Random number generator with digital feedback

• US 4,853,884 A
• Filed: 09/11/1987
• Issued: 08/01/1989
• Est. Priority Date: 09/11/1987
• Status: Expired due to Term

• Current Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Original Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Inventors: Dabbish, Ezzat A., Brown, Daniel P., Danielsen, Carl M.
• Primary Examiner(s): Ruggiero, Joseph
• Application Number:
  US07/095,088
• Time in Patent Office: 690 Days
• Field of Search: 364/602, 364/607, 364/600, 364/717, 331/78, 380/46
• US Class Current: 708/3
• CPC Class Codes: G06F 7/588 Random number generators, i.e. based on natural stochastic processes

Abstract

A zener diode random number generator circuit is described which produces a random binary number output having a statistical distribution exhibiting a controlled degree of randomness determined in response to an input control signal. A microprocessor feedback circuit monitors the random number output and produces the input control signal in response to the difference between the degree of randomness of the output signal and that of a pre-determined statistical distribution. The digital feedback automatically adjusts the zener diode biasing point and the limiter threshold such that part-to-part tolerance, component aging, temperature variations, or voltage fluctuations will not adversely affect the randomness of the bit stream output. In the preferred embodiment, the microprocessor tests the ratio of ONES bits to ZERO bits of the random number such that a desired 1:1 ONES/ZERO ratio is approximated.

Claims

1. A random number generator having a controlled degree of randomness comprising:

noise source means for producing an analog noise signal having a statistical distribution exhibiting a particular degree of randomness determined in response to an input control signal;

pulse shaping means for generating random binary data in response to said analog noise signal;

feedback means for monitoring said random binary data and for producing said input control signal in response to the difference between the degree of randomness of said binary data and that of a predetermined statistical distribution; and

control means for outputting said random binary data only when it is within prescribed limits of said degree of randomness of said predetermined statistical distribution.

2. The random number generator according to claim 1, wherein said noise source means includes a zener diode having a bias point at the knee of the avalanche breakdown region of its characteristic curve so as to generate random noise, said bias point being responsive to said input control signal.

3. The random number generator according to claim 1, wherein said pulse shaping means includes a limiter device having an input switching threshold defined in response to its supply voltage, and wherein said feedback means includes means for varying said limiter supply voltage.

4. The random number generator according to claim 3, wherein both said analog noise signal and said limiter supply voltage are responsive to changes in said input control signal.

5. The random number generator according to claim 1, wherein said feedback means includes a microprocessor which produces said input control signal in a digital format, and further includes a digital-to-analog converter which produces a variable bias voltage in response to said digital input control signal.

6. The random number generator according to claim 1, wherein said predetermined statistical distribution exhibits a ratio of ONES bits to ZEROS bits which is approximately equal to 1:

1.

7. A circuit for generating random binary numbers which meet a given criteria for randomness comprising:

circuit means for producing a bias voltage in response to a control signal;

noise source means for producing an analog noise signal having a random noise component and an average DC component, both of which are responsive to said bias voltage;

pulse shaping means for amplifying and limiting said analog noise signal to product a digital pulse stream having randomness properties defined by both said random noise signal; and

controller means for sampling said digital pulse stream for a particular time interval to provide a random binary number, and for producing said control signal in response to the difference between desired randomness criteria of said random binary number and said given criteria for randomness.

8. The circuit according to claim 7, wherein said controller means produces said control signal in a digital format, and wherein said circuit means includes a digital-to-analog converter for varying said bias voltage in response to said digital control signal.

9. The circuit according to claim 7, wherein said noise source means includes a zener diode having a bias point at the knee of the avalanche breakdown region of its characteristic curve so as to generate random noise, said bias point being determined by said bias voltage.

10. The circuit according to claim 7, wherein said pulse shaping means includes a limiter device having an input switching threshold defined in response to its supply voltage, and wherein said controller means includes means for varying said limiter supply voltage.

11. The circuit according to claim 10, wherein both said analog noise signal and said limiter supply voltage are responsive to changes in said bias voltage.

12. The circuit according to claim 7, wherein said controller means is a microprocessor which determines the acceptability of said random binary number with respect to the desired randomness criteria of a predetermined statistical distribution and outputs said random binary number only if it is within prescribed limits of said desired randomness criteria.

13. The circuit according to claim 7, wherein said desired randomness criteria is the difference between the number of ONES bits and ZEROS bits of said random binary number.

14. The circuit according to claim 13, wherein said predetermined statistical distribution exhibits a ratio of ONES bits to ZEROS bits which is approximately equal to 1:

1.

15. The circuit according to claim 13, wherein said controller means determines the acceptability of said random binary number by testing whether the difference between the number of ONES bits and ZEROS bits of said random binary number is less than a predetermined limit.

16. A method of generating random binary numbers which meet a desired criteria for randomness comprising the steps of:

generating an analog noise signal having a random noise component and an average DC component, both of which are responsive to a variable bias voltage;

amplifying and limiting said analog noise signal to produce a digital pulse stream having randomness properties defined by both said random noise component and said average DC component of said analog noise signal;

sampling said digital pulse stream for a particular time interval to provide a random binary number;

measuring desired randomness criteria of said random binary number;

determining the acceptability of said random binary number with respect to the desired randomness criteria of a predetermined statistical distribution;

outputting said random binary number only if it is within prescribed limits of said desired randomness criteria; and

varying said bias voltage in a manner to direct the randomness properties of said digital pulse stream to approach said desired randomness criteria if said random binary number is not within said prescribed limits.

17. The method according to claim 16, wherein said noise signal generating step is performed by a zener diode having a bias point at the knee of the avalanche breakdown region of its characteristic curve so as to generate random noise, said bias point being determined by said variable bias voltage.

18. The method according to claim 16, wherein said limiting step is performed by a limiter device having an input switching threshold which is also responsive to said variable bias voltage.

19. The method according to claim 16, wherein said determining step determines the acceptability of said random binary number by testing whether the difference between the number of ONES bits and ZEROS bits of said random binary number is less than a predetermined limit.

20. A random number generator having a controlled degree of randomness comprising:

noise source means for producing an analog noise signal having a statistical distribution exhibiting a particular degree of randomness determined in response to an input control signal, said noise source means including a zener diode having an adjustable bias point near the knee of the avalanche breakdown region of its characteristic curve so as to generate random noise, said adjustable bias point being responsive to said input control signal;

pulse shaping means for generating random binary data in response to said analog noise signal; and

feedback means for monitoring said random binary data and for producing said input control signal in response to the difference between the degree of randomness of said binary data and that of a predetermined statistical distribution.

21. The random number generator according to claim 20, wherein said pulse shaping means includes means for limiting said analog noise signal to provide said random binary data, said limiting means having an adjustable input switching threshold.

22. The random number generator according to claim 21, wherein said feedback means includes means for adjusting both said zener diode bias point and said limiting means input switching threshold.

23. A random number generator having a controlled degree of randomness comprising:

noise source means for producing an analog noise signal having a statistical distribution exhibiting a particular degree of randomness determined in response to an input control signal;

pulse shaping means for generating random binary data in response to said analog noise signal, said pulse shaping means including means for limiting said analog noise signal to provide said random binary data, said limiting means having an adjustable input switching threshold responsive to said input control signal, wherein said random binary data has a statistical distribution exhibiting a particular degree of randomness determined in response to said input control signal; and

feedback means for monitoring said random binary data and for producing said input control signal in response to the difference between the degree of randomness of said binary data and that of a predetermined statistical distribution.

24. The random number generator according to claim 23, wherein said noise source means includes a zener diode having an adjustable bias point near the knee of the avalanche breakdown region of its characteristic curve so as to generate random noise, said adjustable bias point being responsive to said input control signal, wherein said analog noise signal has a statistical distribution exhibiting a particular degree of randomness determined in response to said input control signal.

25. The random number generator according to claim 23, wherein said feedback means includes means for determining the acceptability of said random binary data with respect to the desired randomness criteria of said predetermined statistical distribution, and for outputting said random binary data only when it is within prescribed limits of said degree of randomness of said desired randomness criteria.