Tobias wrote:

> Can you differentiate if the koax heats the toroid or the
> toroid itself gets warm ?

Starting from cold, if I do a short run at full power, the wires
get hot but the core only gets slightly warm.  I conclude the
heat is coming from the wire not the core.

Johan wrote:

> No fancy Litz here but the flat geometry may have  helped to
> reduce the skin effect loss.

Yes I'm getting keen on flat wire.  I think proximity loss may be
a big factor here too.  The secondary wire gets hot but the
wire leading away from the secondary only gets slightly warm and
the air cored coils of my LPF which are carrying about the same
current also only get a little warm and they have thinner wire.
So I reckon the strong field in the toroid is giving a lot of
proximity effect.

I dare not put any more current through this transformer until
I've got some suitable wire.  Thanks to all the advice I'm
beginning to understand what I need for a good transformer.

But I need a few more dB to try to reach Italy so I've connected
the secondary windings in parallel instead of series and
doubled the supply voltage.  DC supply is now 62.5V and 3.45A
compared with 27V and 3.78A.  The output voltage has increased
considerably too.  So I think I might be about 3dB stronger
for no extra heat in the toroid.

Also, tonight I will go to 1 second symbols for a 5 character
message.  That will gain at least another 3dB.

Proposing

 Freq: 477690.00000 Hz
 Start: 22:00:00 UT
 Coding: 8K19A
 Symbols: 512 at 1 second
 Chars: 5
 Repeat: XX:00 XX:15 XX:30 XX:45

I'll run now for an hour to see if the PA is reliable with this
setting, so EbNaut as above from 17:00 UT for a while.

--
Paul Nicholson
--