350 and 400.
If you do not find the indicated value for c10 you can use a 10 / 47pF capacitor with a fixed 47pF capacitor in parallel, of course you have to try without the 47pF capacitor and if you don't get an increase in the signal at the output of the sensor stage then put the capacitor in parallel. R4/R5 have been reduced from 330 to 33 ohms, With the P2 / P3 potentiometers it is easier to set up, the same is true for P1 which allows you to dose the amplitude of the signal from the oscillator to the mixer.

R1 = 22K        C1 = 100pF           C16 = 33nF
R2 = 47K        C2 = 22pF            C17 = 4.7nF
R4 = 33 ohm     C5 = 100 uF/25 V     C18 = 4.7nF
R5 = 33 ohm     C6 = 100nF           
R6 = 1M         C7 = 100nF           
R7 = 6.8K       C8 = 100 uF/25 V     Q = 20Mhz quartz
R8 = 1M         C9 = 560pF           D1 = D2 = 1N4148 
R9 = 4.7K       C10 = 10/90pF        or similar
R11 = 1M        C11 = 560pF          TR1-TR5 = BC183C or
R13 = 4.7K      C12 = 560pF          similar BC...C
R14 = 4.7K      C13 = 560pF          P1 = P2 = P3 = 1K trimmer
R15 = 3.3K      C14 = 560pF          L1 = 3 turns air core 
R16 = 3.3K      C15 = 33nF           10 mm diameter
  
Regarding the pcb I used the one of the 8 Mhz version, since the changes mainly concern the value of the components. If there is little space, you can remove P2 and P3 and put a fixed resistance, once the right value has been established. As for the 8 Mhz version, the realization is quite critical because a lot of gain is needed and it is necessary to avoid that the sensor stage self oscillates before having sufficient gain. So a double-sided pcb is recommended and solder the components directly on the tracks. Alternatively, you can use a single-sided pcb and apply a screen on the lower face, for example an aluminum sheet or an unengraved pcb. First you need to check with the oscilloscope that the quartz oscillator is working, a signal on the emitter of TR1 must be in the range 3 - 6 V peak to peak, better if the signal is not a perfect sine wave but is distorted, in such a way there will also be many harmonics. By changing the value of C1  it is possible to change the amplitude and distortion. At this point it is necessary to move the cursor of P1 all towards ground to cancel the signal that goes to the base of TR2 and to check the DC output of the sensor stage, by varying P2 and P3 it is necessary to have a positive signal value (any value is not critical ), this means that the sensor stage self oscillates. The gain (in DC voltage) of TR3 and TR4 depends on the value of the resistance between the emitter and ground and the value for the alternating voltage also depends on the capacitors C13 and C14.