If it comes to it could make a big radial head from some large heat sinks I have.
We have made two versions of the copper head. Version one had a 1.5mm layer of copper plazma sprayed onto and into the combustion chamber of a modified GP125 head, which I would think would make a good thermal contact, and then teamed it up with a copper fin.
The thinking with this head was that the thick copper combustion chamber shell would be better at transmitting waste heat to the outside but the tarnishing of the copper became a problem, to be solved later.
The second one is basicaly an ordinary alloy head and like before its fitted with an oversized copper head gasket that extends into the combustion chamber to form the squish band and extends out the sides to form an extra cooling fin.
With the extra cooling the copper fin gives, the head runs at a measurably lower temperature. So I recon its a Keeper. The next part of the project is to reduce the heat uptake to the cooling system from the exhaust and add some air ducting.
This engine started with a 90 main jet and now runs a 112.5. Its funny, as we developed the engine we had to keep leaning the needle and richening the main jet.
F5 Dave suggested that because jetting up now, does not seem to richen it up, the needle could be controlling the fuel on full throttle and not the main jet.
We followed his suggestions, removed the main jet, ran the bike and it bogged from 1/2 throttle on, so this tells us that the needle and needle-jet combo are ok.
Looking at the piston wash pictures its clear like you say, the GP is running lean.
I have seen the sand blasted look of detonation before, mostly in other peoples TZ's. What looks like detonation in the picture is in fact the piston touching the squish band and polishing the combustion deposits on the edges of the piston. I ran my TZ this way and it worked OK, very well in fact, but I have to be mindfull that the MSV produced suits the GP.
From the piston wash pictures its clear that I could run the bike much, much richer and the wash pattern would tell me something about the transfer stream behaviour. Then following your sugestions and what I can find out from other sources make improvements.
.
Bookmarks