Browse Prior Art Database

HIGHLY LINEAR BROADBAND DISCRIMINATOR FOR MASS PRODUCTION

IP.com Disclosure Number: IPCOM000007274D
Original Publication Date: 1994-Oct-01
Included in the Prior Art Database: 2002-Mar-11
Document File: 5 page(s) / 261K

Publishing Venue

Motorola

Related People

Alan Rottinghaus: AUTHOR [+3]

Abstract

In any communication environment, the linearity of the receiver is paramount to the minimization of distortion and/or the reduction of the probability of bit error. Narrowband and "medium-band" receiv- ers have information bandwidths which readily lend themselves to sophisticated digital signal processing techniques which practically eliminate this problem. Analog cellular is typical of a narrowband applica- tion, while digital cellular (TDMAKDMA) is typi- cal of a "medium-band" application. Broadband receivers however, usually rely on conventional ana- log techniques.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 28% of the total text.

Page 1 of 5

0 M MO-LA

Technical Developments Volume 23 October 1994

HIGHLY LINEAR BROADBAND DISCRIMINATOR FOR MASS PRODUCTION

by Alan Rottinghaus, Steve Hilton and Gary Schulz

  In any communication environment, the linearity of the receiver is paramount to the minimization of distortion and/or the reduction of the probability of bit error. Narrowband and "medium-band" receiv- ers have information bandwidths which readily lend themselves to sophisticated digital signal processing techniques which practically eliminate this problem. Analog cellular is typical of a narrowband applica- tion, while digital cellular (TDMAKDMA) is typi- cal of a "medium-band" application. Broadband receivers however, usually rely on conventional ana- log techniques.

  For broadband non-coherent receivers, a tuned frequency discriminator is predominantly the ele- ment of choice for demodulation. In a single tuned network, linearity is generally improved by reduc- ing the Q of the network. As the network Q is decreased (improving the linearity), the gain of the discriminator is reduced. If extreme linearity is desired, the gain is significantly compromised resulting in higher noise output. An example of a single tuned network is shown in Figure 1. In this network, C, is the source capacitance with R, L, and C creating the tank element.

  Single tuned circuits produce differential gain curves similar to Figure 2. The curve shown is nota- bly parabolic due to the limited polynomial order. Asymmetric distortion of the curve is referred to as tilt, which results in an increase in even harmonics (when a tone modulated carrier is discriminated). Symmetric distortion of the curve is referred to as bow, which results in odd harmonic increases.

Double tuned networks allow for better control ofthe distortion than do single tuned networks, with-

out compromising the discriminator gain. The added distortion control comes from the higher polyno- mial order of the network, allowing more degrees of freedom in tuning. This network, shown in Figure 3, does not provide a simple solution however; since, the large number ofelements make adjustment more

0 Motorola, Inc. 1994

of an art rather than a methodology. And while achieving high linearity for an individual circuit is difficult (from the tuning standpoint), mass produc- tion repeatability is nearly impossible. To achieve a highly linear differential gain without manual adjust- ment, each component (including the demodulator's source impedance), must be repeatable to within one or two percent. Practically speaking, this task is impossible for inductors and the source impedance.

  The dual tuned network can be improved; but first, a general understanding of the circuit is necessarry. Even though the total transfer function is a result ofa lumping ofall elements, the effects of each individual resonant element (primary-Rl, Cl, and Ll and secondary-R2, C2, and L2) on the total transfer fimction can be seen. In fact, for qaximal linearity, a re...