Date of Award

Spring 1985

Document Type

Dissertation

Degree Name

Doctor of Engineering Science in Electrical Engineering

Department

Electrical Engineering

First Advisor

Jacob Klapper

Second Advisor

Sol Rosenstark

Third Advisor

Roman Wolodymyr Voronka

Fourth Advisor

Joseph Frank

Abstract

The phase lockedloop is widely used as an FM demodulator known as a phaselocked detector (PLD). It is particularly useful in the detection of weak signals because lower threshold is possible with a PLD than with a conventional limiter discriminator. An enhanced version PLD known as an extended range phaselocked detector (ERPLD) is capable of further threshold reduction. The most popular PLD is a second order type I system employing a loop filter having one real pole and one real zero.

One implementation of an ERPLD calls for the addition of a differentiator to the standard loop filter which replaces the real zero by a pair of complex zeros. Previous studies have assumed an ideal differentiator which results in an unbounded noise bandwidth and therefore requires the use of a predetection filter. But this is an idealization which is not physically realizable because there will always be an additional pole due either to design or to stray effects. This dissertation is a study of the effect of an additional pole upon the threshold performance of these detectors.

A number of different PLDs and ERPLDs were selected for study. They are characterized by the presence or absence of an additional pole, the presence or absence of a predetection filter, and either test tone or voice modulation. In all, ten systems are analyzed and each is optimized with respect to threshold by a determination of the set of loop parameters which results in the minimum threshold carrier-to-noise-ratio,(CNR)TH. Mathematical models are developed for (CNR)TH in each case and the optimization is performed by digital computer algorithm.

The calculations reveal that for a PLD with no predetection filter and for an ERPLD with a predetection filter, threshold decreases with increasing pole frequency, becoming optimum as the pole location approaches infinity (i.e. no second pole). However, only a small penalty is incurred if the pole exists at a sufficiently high frequency. Thus for 1 KHz test tone modulation and 10 KHz frequency deviation, a second pole at 140 KHz increases the minimum (CNR)TH of the PLD by only about 0.1 dB.

With the second pole added, the ERPLD can be optimized for use with no predetection filter and its minimum (CNR)TH is about 0.5 dB below that for the optimum standard PLD. In this design the real zero in the loop filter is located at infinity (i.e. eliminated).

The model is shown to be inapplicable to the case of a standard loop when used with a predetection filter and therefore no optimization is achieved for this case.

Practical optimum loop filter designs with components having ordinary tolerances are feasible because the minimum threshold levels are not very sensitive to small variations in parameter values.

Experimental verification of the computer results is given for several systems with test tone modulation. The theoretically derived optimum systems were implemented in the laboratory with high gain active filters and discrete phaselocked loops and exhibited measured threshold levels consistent with the calculated values.

Share

COinS