<< 
Feeding and matching a 1/4 lambda circumference loop antenna. 
Magnetic loop calculation software presumes that the
RFcurrent has a constant value everywhere around the
radiator.
This assumption gives useable results for very small loop antennas, with a circumference smaller than 1/20 lambda.
What if the circumference is 1/8 lambda to 1/4 lambda?
Example : A tuned 1/4 lambda circumference loop antenna.
 7.1 MHz
 100Wrms
 Circumference 10m
 Radiator 80mm (!!) thick:
Calculation example using 66pacific calculations : 
These calculations are wrong.



By research, G0CWT found that the current around the circumference of a 1/4 lambda loop is NOT constant, but varies largely.
This
has large consequences for practical (real) loop properties.
Calculation example using practical impedance measurements
by G0CWT. From this and the 100W transmitter power we can calculate ( I^{2} = P / R) the loop current as
4.26A. G0CWT measured
a feedpoint impedance of 22.5 Ohms at the tuning capacitor. The capacitor voltage recalculated. In practice my loop has on 40m an
"VSWR < 1.5
bandwidth" of 85kHz. 
Still not convinced ? Compare these results with calculations after DL4CKJ.
Matching a 1/4 lambda circumference loop antenna.
G0CWT uses a broadband ferrite core transformer to match the 50 Ohms coax feeder to the loop.
Coupling between his primary and secondary windings is mechanically tight, with the
drawback of also capacitive coupling.
A suitable transformer can be inserted in any point into the loop :
 in the current maximum or
 in the current minimum or
 inbetween.
Connecting the matching
transformer to the tuning capacitor has a big advantage : 
For these
three cases, the transforming ratios will differ, as the feed point impedances
will differ. G0CWT measured an impedance of 22.5 Ohms near the tuning capacitor.
Zpri : Zsec = 50 :
22.5 = 2.25 : 1 => turns = 1.5 : 1
According to my experience, capacitive coupling must be absolutely minimal, to prevent Efield radiation and reception, TVI / BCI and manmadenoise reception.
To avoid capacitive coupling and capacitive unbalance to the loop, i constructed a transformer with very low coupling capacitance between the primary and secondary windings.
After experiments with several transformers i found :  To small winding inductance has negative influences to the tuning behavior, selectivity and performance of the loop.  A to high secondary winding inductance introduces, in parallel with the total tuning capacitance of the loop, a blocking parallel resonance. An FT24061 ring core with sec. 8t and
pri. 12t #61 material is chosen for its low losses, and flat frequency and temperature responses. 
Matching on 80m.
On 80m the circumference of the loop is only 1/8
lambda. Its feed point impedance drops by a factor 9.
By adding an auto transformer with a transformation ratio of pri : sec = 3 : 1 good match is achieved again. This is done by winding a transformer on a FT24061 or FT18561B ringcore, wound with 15 turns, and a tap on 5 turns from the ground connection..