2kHz would cover all voice requirements, except possibly somebody’s ex wife
Then 18kHz should be sufficient if she doesn’t operate at bat’s frequencies.
I doubt they will use 8,33 kHz or a fraction of that as reference (not sure though), I would expect either the much more common 12,5 kHz or 25 kHz.
On band pass filtering, it is true that you won’t be transmitting 25 kHz HIFI audio, yet the KX-165A with 8,33 kHz has additional RX and TX filtering compared to the 25 kHz version. And they use a different local oscillator design, which also seem to suggest that as a standard it’s not just sending another N value to it.
that’s necessary only if you are making e.g. a signal generator which goes up in 1Hz steps
Not necessarily. If you use integer N, you’ll have to use max. 8.33kHz phase comparison frequency, thus much lower loop bandwidth, which means settling time will be long (not an issue for a voice radio), and your VCO will have to use a high Q inductor and very low 1/f noise. During the last 20 years or so fractional N has been absolutely standard. But then again the KX165 design is probably much older…
But Fractional N PLLs have been popular for decades…
I don’t think these radios use fractional-N. That is something that’s necessary only if you are making e.g. a signal generator which goes up in 1Hz steps. And you get some filtering challenges – lots of patents around that filed by Marconi c. 20-30 years ago. Generating ~130MHz in 8.33kHz steps is absolutely trivial, using an off the shelf classical PLL chip.
I doubt 8.33 radios really switch TX bandwidths
I doubt that too. 2kHz would cover all voice requirements, except possibly somebody’s ex wife 
Modification of a PLL to go to a thirth is quite difficult, going to 12.5 would like be fare more easy.
Depends on the design. But Fractional N PLLs have been popular for decades…
You will also have to reduce the bandwith for both RX and TX, as suchs it would require different filtering
Why? I doubt 8.33 radios really switch TX bandwidths, they are bandlimited anyway to max 6kHz, and nobody has been advertising aviation radios with “HD Audio” 8-)
Modification of a PLL to go to a thirth is quite difficult, going to 12.5 would like be fare more easy.
Simpple modification requires new PLL tuning, new software, new read out, and than you are only done for the frequency bit. You will also have to reduce the bandwith for both RX and TX, as suchs it would require different filtering.
If true it is a very nice project of a radio amateur to learn about technology, though no way economical as a modification / upgrade kit.
It does require a new front panel.
The mod was developed by a private person
Yes you are probably right about the FCC approval.
However I am pretty sure that a perfectly usable 8.33 operation can be achieved simply by modding the DPLL to step in 8.33 steps. If that was not true, any use of existing non-8.33 radios would wipe out all the intermediate 8.33 channels. Yet, we have lots of non-8.33 radios and will probably, frankly, have them for ever, existing within the GA community, in areas where radio checks are not needed (some homebuilts, etc). And 8.33 does work right now, with non-8.33 radios operating on the 25kHz channels right next to the 8.33 channels.
Detection is very unlikely since so much GA flies with radios so crappy they barely work.
It would require modification on the receiver, transmitter, audio, oscillators, display control, I doubt this true, as it is quite some work, would be more sensible to start from scratch. I am also quite sure the unit is not allowed on homebuild aircraft, as it will violate the excisting FCC approval for the KX165
I am wondering how they release that modification ?
Even with an official exchange unit you need an EASA approval because it is an equipment change (other Partnumber)
I am reliably informed that somebody has found a way to do this with a KX165
I don’t have the details, but presumably it would be legal only on a homebuilt aircraft
