The modification is so simple. Just

(1) de-activate FM detector by MC3357

(2) change local oscillator crystal to 10.6985 or around

(3) change crystal filter from FM type to SSB/CW narrower bandwidth filter

That's all!

First, just we need to make path just after pin #3 of MC3357 (output of 2nd mixer) to the detector output. I did not pull out any component, so if I am tired or have interest on FM communication, just remove jumper and recover the cut patter will make it FM operation again.

Fortunately I have a small size 10.7 MHz crystal filter (from KENWOOD rig) and simply inserted instead of wide band FM crystal filter.

That's all. I confirmed -136 dBm CW signal detection by this modification. 

The fault point is, the PLL has 10kHz step, meaning  for example it can't receive 144.075 MHz but 144.080 will be ok.

So, So, so far, I've touched "receiver" not transmitter. How we make TX?

In this era, the FM modulation is realized by direct modulation to VCO in PLL loop. The FM signal without input is, yes, continuous wave with single frequency component. So I just cut a purple line connecting microphone amplifier to VCO.

After this modification, PTT switch turns into Morse code keyer and microphone will not work anymore!