Hi Stefan,

>> Third test:
>> Moved dummy load to output of ferrite matching transformer with
>> ratio set for 1:1 impedance, so
>> PA>  LPF>  ScopeMatch sensor>  long coax>  1:1 xfmr>  dummy load. 15
>> minute RF carrier at 150W. At start, voltage and current on the
>> scope equal amplitude and in phase. After 15 minutes no change.
>>    
> All right, so now the PA and the cable and the xfmr is not the problem, 
> as long as the xfmr sees Z=R=50 Ohm.
> Did the xfmr stay completely cool?

Yes, it stayed cold.

>> Fourth test:
>> With coil and antenna, so
>> PA>  LPF>  ScopeMatch sensor>  long coax>  xfmr set to match initial
>> antenna R to 50 ohms>  Loading coil>  antenna. Started RF carrier at
>> 150W, voltage and current on the scope equal amplitude and in phase.
>> After 45 seconds, voltage started to decrease, current started to
>> increase, and phase began shifting with voltage leading. After 3
>> minutes the phase stopped changing and remained with voltage leading
>> by a few degrees. After 6 minutes, voltage and current reached a
>> steady state with current much higher than initial. RF power had
>> increased significantly due to the lower Z load on Class-E PA. It
>> remained steady for the remaining 9 minutes of the test.
>>    
> Did the xfmr become hot?

No. As before it was completely cold.

> It provides galvanic decoupling from the shack 
> earth and antenna earth, right?

Yes. The primary of the xfmr is connected between center and shield
of coax, but not to antenna earth. The secondary of the xfmr
connects between the bottom of the loading coil and the antenna
earth system.

> Obviously your resonated antenna does not provide a Z=R=50 Ohm 
> impedance. Otherwise there would be no difference to the dummy load. 
> BTW, the output signal is usually a good sine wave i assume?

As far as I can tell it is a perfect sine wave. Looking at the RF
after the filter with a spectrum analyzer shows very low harmonic
levels.

> If the xfmr has just enough primary turns not to saturate at 50 Ohm, it 
> may thermally run away when connecting something higher than 50 Ohm 
> because then it starts to saturate.

So what if I could make a dummy load with the same R as that of my
antenna / earth and do the test again?

> It must be the core i think, because if the antenna detunes, the current 
> would decrease. But if the coil becomes hot, then it will!

I'm not sure I understand.

BUT... I found ONE thing that gets warm. It is that bottom-most turn
on the loading coil. The one that is a shorted turn because of the
way the tap is connected. I do not dare put my hand in there when
the transmitter is running! It takes me about 45 seconds after
shutting down the transmittter to make SURE the TX is disabled, go
outside and feel the coil. At that point, with air temperature being
about 0C today, that bottom turn (or two) is still noticeably warm.
The rest of the coil feels cold.

I have just brought the big coil inside for inspection and a
modification to eliminate the shorted turn at the bottom.

> The xfmr is switched between antenna earth and loading coil, in series, 
> and then the coil connected to the wire, right?

Yes.

73,
Paul