August 26, 1996
To: Jerry Black, WSOC-TV
Dear Jerry,
Per telco of August 21, 1996, here are some photos of what is happening in the 2 Ghz LNA of the Ultrascan antenna.
Setup:
To determine "typical" signal levels that the Ultrascan will see, a single PCS site at a distance of 2 miles was chosen as an example. This would involve 12 to 15 transmitters with up to 1600 watts (+62 dbm) each. For this test, we shall use two +60 dbm signals, one at 1960 Mhz, and the other at 1989.8 Mhz.
The filter in the Ultrascan passes both of these signals with less than 0.3 db rejection (referenced to a mid-band frequency of 2200 Mhz, minus filter insertion loss). The path loss is 108 db. This leaves an RF wavefront level of -42 dbm at a distance of 2 miles
(-38 dbm at 1 mile, and -34 dbm at 1/2 mile). The gain spec of the antenna is 18 dbi. The dbd rating is then 16. Putting this together, we have +60 dbm ERP, 108 db path loss, and 16 db antenna gain, or (+60) - (108) + (16) = -26 dbm signal level fed to the LNA.
To simulate this, two separate Gigatronics signal generators were fed into a combiner, and the RF power level of each generator was adjusted for -26 dbm as displayed on a calibrated Tektronix 2782 Spectrum Analyzer (picture 1). Output levels were then individually verified with an HP 436A digital power meter, and agreed within 0.2 db.
Picture 1 - Generators are spectrally pure.The combined signals were then applied to the input of the amplifier for WLOX-TV. The amplifier was operating in the bypass mode, so only the first MMIC amplifier was active.
The resulting intermod is shown in picture 2.
Picture 2 - The WLOX LNA overloads with -42 dbm at antenna.
This is with 2 signals. The mix from multiple
transmitters on a PCS site, would fill this screen!
There is no getting around this. In speaking with people at Radio Waves, the maker of the Microwave Radio’s LNA, I was assured that the first stage was clean, but it clearly overloads at moderate signal levels.
In the real world, this amplifier will overload from moderate ENG, ICR, or STL signals, raise the noise floor, and produce intermod. Add PCS signals to this, and you have a disaster!
Circuit losses are very high in the circuit design of this LNA; greater than 15 db through the board, with no amplification.
The gain spec on this unit is 26 db. The actual measured gain (without bypass) on the "Vert" channel was 21.4 db.
This should give you a good perspective of what is happening.
Respectfully,
LLoyd Phillips
