Method for Quantization of Binary Signals


This patent refers to a method for quantization of binary signals. Such signals might be e. g. the clipped IF (e. g. 10.7 MHz) of an FM receiver, or the oscillator signal from an FM condenser microphone.

The text describes how these signals can be synchronized to a high-frequent system clock, while the time quantization noise introduced by the synchronization process is shifted to higher frequencies well above the baseband.

Corresponding principles are well-known with voltage-domain ADCs as noise-shaping. This patent introduces a noise-shaping method in the frequency domain. The mixed analog-digital frontend utilizes noise-shaping of 2nd or higher order.

Once the signal is synchronized, it is a synchronous binary data stream, which can be further processed (e. g. FM-demodulated) by digital methods, implemented in ASICs, FPGAs, or DSPs.

This method allows direct digital demodulation of FM signals, while it does not require any conventional voltage-ADCs.

In contrary to analog FM-demodulation methods, FM demodulators after this patent have a much longer conversion slope (several MHz), which also is nearly ideal. This results in extremely low distortion of the demodulated (then digital) audio signal.

Typical applications of the patent's quantization method are:

These quantizers are also very well suited for space-use, since they can be built as hybrid modules, made from rad-hard parts.

The patent text may be requested from the Deutsches Patent- und Markenamt. or retrieved from Europe's Network of Patent Databases.

The author is happy to provide more details on request.

Requests about the European Patent may also be sent to company von Hoerner & Sulger GmbH, e. g. by E-Mail.

Copyright © 2001 - 2017 by Hartmut Henkel

Last file change: 2002/01/07