JSC 1318

 4.2296 3.11
 8.4655 3.86
12.7028 4.79
16.9407 5.57
21.1772 6.41
25.4190 6.97
29.6600 7.73
33.8981 8.14
38.1429 8.89
42.3888 9.11

12.2964824  207.368060736293  2424.31892969673 5.33222740327061 -58.4734516928983
12.4824121  613.191835297415  4172.12464852922 5.33145577102593 -30.5750121668707
12.6683417 6924.16636110013  12439.8056909876  5.34964419981199  -3.06964395870417
12.8542714 2520.92184257022  -8190.69094405199 5.45947830037164  24.3919472010177
13.0402010  417.895587911188 -3445.71636197857 5.59177485432348  51.9923972818836

Several steps were performed for selecting the transformer Balun. The terminated line Smith Chart was as an indication of proper operation. With the window line directly connected to the 50 Ohm unbalanced calibrated plane, the smith chart had ovals with about a 2x1 shape factor. Added six inches of 43 material snap-its changed the oval shape, but still an oval.

Window line, being symmetrically balanced, is a good candidate for a transformer Balun with a grounded center tap. This type of Balun has negligible common mode rejection. A symmetrically balanced load has negligible common mode current.

The transformer inner windings are 11T red color #24, with a center tap. Care was taken to layer the windings. The outer winding, four turns #22 green, is a single layer. S11 measurement with an appropriate load indicated no resonant issues. The trace was rather flat, but decided to add the small capacitor to compensate out leakage inductance. The range of the vintage 422 ARCO is 40 pF. The capacitor was adjusted for a near flat response from 3 to 40 MHz. The leakage inductance takes over above 40 MHz. Below 3 MHz suffers from low winding inductance. Suspect the capacitor to be at about 30 pF.

OSL calibration was performed at the 11T solder connections. The load is a VNWA measured 390 Ohm 1/4 watt beige-bodied, probably helical cut carbon film, which measured 380 Ohm with an inductance of 30 nH. The resistor values entered into the VNWA calibration kit window.

The resulting smith chart above has nice circles, though they do spiral up a little. This is a good indication that the selected Balun is performing the intended function.

Correctness can also be evaluated by observing the loss computations. As long as copper skin is fully involved, TL loss will follow a k1*sqrt(F) + k2*F line. Notice in the top loss plot that at other then the 1/4 and 1/2 wavelengths, the 10 to 40 MHz plot follows a k1 k2 line very nicely. The 11 to 14.5 MHz loss plot includes marked measurement points. Other then the three points about 12.7 MHz, 3/4 wavelength, the measurements are very close to the k1 k2 line.

Chipman computations are not valid at Z equal zero conditions. The detailed Zo and Loss plots centered at the 3/4 wavelength, 12.7 MHz, demonstrate the resulting values are invalid near zero Z values.