Discrete PLL in this oscillator/converter, how does it work?

I have found this circuit in an old magazine. It is a harmonic converter (receive transverter). It can convert any incoming RF signal into any IF within the HF range. It is not well filtered but it does the job as a lab gear. What is interesting is the PLL section.

I suspect it works by injection locking of the VCO signal to one of the harmonics of the 1MHz harmonic generator, the harmonic that is more close to the VCO signal each time.

Have you got any suggestions of how this part of the circuit actually works?

I would like to eliminate the IC. I see that the diodes circuit is fed with opposite phase by the IC gates. Maybe a crystal controlled multivibrator can do the job instead of the IC?

The phase detector is a sampling detector. The pulses that feed the diodes should have a pluse width that is significantly sorter than the period of the smpled waveform. A logic IC is used to get fast edges for the sampling. The RC time constant from the logic gates to the diodes should be short to give as narrow a pulse as possible to turn the diodes on.
There is plenty of inforamtiuon about sampling phase detectors available online. Most of them are for microwave use and contain a step recovery diode to genreate the samling pulses from a VHF input.

I agree with G4BCH, any attempt to use an analog oscillator in place of the IC will fail. It needs fast square wave edges and one of the two outputs must be inverted with minimal additional phase shift.

One comment I would add: if you build it, add a 5.1V Zener diode across the supply pins of the 7400. I wouldn't trust a resistor alone (bottom left corner) to drop 12V to a stable 5V.

Brian.

So It could be possibly replaced by a step recovery diode circuit to generate pulses on HF range?

The circuit generates FIXED frequencies, it locks the VCO to the harmonics of the IC oscillator. So it cannot be used as a tunable PLL VCO am I right? It is intended to be the first fixed LO in a receiver to downconvert a whole band into a wideband IF. Am I right?

You could replace the chip with an SRD if you can find one that works at HF. I've seen them used in the past with inputs as low as 5MHz for locking cavity oscillators in the 1GHz region. SRDs are not that common there are still a few suppliers but they tend to be for microwave diodes.
The circuit as it is is not very suitable for a tuneable LO, you would have to tune both the reference oscillator and the VCO, or make up some sort of locking circuit to keep them both in lock. This type of circuit is quite old technology, but still has its place in applications such as phase locked DROs
If you need a stable tuneable oscillator it would be easier and beter to replace it all with something like a digital PLL or DDS.

D3 is a voltage tuned varactor diode, where a DC tune voltage controls the VCO frequency. I do not see any "injection locking" going on in that circuit