(SWS)
(CIGAR)
(DESIGN)
(1G3)
(2G4)
(3G4)
(HOME)
Cigar antennas for 1.2GHz, 2.4GHz and 3.4GHz
Matjaz Vidmar, S53MV
4. 23-element cigar antenna for 1.3GHz
The first experiment was to scale the original Russian cigar "BDM-2" down to 1.3GHz. At lower frequencies more accurate measurements could be done. At 1.3GHz the scaled "BDM-2" cigar together with a standard TV-antenna mast clamp fitted efficiently an available 1m long M10 threaded rod:
The many disk elements were cut from 1mm aluminum sheet using handheld sheet-metal scissors. Spacers were made from 12mm/10mm aluminum tube. Unfortunately the latter sometimes had unacceptable tolerances. Some samples had their internal diameter less than 10mm thus not fitting on a M10 threaded rod.
A standard female N connector with flange UG-58/U was used for the feed. The flange was attached to the reflector disk "R" with four screws M3x6mm. The center pin was soft soldered to a M3 screw using an approximately 11mm long copper tube (shield of UT141 semirigid cable). Finally the screw was attached to the driven element "S" with a M3 nut and lockwasher.
Computer simulation of a similar disk-Yagi without the boom calculated the center frequency too low by about 100MHz! This error was corrected by adding an accurate electromagnetic simulation the boom. The corrected simulation more than doubled the number of segments in the 4NEC2 model.
Radiation-pattern measurements required a suitable reference antenna. A 17-element NBS-Yagi designed for 1.3GHz did not provide the required bandwidth: above 1.4GHz the NBS-Yagi already reverts its main beam backwards. A 12dBi cup dipole was found a suitable reference antenna including its good forward/back ratio to suppress unwanted reflections. Finally, when more cigar prototypes for 1.3GHz became available, they also performed as excellent reference antennas further suppressing unwanted reflections of the antenna test range.
Since radiation-pattern measurements were performed in an open field, out-of-band interference from high-power cellular-phone towers had to be suppressed by a microstrip bandpass filter 1100MHz-1500MHz in front of the diode detector. Radiation-pattern artifacts caused by interference were further suppressed by a careful choice of the lock-in frequency. For accurate results, the diode-detector response was calibrated by a precision step attenuator.
The measured radiation patterns of the final version of the 23-element cigar antenna for 1.3GHz (including unwanted reflections of the antenna test range!) are presented in the following gif animations (scanning all frequencies and both polarization planes) in both linear (field) and logarithmic (dB) scales:
The directivity was calculated from both radiation patterns in the E and H planes measured at each particular frequency in the 1120MHz-1480MHz range. At each frequency two different algorithms were used for the directivity: average of inverse partial directivities <Dj> and geometrical average of partial directivities <logDj>. The first is mathematically correct for a large number of cuts (measurement planes) of the radiation pattern while the second provides better results with just two cuts (E and H planes) for fan-beam antennas. Since cigar antennas produce rotationally-symmetrical pencil beams, both algorithms provide similar results:
The directivity of the 23-element cigar antenna for 1.3GHz was measured about 2dB higher than a 17-element NBS-Yagi with a similar boom length and about 1dB higher than a 50cm diameter SBFA for 1.25GHz. The slow-wave-structure bandwidth of the 17-element NBS-Yagi is much narrower. The useful bandwidths of both the 23-element cigar and the SBFA are mainly limited by the feed bandwidth.
The impedance matching of the scaled version for 1.3GHz was not particularly good. The return loss could not be found better than -10dB at any frequency (SWR worse than 1:2). Such an impedance mismatch implies an efficiency/gain loss of about -0.5dB. The feed eccentricity "e" had to be reduced from the 43mm scaled value down to 30mm to obtain good impedance matching to 50Ω. The suggested driven-element "S" diameter of 124mm provided the best impedance matching around 1250MHz. A smaller driven element "S" of 121mm provided the best impedance matching around 1300MHz.
Comparison tests confirmed that thanks to the excellent radiation efficiency and good impedance matching, the gain of the 23-element cigar antenna for 1.3GHz is practically identical to its directivity.