DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
  s=s1024; d=btopenworld.com;
  h=Received:X-Yahoo-SMTP:X-YMail-OSG:X-Yahoo-Newman-Property:Message-ID:From:To:References:In-Reply-To:Subject:Date:MIME-Version:Content-Type:Content-Transfer-Encoding:X-Priority:X-MSMail-Priority:X-Mailer:X-MimeOLE;
  b=DVIGOLQq0vFAgD/W2ZXO+oXjNvD7pmUluFUNhjgBr8vI3DVk2T0GrKrWfEFaNY00822QFLD072uYlccF6Ob1tBWqVgCe5wzHz63oUklHZCg+9RIpPsETUlqP4snvWc6WKRpm7a0ov2eBdnXQt+f30w/1K6Q2epgvr6esjI8aeUA=  ;
Message-ID: <9DC81BDAB5CF49D8BB66CF0F3868CC6F@JimPC>
From: "James Moritz" <james.moritz@btopenworld.com>
To: <rsgb_lf_group@blacksheep.org>
References: <1MTPIE-2555Oq0@fwd08.t-online.de> <4A674683.1000301@usa.net>
In-Reply-To: <4A674683.1000301@usa.net>
Date: Wed, 22 Jul 2009 22:33:27 +0100
MIME-Version: 1.0
DomainKey-Status: good (testing) 
Subject: Re: LF: Loop (was Re: IGBT in 136 KHz TX?)
Content-Type: text/plain;
	format=flowed;
	charset="iso-8859-1";
	reply-type=original
Content-Transfer-Encoding: 7bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org


Dear LF Group,

Since I have used several designs of loop antennas almost exclusively for 
136k and 500k reception, here are some thoughts:-

For small VLF - MF RX loops, the available signal power from the loop 
depends mostly on the area of the loop and the unloaded Q of the loop (the 
shape, and in multi-turn loops the distributed capacitance have some smaller 
effect too). You get maximum power delivered to the receiver when the loaded 
Q with the receiver connected is 1/2 the unloaded Q. This happens when the 
impedance connected to the loop has a resistive component equal to the loss 
resistance, and a capacitive reactance equal to the inductive reactance of 
the loop, i.e. resonance and conjugate match. To a first approximation, the 
number of turns does not affect the signal power level, only the design of 
the impedance matching. With very small area loops you want to try to 
achieve conjugate matching in order to maximise sensitivity. But with 
relatively large area loops (of the order of 1 square metre or more), and a 
suitable low-noise receiver or preamp, etc. you can get enough signal so 
that band noise is the limiting factor even with quite a large mismatch. In 
particular, mismatching to reduce the Q increases the bandwidth, which is 
often very useful. At LF/MF, it is usually possible to make preamps whose 
noise level is either below the band noise, or below the thermal noise level 
generated by the loss resistance of the loop, so the loop itself is normally 
the limiting factor. With several square metres area, you can get ample 
signal level without bothering to tune to resonance, and have really 
wideband loop antennas. The EMF induced in a given  loop is proportional to 
frequency, so the output decreases at low frequencies, but since the band 
noise level increases at low frequencies, this isn't usually a problem

How many turns the loop has is largely a matter of convenience for the 
tuning and impedance-matching point of view. The traditional high-Q loop has 
many turns so that it can be tuned to resonance with available variable 
capacitors; also the resulting high parallel resistance gives a low noise 
figure when connected directly to a valve input stage, although this is not 
so much of an advantage these days. There are several disadvantages , such 
as being easily de-tuned, susceptible to capacitive noise pick-up, requires 
remote tuning, mechanically complicated.

Small single-turn tuned loops have the drawback of requiring a large tuning 
capacitor. For instance, my 136k, 1m^2 "bandpass loop" design has a 0.4uF 
tuning capacitance. But since a fairly narrow band is to be covered, the 
tuning is fixed, and the loaded Q is quite low this isn't really a problem, 
and the loop is a simple square of tubing that is not significantly de-tuned 
even if laid on the ground.

As has been pointed out, you can add a 1:n step-up transformer to a 
single-turn loop, and the combination essentially behaves as a n turn loop, 
with loop EMF increased by a factor n and resistance/inductance n^2 times 
that of the loop element itself, so this is a flexible way of getting the 
impedance/inductance level you want. The transformer will add some loss 
resistance, but usually this can be made small enough not to make much 
difference. Also, the transformer winding inductance will be effectively 
shunted in parallel with the transformed loop inductance, and a small 
leakage inductance appear in series. In Alberto's case, with a single-turn 
winding through the core, the problem is likely to be getting sufficient 
primary inductance. You would want to choose a core where a single turn has 
a relatively large inductance compared to the inductance of the loop, i.e. 
with AL >> loop inductance. Small single turn loops have inductance of the 
order of a few uH - this requires quite a large core in high permeability 
material, or several stacked smaller cores. Or you could incorporate the 
shunt inductance into the circuit design. It would be harder to optimise the 
transformer design if you are restricted to a single turn primary, which 
might cause problems if you were trying to make a high Q design.

Some people have made quite large, high Q tuned loops - the main 
justification for this would seem to be to get a large signal output so that 
a receiver with poor sensitivity can be used directly without a preamp. The 
high loop Q helps to prevent the RX being overloaded by out-of-band signals. 
The same result could be achieved with a small/wideband loop with a suitable 
preamp/preselector.

A large area wire loop can be connected directly to a low-impedance receiver 
input with reasonable results as a wideband antenna- or a considerable 
improvement in matching can be obtained using a wideband transformer as 
PA0RDT and DJ1ZB have done. The loop inductance can be incorporated as part 
of a filter/matching network, as I have done in the "bandpass loop" designs. 
Low-pass designs are also quite feasible - I am currently using 4 x 5m 
single-turn wire loops incorporating low-pass filters matched to 50ohms, 
with cut off frequencies around 550kHz to reduce the local broadcast 
signals. These give plenty of signal over the 10 - 550kHz range without 
re-tuning.

Small loops have the advantage that they are easier to re-position away from 
noise sources. At LF , mains wiring seems to be a potent source of H-field 
noise, so it is often essential to experiment to find the best location. The 
height above ground, or the presence of trees and buildings has little 
effect on the signal level. Whether the loop is big or small, narrow- or 
wideband, it has the same figure-of-eight pattern, so the nulls can be used 
to reject local or distant QRM. Unless your real estate is big enough for LF 
phased arrays, the loop is the only feasible LF directional antenna. At 
M0BMU, 500kHz reception would currently be almost impossible without loops 
due to wideband QRM radiated by the broadcast stations.

Cheers, Jim Moritz
73 de M0BMU




----- Original Message ----- 
From: "Alberto di Bene" <dibene@usa.net>
To: <rsgb_lf_group@blacksheep.org>
Sent: Wednesday, July 22, 2009 6:04 PM
Subject: Re: LF: Loop (was Re: IGBT in 136 KHz TX?)


>I would like especially to comment on using a one
> winding loop combined with a step-up transformer:I am interested in 
> knowing opinions about using the loop itself as
> a primary of the transformer, as in this picture (heavily compressed
> to stay into the size limits). Here it is a three-turn loop, but it could
> be a single turn one, as discussed.
>
> 73  Alberto  I2PHD
>
>
>