Yes, Rik. That's the problem. There will ever be very high voltages, even with 
a 300m vertical. So, one can start the calculation by defining the maximum 
allowed voltage at the antenna wire (to avoid PD). Then, with a given C of the 
wire the maximum current is achieved. With the coil, that is fixed in my case 
(disadvantage) but well dimensioned (high Q, advantage) i have to spend that 
parallel C that will increase the coil surrent und thus the losses, and so on 
and on... 
The alternative would be to wound a individual high L coil but this is 
material- and time intensive, maybe 1km of not to thin wire. If Q should be 
high it may become a problem to put that coil into a car to bring it to the 
field ;-) And the high voltage aspect makes it even more difficult. So, it 
seems easier to me to take a smaller but available coil (85mH is not too bad) 
and Cs and spend 7dB more TX power to generate the losses ;-) 
Yes, the small bandwith may cause additional problems. Perhaps it is an idea to 
inject the power by an inductive coupled primary coil, like in a tesla 
transformer. Then one could resonate the system at the low voltage input... 

Tomorrow i will get such a coil and will measure the impedance and Q @ 8,9kHz. 
In a next step i will resonate it (at home, without an antenna) and look which 
power is needed to achieve 20kV rms, and what will happen with the Cs! Then i 
will built a class D PA for abt 100...300W at 12V (2x 2x IRF1405 or so. FETs 
will be driven with my DDS-VFO and a suitable driver, like ICL7667...). I will 
report here then... 

What about your tests, Rik. Any improved arrangements? :-) 

73, Stefan 

________________________________

Von: [email protected] im Auftrag von Rik Strobbe
Gesendet: Do 25.02.2010 10:48
An: [email protected]
Betreff: LF: 9kHz


Stefan,

did you have a look at the graph Jim (M0BMU) showed yesterday ? This may 
indicate that in an open area you might be better off than 1/f, maybe in the 
50..100 Ohm range.
Last night I was mulling about Alexanders remark that a parallel C would 
increase the losses. At first though I would say that the parallel C just 
performs some kind of impedance transformation without inducing additional 
losses. And this is true if the loading coil would be perfect (no losses).
But for a real world situation (ie. assuming a constant Q = 250 for the loading 
coil) Alexander is right: the transformation caused by a large parallel C 
causes extreme high currents in the loading coil, and thus higher losses.
Simulation (with SIMetrix) confirms this.
Taking your antenna (580pF / 50 Ohm)
1. Without parallel C you would need a 550mH loading coil. At Q = 250 the coil 
loss would be 130 Ohm. Putting 100W into this system would result in an antenna 
current of about 750mA and antenna voltage of 23kV. 750mA trough the loading 
coil means a dissipation of 73W.
2. With a 3.2nF parallel C you would need a 85mH loading coil. At Q = 250 the 
coil loss would be 20 Ohm. But putting 100W into this system would result in 
only 330mA antenna current, while the current through the loading coil would be 
2.1A (= 88W dissipation). The reduction in antenna current (750mA to 330mA) 
would lower the ERP by 7dB.
In addition the parallel C will decrease the antenna bandwidth significantly, 
making it difficult to tune.

73, Rik  ON7YD - OR7T

At 01:14 25/02/2010, you wrote:


    Here i agree. And here i understand "dreamers" in a positive and optimistic 
way, just like children are dreamers!
    We have written so many mails (including me, of course ;-) ) und could 
start now just to try what is possible to reach more than 10km if not done yet. 
    
    On the LF pages of DK8KW i found a notice of a successful 2 way QSO over 
10km @ <9kHz. Does anybody know about more than 10km by amateurs?
    
    Yesterday in the night i talked more than 1 hour to Markus/DF6NM and he 
persuaded me to try it with the 100m vertical in a first test. Maybe i get that 
temp licence for a 300m kite vertical but the public autoritys are still 
discussing if i need a special casualty insurance...
    
    Do you, the group, think that it makes sense to try it if the next 
listening station is >50km or even more away? What could be possible if this 
would be the arrangement with the 100m vertical:
     
    The wire has abt 580pF
    I can get a air coil with L=85mH out of a 20kV field. DC resistance abt 
150mOhm, continous current abt 30A, voltage over coil abt 30kV rms (if it 
remains dry(!)), measured self resonance frequency abt 50kHz.
    So, if f=8,9kHz, a parallel C of abt 3,2nF is needed (i can get some 
470pF/40kV Cs, no problem)
    The voltage across my thin wire should be limited to abt 20kV rms, that 
means, antenna current should not exceed 650mA.
    If Rik's 1/f dependency is valid, then my ground losses will be in the 
region of 250 Ohm (no houses and trees there on the hill in a radius of abt 
100m; 15Ohm @ 137 kHz).
    So, 100W TX power would radiate 1,4mW and thus the efficiency would be 
0,0014%. In that case, i mustn't apply more than 100W since the voltage across 
the wire would be to high, so no need for a QRO PA...
     
    If that calculation is reasonably correct, what distance could be reached 
with 1,4mW @ 8,9 kHz in e.g. QRSS10 and the noise level that can be expected on 
a sunday morning? :-)
     
    73 es wish gl to all those, who are trying in these days ;-)
     
    Stefan/DK7FC

<<winmail.dat>>