Browse Prior Art Database

Capacitor Card Threshold Circuit

IP.com Disclosure Number: IPCOM000096480D
Original Publication Date: 1963-Jun-01
Included in the Prior Art Database: 2005-Mar-07
Document File: 2 page(s) / 52K

Publishing Venue

IBM

Related People

Kiessling, CE: AUTHOR [+2]

Abstract

The circuit 10 is a digital simulation of a network of threshold logic circuits. As such, it can be used to perform the functions of such networks, as, for example, used in pattern recognition, speech recognition or Boolean Logic implementation. Capacitor cards 14a and 14b can be of the type described on page 96 of the March 1963 issue of the IBM Technical Disclosure Bulletin, Vol. 5, No. 10. Weighted values are provided in each card column. This is by selectively punching out different capacitor elements provided by conductive spots on the cards, a so as to leave elements only at desired locations. The weighted values in each column can be determined by specific values assigned to unpunched areas in accordance with either decimal or binary code.

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 58% of the total text.

Page 1 of 2

Capacitor Card Threshold Circuit

The circuit 10 is a digital simulation of a network of threshold logic circuits. As such, it can be used to perform the functions of such networks, as, for example, used in pattern recognition, speech recognition or Boolean Logic implementation. Capacitor cards 14a and 14b can be of the type described on page 96 of the March 1963 issue of the IBM Technical Disclosure Bulletin, Vol. 5, No. 10. Weighted values are provided in each card column. This is by selectively punching out different capacitor elements provided by conductive spots on the cards, a so as to leave elements only at desired locations. The weighted values in each column can be determined by specific values assigned to unpunched areas in accordance with either decimal or binary code. The weighted values are obtained from a adaptive procedure simulated on a digital computer.

In this character recognition application, the character is represented by a binary array of thirty-five bits. Using a one-out-of-n coding scheme, n being the number of characters to be recognized, there is one capacitor card provided for each character. Each bit has a weight attached to it as a result of the adaptive process. Each weight is stored in a card column. Thus, thirty-five columns are used per card. Weighted digit values from each or the columns or a plurality or changeable capacitor cards 14a, 14b, etc., are read into corresponding counters 12a, 12b, etc. The cards are scanned in parall...