References: <AANLkTi=j-Bp7+YC+527yQ6EBahLUdq6WqYXSGNTk33m7@mail.gmail.com>	<AANLkTimcd822cDsP-X1ieWDUiu89xjVWwOWhXsjvNf-u@mail.gmail.com>	<4D83D475.9090609@telus.net>	<AANLkTinbMc5CWxUF6pSeXmS_R3AkdZGZaCy7i5EUc+Jk@mail.gmail.com>	<BF4A524700075746A6467658DFC7102C88B2129D4C@ICTS-S-EXC2-CA.luna.kuleuven.be>	<000b01cbe622$6a23f8d0$8d01a8c0@JAYDELL>,<AANLkTik_NCrqaJTim47c5n4TC2jsCuCWwapEO7O=rsOi@mail.gmail.com> <BF4A524700075746A6467658DFC7102C88B2129D4D@ICTS-S-EXC2-CA.luna.kuleuven.be>,<4D86BD6F.8020006@telus.net> <BF4A524700075746A6467658DFC7102C88B2129D51@ICTS-S-EXC2-CA.luna.kuleuven.be> <4D8736CB.8040509@telus.net> <4D87E247.5020100@iup.uni-heidelberg.de>
To: rsgb_lf_group@blacksheep.org
Date: Tue, 22 Mar 2011 07:49:10 -0400
In-Reply-To: <4D87E247.5020100@iup.uni-heidelberg.de>
MIME-Version: 1.0
From: Markus Vester <markusvester@aol.com>
Message-Id: <8CDB69E4AD7AA3C-1490-33A81@webmail-m018.sysops.aol.com>
Subject: VLF: Resonate the antenna, or not?
Content-Type: multipart/alternative; 
 boundary="--------MB_8CDB69E4AE12FC4_1490_71A4A_webmail-m018.sysops.aol.com"
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org


----------MB_8CDB69E4AE12FC4_1490_71A4A_webmail-m018.sysops.aol.com
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset="utf-8"


Hi Stefan,=20

you are raising the interesting question of how important it is to resonat=
e the antenna. I think it boils down to a comparison of the amplifier effi=
ciency versus the Q factor of the matching element.

Without the matching element, the amplifier (really a switchmode DC-to-AC=
 converter) has to transfer the reactive power back and forth between the=
 load and the power supply capacitors. With ideal switches, indeed nothing=
 would be lost. But with finite efficiency, you will loose a portion of (1=
 - eta), eg. 2 % per cycle if your amplifier had 98% efficiency measured=
 into a real load. On he other hand, if your loop was resonated using a ca=
pacitor with Q-factor 500, you would dissipate only 0.2% per cycle in it.
=20
Jim's small VLF loop example (0.1 + j 2.2 ohms) had a Q of only 22, ie. 1/=
Q =3D 4.5% of the energy is lost per cycle. So the total power comparison=
 would be 6.5% for the amplifier version versus 4.7% for the capacitor -=
 really not that much difference.=20

But as Stefan says, the amplifier would have to be big enough to handle al=
l the reactive power (VxA). If fed by a large enough power supply, the ver=
y same amplifier could deliver 2.2 kW and raise the radiated power by a fa=
ctor of Q, ie. 13 dB.


BTW There exists a similar criterion for active receive antennas: If you=
 have a small capacitive probe and resonate it with a coil, the coil losse=
s will add in some resistive noise. In a ferrite loopstick, a (nearly loss=
less) capacitor is typically used to tune out the inductive reactance, but=
 this makes the antenna narrowband.=20

If on the other hand you do not compensate the reactance, you can still no=
ise-match the preamp to the modulus of the source impedance, at the penalt=
y that it's own noise temperature will increase by a factor of Q. For very=
 low preamp noise figures (ie. Tpreamp < Tantenna / Q), adding the tuning=
 element brings little or no benefit for the system noise figure.

Best wishes,
Markus (DF6NM)



-----Urspr=C3=BCngliche Mitteilung-----=20
Von: Stefan Sch=C3=A4fer <schaefer@iup.uni-heidelberg.de>
An: rsgb_lf_group@blacksheep.org
Verschickt: Di., 22. Mrz. 2011, 0:41
Thema: Re: LF: Loop TX antennas at VLF...


Hi Scott,=20

I have some comments, yes.=20

I would try not to resonate the PA but drive it with a H bridge PA that ca=
n handle pretty much reactive power. If  you think about Rik's and Jim's=
 calculation and assume 70 V across the loop, while taking 32 A, you ( ah,=
 you have 115 V mains?) you may directly rectify the mains and drive a H=
 bride with it and feed the loop with that switched rectified voltage. So=
 you would get about 75 A. That's not a problem for some well choosen FETs=
 and caps (on the DC side!). Jim and Wolf already suggested to try that.=
=20

I know it from my kite antenna: When the wind is poor or if the wire gets=
 disconnected from the hot end of the coil (this happened 1x), the Z of th=
e primary winding (unresonated) is the current limiting factor. I am still=
 switching at 300 VDC and 8970 Hz to a bundle of wire of 50 turns on a wat=
er barrel (say 70 cm diameter on the lower side). There will be heavy reac=
tive currents but that is no problem for the FETs! The resistive losses in=
 the whole circuit (measured on the 230V AC side by a true rms multimeter)=
 are below 25 W in such a situation, including the driver/exciter power.
I think it wouldn't be a good idea to try this with a standard audio amp..=
..

My FETs are simple IRFP460, about 2.5 EUR each. There are no ferrite parts=
 in the whole circuit!

But, unfortunately, i think there is another problem: If you try that with=
 your loop, who will be the next receiving station? What distance? Even if=
 you run 100A rms into that loop (BTW, what is the area?) you will radiate=
 probably below 1 mW ERP...=20

Now reading the above comments from the others ;-)

Best 73, Stefan EI/DK7FC, currently in IO62SI41IS



Am 21.03.2011 12:30, schrieb Scott Tilley:=20
Hi Rik

Thanks for this!

I may be comparing apples to oranges alittle (maybe alot).  The loop I hav=
e would only require a 1.8uF cap and this coupled with a higher Q makes fo=
r a little dicier tuning scenario.  I concede this scenario is not as bad=
 as I first thought when I thought about this briefly when Stephan enquire=
d about a possible test from here many months ago. A careful review of you=
r summary Rik was very helpful, thanks to all that contributed. =20

The two big , and one minor design issues to get maximum potential from a=
 loop down at 9KHz will be:
1) Current handling capability of the caps.
2) Tuning capability.
3) Ferrite saturation for the matching xfmr...  Or should we use ferrite?

#1 can be resolved, I think, as you have suggested with the use of LOTS of=
 polypropylene caps...

However, #2 will likely be the difficult one and be abit of an engineering=
 challenge as the small loop proposed below may just get away with course=
 tuning but as you get larger and reduce the Rac the Q goes up and so does=
 the tuning criteria.  This is where I got stuck with my original glimpses=
 of thought on this.  We can't rely on a good old vacuum variable so some=
 sort of other idea...=20

Hmmm, how about a special decade type box made of Polypropylene caps?  The=
 box could be designed once the loop's characteristics after installation=
 have been bracketed in...

The other downside of polypropylene caps will be there tendency to drift=
 once you put alot of current into them, so you'll need lots!  This drifti=
ng issue could be very problematic as I discovered on 137 with anything ot=
her than good transmitting micas with high current ratings at the operatin=
g frequency. =20

Based on the results of others using mains transformers on 9KHz maybe item=
 #3 is not a big problem after all?  Anyone have a comment here?

Looking forward to more ideas as this is tweaking my interest into buildin=
g another tuner.

73 es TU
Scott
VE7TIL
...

----------MB_8CDB69E4AE12FC4_1490_71A4A_webmail-m018.sysops.aol.com
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html; charset="utf-8"

<font color=3D'black' size=3D'2' face=3D'arial'>
<div>Hi Stefan, </div>


<div>&nbsp;</div>


<div>you are raising the interesting question of how important it is&nbsp;=
to resonate the antenna. I think it boils down to a comparison of the ampl=
ifier efficiency versus the Q factor of the matching element.<br>
<br>
Without the matching element, the amplifier (really a switchmode DC-to-AC=
 converter) has to transfer the reactive power back and forth between the=
 load and the power supply capacitors. With ideal switches, indeed nothing=
 would be lost. But with finite efficiency, you will loose a portion of (1=
 - eta), eg. 2 % per cycle if your amplifier had 98% efficiency measured=
 into a real load.&nbsp;On he other hand, if&nbsp;your&nbsp;loop was reson=
ated using a capacitor&nbsp;with Q-factor 500, you would dissipate only 0.=
2% per cycle in it.</div>


<div>&nbsp;</div>


<div>Jim's small VLF loop example (0.1 + j 2.2 ohms) had a Q of only 22,=
 ie. 1/Q =3D 4.5% of the energy is lost per cycle. So the total power comp=
arison would be 6.5% for the amplifier version versus 4.7% for the capacit=
or - really not&nbsp;that much difference. </div>


<div>&nbsp;</div>


<div>But as Stefan says, the amplifier would have to be big enough&nbsp;to=
 handle all the reactive power (VxA). If fed by a large enough power suppl=
y, the very same amplifier could deliver 2.2 kW and raise the radiated pow=
er by a factor of Q, ie. 13 dB.</div>


<div>&nbsp;</div>


<div>&nbsp;</div>


<div>BTW There exists&nbsp;a similar criterion for active receive antennas=
: If you have&nbsp;a&nbsp;small capacitive probe and resonate it with a co=
il, the coil losses will add in some resistive noise.&nbsp;In a ferrite lo=
opstick, a (nearly lossless) capacitor is typically used to tune out the=
 inductive reactance, but this makes the antenna narrowband. </div>


<div>&nbsp;</div>


<div>If on the other hand you do not&nbsp;compensate the&nbsp;reactance,=
 you can still noise-match the preamp to the modulus of the&nbsp;source im=
pedance,&nbsp;at the penalty that&nbsp;it's own noise temperature will inc=
rease by a factor of Q.&nbsp;For very low preamp noise figures (ie. Tpream=
p &lt; Tantenna / Q), adding the&nbsp;tuning element&nbsp;brings little or=
 no benefit for the system noise figure.</div>


<div>&nbsp;</div>


<div>Best wishes,</div>


<div>Markus (DF6NM)</div>


<div>&nbsp;</div>


<div>&nbsp;</div>


<div>&nbsp;</div>


<div style=3D"FONT-FAMILY: arial,helvetica; COLOR: black; FONT-SIZE: 10pt"=
>-----Urspr=C3=BCngliche Mitteilung----- <br>
Von: Stefan Sch=C3=A4fer &lt;schaefer@iup.uni-heidelberg.de&gt;<br>
An: rsgb_lf_group@blacksheep.org<br>
Verschickt: Di., 22. Mrz. 2011, 0:41<br>
Thema: Re: LF: Loop TX antennas at VLF...<br>
<br>


<div>Hi Scott, <br>
<br>
I have some comments, yes. <br>
</div>


<div>I would try not to resonate the PA but drive it with a H bridge PA th=
at can handle pretty much reactive power. If&nbsp; you think about Rik's=
 and Jim's calculation and assume 70 V across the loop, while taking 32 A,=
 you ( ah, you have 115 V mains?) you may directly rectify the mains and=
 drive a H bride with it and feed the loop with that switched rectified vo=
ltage. So you would get about 75 A. That's not a problem for some well cho=
osen FETs and caps (on the DC side!). Jim and Wolf already suggested to tr=
y that. <br>
</div>


<div>I know it from my kite antenna: When the wind is poor or if the wire=
 gets disconnected from the hot end of the coil (this happened 1x), the Z=
 of the primary winding (unresonated) is the current limiting factor. I am=
 still switching at 300 VDC and 8970 Hz to a bundle of wire of 50 turns on=
 a water barrel (say 70 cm diameter on the lower side). There will be heav=
y reactive currents but that is no problem for the FETs! The resistive los=
ses in the whole circuit (measured on the 230V AC side by a true rms multi=
meter) are below 25 W in such a situation, including the driver/exciter po=
wer.<br>
I think it wouldn't be a good idea to try this with a standard audio amp..=
..<br>
</div>


<div>My FETs are simple IRFP460, about 2.5 EUR each. There are no ferrite=
 parts in the whole circuit!<br>
<br>
But, unfortunately, i think there is another problem: If you try that with=
 your loop, who will be the next receiving station? What distance? Even if=
 you run 100A rms into that loop (BTW, what is the area?) you will radiate=
 probably below 1 mW ERP... <br>
<br>
Now reading the above comments from the others ;-)<br>
<br>
Best 73, Stefan EI/DK7FC, currently in <A href=3D"http://no.nonsense.ee/qt=
hmap/?qth=3DIO62SI41IS" target=3D_blank>IO62SI41IS</A><br>
<br>
<br>
<br>
Am 21.03.2011 12:30, schrieb Scott Tilley: </div>

<BLOCKQUOTE type=3D"cite">Hi Rik<br>
<br>
Thanks for this!<br>
<br>
I may be comparing apples to oranges alittle (maybe alot).&nbsp; The loop=
 I have would only require a 1.8uF cap and this coupled with a higher Q ma=
kes for a little dicier tuning scenario.&nbsp; I concede this scenario is=
 not as bad as I first thought when I thought about this briefly when Step=
han enquired about a possible test from here many months ago. A careful re=
view of your summary Rik was very helpful, thanks to all that contributed.=
&nbsp; <br>
<br>
The two big , and one minor design issues to get maximum potential from a=
 loop down at 9KHz will be:<br>
1) Current handling capability of the caps.<br>
2) Tuning capability.<br>
3) Ferrite saturation for the matching xfmr...&nbsp; Or should we use ferr=
ite?<br>
<br>
#1 can be resolved, I think, as you have suggested with the use of LOTS of=
 polypropylene caps...<br>
<br>
However, #2 will likely be the difficult one and be abit of an engineering=
 challenge as the small loop proposed below may just get away with course=
 tuning but as you get larger and reduce the Rac the Q goes up and so does=
 the tuning criteria.&nbsp; This is where I got stuck with my original gli=
mpses of thought on this.&nbsp; We can't rely on a good old vacuum variabl=
e so some sort of other idea... <br>
<br>
Hmmm, how about a special decade type box made of Polypropylene caps?&nbsp=
; The box could be designed once the loop's characteristics after installa=
tion have been bracketed in...<br>
<br>
The other downside of polypropylene caps will be there tendency to drift=
 once you put alot of current into them, so you'll need lots!&nbsp; This=
 drifting issue could be very problematic as I discovered on 137 with anyt=
hing other than good transmitting micas with high current ratings at the=
 operating frequency.&nbsp; <br>
<br>
Based on the results of others using mains transformers on 9KHz maybe item=
 #3 is not a big problem after all?&nbsp; Anyone have a comment here?<br>
<br>
Looking forward to more ideas as this is tweaking my interest into buildin=
g another tuner.<br>
<br>
73 es TU<br>
Scott<br>
VE7TIL<br>
...</BLOCKQUOTE><!-- end of AOLMsgPart_2_f6175cf1-9f8d-45b3-ba7e-642d26934=
e06 --></div>
</font>

----------MB_8CDB69E4AE12FC4_1490_71A4A_webmail-m018.sysops.aol.com--