Re: LF: 12 km on Dream(ers) Band - TX loss calculation

[Markus Vester <markusvester@aol.com>]

Dear Rik, LF,

 

the reason why I prefer series feeding is that the amplifier current drops when the antenna falls out of resonance, either due to arcing, or detuning by wind (or pieces falling off). With a tap at a few turns or light inductive coupling, the detuned load would be seen nearly as a short circuit. There are ways around this, either by parallel-resonating the small coupling coil with a large capacitance (ie. bandfilter coupling, which also provides a convenient variable impedance match), or inserting a lowpass filter having about 90 degrees phase shift.  

 

In the first round of 9 kHz experiments, Walter (DL2LF), Geri (DK8KW) and myself all have attempted to use large ferrite cores. In principle you can come to a more compact and higher Q coil. But when going to large voltages, there are two types of difficulties: Core saturation introduces additional hysteresis losses, accompanied by current-dependent inductance which made tuning "jumpy". To mitigate that, you basically need a large iron cross section and a large air gap. The other issue was voltage breakdown between the "hot" end and the partly conductive core material.

 

I think my coil wire is actually .2 mm copper diameter, about 0.55 ohms/m. The ten sections were wound separately on their food containers using a "converted electrical lemon squasher", and series connected externally after assembly.

 

Best 73,

MArkus

-----Urspruengliche Mitteilung-----
Von: Rik Strobbe <rik.strobbe@fys.kuleuven.be>
An: rsgb_lf_group@blacksheep.org <rsgb_lf_group@blacksheep.org>
Verschickt: Mo., 8. Mrz. 2010, 17:05
Thema: Re: LF: 12 km on Dream(ers) Band - TX loss calculation

Hello Marcus,

could the loss in the impedance transformer be avoided by tapping the loading coil ?
I just had a look at the "copper wire tables" and 0.2mm OD wire (actually 0.15mm copper) has a resistance of 1 Ohm/m. Quite a  challenge to get 830m wire on the coil without breaking the wire :-).
Would it also work with less (but thicker) wire and a ferrite core ?
The ferrite core will introduce some additional loss, but on the other hand you will have less (and thicker) wire reducing the copper losses. But I am not sure what the overall effect will be.

73, Rik  ON7YD


At 11:44 8/03/2010, you wrote:

Dear Rik, LF,
 
yes these numbers are pretty much where I arrived in my calculations. There is some loss in the transformer, about  1024* 0.25 ohm from the primary and 60 ohm in the secondary, adding up to 315 ohm series loss. Thus antenna resistance at 9 kHz would be more like 700 ohm. The same antenna has a much lower resistance on 137 kHz, around 35 ohms normally, and down to 28 ohms in cold and dry weather. This includes the LF coil resistance (RF litz wire, Q ~ 700, Rcoil ~ 7 ohm). The large difference indicates that capacitively coupled losses from high-resistance objects (trees, roof etc.) are probably dominating at 9 kHz.
 
BTW I just noticed an error in the the 9 kHz coil description, it actually has 10 sections, 500 turns each (back to kindergarten then to learn counting to ten ;-). Skin effect is probably still neglegible with the thin 0.2 mm wire, but there may be some eddy current loss (proximity effect) in the overlapping layers.

Best 73,
Markus


-----Urspruengliche Mitteilung-----
Von: Rik Strobbe <rik.strobbe@fys.kuleuven.be>
An: rsgb_lf_group@blacksheep.org <rsgb_lf_group@blacksheep.org>
Verschickt: Mo., 8. Mrz. 2010, 9:22
Thema: Re: LF: 12 km on Dream(ers) Band

Hello Marcus,

first of all congrats with your achievement.
Based on the numbers you gave I tried to figure out the loss resistance at 9kHz:
- 35W and 0.135A would give 1920 Ohm.
- with a 32/1 transformer the TX (audio amp) would be loaded with 1.8 Ohm
Assuming that the loss at 9kHz is slightly more than the DC loss (let's say 900 Ohm) the "ground loss" (including greenery and buildings) would be about 1 kOhm.
Can you confirm this figures ?

73, Rik  ON7YD