That is exactly what I did.
I ran a bare copper wire 14 gauge around the length of my layout and clearly marked it as the ground for K83 devices. This allowed me to then use a single cat 5 network cable for each K83 since the cat5 network cable has 4 pairs of wire in it. Makes for a neater install in my opinion.

It would look something like this:

(picture shows 5 turnouts instead of 16)

Advantage is that you only need to bring the blue leads of each turnout to the decoder, and connect the yellow to a nearby point on the yellow common line.

Good scheme - thanks very much, I'm off to get hold of some Cat5 cable

Here is an example of how I connected the yellow wires to the bare copper wire. The K83 is connected to the bare wire on the left of the arrow and one of the yellow wires is connected via a extension which in this case was some left over brown and white cat5 wire. All connections are soldered on and the extension pieces have shrink tubing over the soldered connections. I leave extra wire under the layout so that if I have to lift a section of track, the slack in wires make for an easier removal.



The second photo shows the connections at the turnout end.

I concur.

I would not tie the yellows of different k83s together

Mmmmh, so to be safe I better run a couple of extra yellow wires.

Thanks everybody for your help.

Best regards,
Gregor

Because the Yellow is a -22V pedestal generated within each k83 there is, at minimum, an additional capacitor and diode between it and the blues of another k83. Tieing the yellows together @ the k83s bypasses this and in some ways "potentially" increasing the effective capacitance for a bigger punch. BUT it is also stressing the weakest links which will be doing most of the work all of the time.

As stated this can work if there is close tolerances of the two units. However as there are at least two (PCB) versions of 6083 as well as the 60830 this cannot be guaranteed.
If you want to coexist with Viesman or other look a likes (some of which do cater for a separate yellow input - which is different from the yellow output) then all bets are off.

This also presumes the normal usage of a k83 as a pulse in the circuit - if you have a short (moisture in the mechanism from over zealous PVA when ballasting) the domain of potential failure , and resulting problem determination, is greatly broadened

Add to this the possibility that (in the future) you may have k83s on layout segments powered by different booster - some of which will not have a common brown let alone a common red.

You may think you understand all these risks but if you entertain comrades to help on your layout or are likely to expand the layout over time and/or introduce new technology as it comes along ......

In the case of a club layout - keep it standard and minimise the risk as much as possible.

I say - Just don't go there