KR101930271B1 - Random Number Generator using Dark Current of SPAD(Single-Photon Avalanche Diode) - Google Patents

Abstract

The present invention discloses a random-number generator using a dark current of a single-photon avalanche diode as simple as possible, which has low power consumption and no safety issues.
The present invention relates to a dark current pulse detection circuit for outputting a dark current pulse in connection with the SPAD, a SPAD for blocking photon input, a dark current pulse detection circuit for outputting a dark current pulse, And a pulse-random number converter composed of a logic circuit, in addition to connecting a post-processor.
Accordingly, the present invention realizes a complete random number generator by utilizing the quantum randomness of the dark current generated in the SPAD, and it can be realized in a CMOS process while providing a random random characteristic with a uniform randomness. Therefore, It is possible to integrate it with a logic circuit on a die, and it is possible to reduce the production cost and miniaturization, and there is a beneficial effect that a further enhanced security can be obtained.

Description

[0001] The present invention relates to a random-number generator using a dark current of a single-photon avalanche diode,

The present invention relates to a random-number generator using a dark current of a single-photon avalanche diode (SPAD), and more particularly, to a random-number generator using a dark current pulse, which is an electrical pulse signal generated with a low probability in a complete light- Or a random number generator using a dark current of a single photoconductor balancer diode capable of generating random numbers according to information on the number of dark current pulses generated within a predetermined time interval.

As will be appreciated, random numbers are needed in many fields such as information encryption, communication algorithms, Monte Carlo computer simulation, test vector generation, probability statistical analysis, and games.

There is also a pseudo random number generator (RNG) that computes random numbers with a given seed and polynomial as the simplest way to generate random numbers.

Such a simulated random number generator can be implemented by a software algorithm in a computer or simply implemented in the form of a hardware logic circuit. However, since the seed and the polynomial are estimated from the random number sequence generated in the prior art, there is a problem that the random number generated thereafter can be predicted. In other words, the calculation method works in a deterministic manner, so it can not produce a true random number.

The best way to get a real random number is to use a physical random number generator (Physical RNG or True RNG). As a method of obtaining a random number by measuring a well-controlled physical process, it is possible to consider using minute changes such as temperature, vibration, and voltage. As a physical random number generator, there are a ring oscillator-based random number generator and a metastability-based random number generator in a typical method that can be easily implemented in a general CMOS semiconductor process.

First, a representative example of a ring oscillator-based random number generator is disclosed in Korean Patent No. 1646506 (the invention is called a random number generator). An odd number of inverters are sequentially connected and the output of the last inverter is connected to the input of the first inverter. And thus, the irregularity is obtained at a frequency which changes in accordance with the temperature change or the voltage noise.

In addition, the metastability-based random number generator uses the randomness from the metastability of two inverters crossed each other. 0 '/' 1 'or' 1 '/' 0 'after a certain period of time through the metastability state when the output nodes of two inverters connected to the other input are similarly pulled up to a logical value' Lt; / RTI > This stabilized value is influenced by temperature noise and has an emissivity.

As described above, the ring oscillator-based random number generator or the metastability-based random number generator has a simple structure and is easy to implement in a CMOS semiconductor process. However, it can not prove the information theoretically, and the obtained random number value tends to be deflected , And it is difficult to control it to have a uniform randomness.

Currently, the most recognized method for generating a physical random number generator is a random number generator based on quantum characteristics and a quantum random number generator (abbreviated as Quantum RNG) in abbreviated form. European Patent Application 06742182.6 (entitled Quantum Random Number Generators) .

Such a quantum random number generator is based on the randomness of the two, and the random number characteristic can be verified theoretically. The quantum random number generator is a typical method using light (photon) and using radioactive decay.

The photon-based method extracts a random number using a device composed of a photon generator, a semi-transmission mirror, and a single photon detector. The probability that a photon generated by a photon generator passes through a transflective mirror and is detected in a single photon detector is an ideal transflective mirror and has a randomness of 50%. The photon-based method guarantees randomness, but it is impossible to produce an ideal semi-transmissive mirror, and thus has a problem that the probability is biased.

The way to use radioactive decay is based on the randomness of particle generation from radioactive isotopes. Since the particles emitted to the radioisotope have quantum uncertainty, random numbers are generated using the time data of the emission particles detected. The method using radioactive decay has a good random number characteristic, but there is a danger of using radioactive material, and there is a restriction to be widely used due to the complicated structure and device configuration compared with the existing physical random number generator.

1. Korean Registered Patent No. 1646506 (entitled "Random Number Generator") 2. European patent application 06742182.6 (entitled Quantum Random Number Generators)

SUMMARY OF THE INVENTION It is an object of the present invention to solve the above problems by providing a single photon oscillator diode (SPAD) that does not use a light source or a radioactive isotope, has low power consumption, The present invention provides a random number generator using a dark current.

Another object of the present invention is to provide a physical random number generator which does not use quantum characteristics by making all constituent devices on a single silicon die in a general semiconductor process such as a quantum random number generator and a physical random number generator based on a meta- And a random number generator using a dark current of a single photoconductor avalanche diode so as to satisfy a level comparable to that of a conventional photoconductor and a production cost.

It is another object of the present invention to provide a system semiconductor that requires a random number and can integrate the random number generator as an IP in the same chip, thereby preventing the random heat generated on the chip from being exposed and consequently enhancing system level security The present invention provides a random number generator using a dark current of a single photoconductor avalanche diode.

In order to achieve the above object, the present invention provides a plasma display apparatus comprising: a single-photon avalanche diode (SPAD) for generating a dark current; a light shielding means for shielding the SPAD from being irradiated with light; And a pulse-random number converter connected to an output of the dark current generator and configured to generate a random number by a pulse, and a pulse-random number converter connected to an output of the dark current generator, A random number generator using a dark current of an avalanche diode is provided. The random number generator includes a SPAD, a light blocking means for blocking the SPAD from being irradiated with light, and a pulse detection circuit connected to the SPAD for outputting a dark current pulse A dark current pulse detection circuit coupled to the output of the dark current pulse detection circuit, Wherein the pulse-to-random converter comprises at least one of an M-bit timer and a pulse counter to generate a dark current pulse generation time, a time interval between the dark current pulses, a predetermined time period And a dark current pulse generated from a single photon avulcanization diode, which is obtained by cutting out the lower M-bits from a value of the dark current pulses generated during the dark current pulse generated during the dark current pulse.

Thus, the present invention realizes a complete random number generator by using the quantum random number of the dark current, which has been regarded as an object of minimization as an obstacle to the photon measurement which is the original purpose of SPAD, and it is a quantum random number generator, By eliminating the use of semitransparent mirrors that are otherwise inevitable to bias with other light sources and enabling them to be implemented on a single semiconductor chip, it can be implemented in a CMOS process while providing a uniform random randomness characteristic that is not biased, It is possible to integrate it with a circuit, which has a great advantage in reducing the production cost and miniaturization.

The random number generator using SPAD's dark current according to the present invention can be utilized as an IP in a semiconductor chip design requiring information security and encryption, and when a random number generator is integrated with IP, There is a useful effect of having more enhanced security because it is not exposed.

1 is a circuit diagram illustrating a conventional ring counter type random number generator.
2 is a configuration diagram illustrating a conventional Quantum random number generator;
3 is a schematic diagram showing a random number generator including a dark current generator and a pulse-random number converter according to the present invention.
FIG. 4 is a schematic diagram showing an example in which a random number generation rate is accelerated using a plurality of random number generators and a post-processor of FIG.
5 is a circuit diagram of a dark current pulse detection in a manual mode.
Figure 6 is a circuit diagram of three embodiments of a pulse-to-random converter.
7 is a timing diagram of a pulse-to-random converter.
8 is a schematic view illustrating a configuration of a post-processor.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

3 and 4 show the overall configuration of a random number generator using a dark current of a single-photon avalanche diode (SPAD) according to the present invention.

The present invention is characterized by using an undesired dark current pulse generated in the SPAD to generate unpredictable, non-biased, and unrelated random numbers.

As we all know, SPAD is a semiconductor device originally designed for light sensing. Reverse-biased SPAD also reacts to one photon and causes avalanche breakdown. The transient amplification current generated at this time appears as a pulse signal through the resistor. By monitoring this pulse signal, photons or minute light can be detected.

As a feature that interferes with this intended function, SPAD generates a dark current with a certain frequency even if it does not absorb the photons or absorbs the photons, and the pulse count number increases accordingly. In addition, the pulse frequency caused by the dark current is called a dark count rate (DCR). Physically, dark current is known to occur due to random generation of electrons and holes in the SPAD by temperature or tunneling.

The present invention utilizes the fact that the generation of dark current is random in the SPAD, and the frequency of dark induction due to the dark current is influenced by the area of the SPAD, the bias excess voltage, and the temperature, but compared with the conventional quantum random number generator It is easy to control so that the generation distribution of the dark current pulse is not distorted or deflected by excluding the use of the semi-transparent mirror which is inevitably deflected with the light source.

3, the SPAD random number generator using the dark current according to the present invention is composed of a dark current generator 300 and a pulse-random number converter 400. The dark current generator 300 includes a SPAD 100 and a dark current pulse detection circuit 102 for commonly used photon detection. Particularly, in the present invention, the SPAD 100 is provided with a light shielding unit 101 Blocking layer or a molding package to completely block the photon input to the SPAD 100. This is in contrast to exposing the SPAD 100 to the outside so that a light source can flow into the SPAD 100 at the time of manufacturing the SPAD 100 for photon detection.

The dark current pulse detection circuit 102 outputs a pulse signal when the light source is absorbed by the SPAD 100 or a dark current is generated. The SPAD 100 is shielded from the inflow of photons by covering the light shielding means 101 Only the pulse (dark current pulse) generated by the dark current is output.

This light shielding means 101 may cover only the SPAD 100 element to shield the SPAD 100 from light, or may cover the whole circuit or the entire system including the SPAD 100. 5 shows a general passive circuit configuration of the dark current pulse detection circuit 102. Since the dark current detection operates with the same circuit as the detection of the dark current, various SPAD circuit structures developed and improved for quantum detection Can be used.

The pulse-to-random number converter 400 generates an M-bit digital random number signal (M is an integer equal to or greater than 1) with the dark current pulse information generated by the dark current generator 300. In order to generate the digital random number signal, the pulse-random number converter 400 uses a method (hereinafter referred to as "Method 1") that utilizes the time of occurrence of the dark current pulse and a method that utilizes a time interval (Hereinafter, referred to as 'method 2') or a method of utilizing the dark current pulse number occurring during a predetermined time period (hereinafter referred to as 'method 3'). This is shown in FIG. 6 and will be described in more detail with reference to FIG. 6 as follows.

That is, in the method 1, an M-bit timer (a timer repeating increment from 0 when the maximum value is reached) is installed in the pulse-to-random number converter 400, and when a dark current pulse is inputted, As a random number. (At this time, the output random number value means the lower M-bit of the timer value of the pulse generation time based on a certain time point.)

In method 2, an M-bit timer (a timer that repeats the increment from 0 when the maximum value is reached) is placed in the pulse-to-random number converter 400, and when the dark current pulse is input, , And initializes the timer value of the M-bit timer to '0'. (At this time, the output random number value means the lower M-bit of the inter-pulse time interval.)

Method 3 also includes setting the M-bit pulse counter in the pulse-to-random converter 400 and incrementing the counter value each time the dark current pulse is input. (In this case, if the counter value reaches the maximum value, The pulse-to-random number converter 400 outputs the counter value of the pulse counter as a random value and initializes the pulse counter value to '0' at a predetermined cycle. (The random number value output from the pulse- Bits of the number of pulses in the cycle). The output value then theoretically follows the Poisson distribution with the average dark pulse frequency for a given period as a parameter.

Particularly, in order to obtain an even distribution of the output values, in the present invention, the method 1 and the method 2 require a timer (using a local timer clock) capable of dividing the dark current pulse average generation period into sufficiently large resolutions , And method 3 should capture the pulse counter value with a cycle that is sufficiently larger than the dark current pulse average generation period (using the RNG generation clock). The above methods 1, 2 and 3 have the best output distribution when M = 1, and ideally, the output probability of '0' and '1' converges to 50%, so there is no possibility that the probability is concentrated.

6 shows an example of the hardware configuration of the pulse-random number converter 400 implemented method 1, method 2, and method 3. The method 1 and method 2 internally use a local timer clock to drive the timer, Has a structure in which the dark current pulse is received to drive the counter and, consequently, these values are transmitted to the final output. Both the timers of methods 1 and 2 and the counter of method 3 implement only the lower M-bits in hardware, so they start from '0' when the maximum value is reached. Method 1 and Method 2 may add a step that synchronizes to a random-generated clock before output (grayed out) as needed for random-numbered clock synchronization.

FIG. 7 illustrates a timing diagram to help understand the hardware configuration and operation of Method 1, Method 2, Method 3. In the method 1 and 2 implementation, the timer operates according to the local timer clock, and the timer value is transferred to the output in synchronization with the dark current pulse generation. The difference between methods 1 and 2 is that method 1 is initiated by a dark current pulse generation while method 2 continues to increase the timer. In method 3, the counter is incremented by the dark current pulse and is transferred to the output in synchronization with the random number generating clock.

FIG. 4 shows an example in which the number of random number generators N (N is an integer equal to or greater than 1) and the post-processor 600 shown in FIG. 3 are used to expand the random number generating bit width or to increase the random number generating speed How to do it. In FIG. 4, the N dark current generator 300 and the pulse-random number converter 400 generate N M-bit random numbers in synchronization with the random number generating clock, and the post processor 600 receives these values as input, Generates an L-bit random number synchronized to the clock.

The exemplary structure of the post-processor 600 comprises a first-in first-out (FIFO) buffer 601 and a hash engine 602, as shown in FIG. The FIFO buffer 601 used in the post processor 600 is implemented using a dual clock memory to differentiate input and output clocks and receives (N x M) -bit values in accordance with CLK_fi (random number generation clock) And outputs the output value in H-bits according to CLK_fo (random number output clock). The buffer FIFO buffer 601 is used for a random number output bit width or a random number output frequency conversion. In this case, buffer overflow or underflow occurs when (N x M) x freq (CLK_fi) = H x freq Not.

In addition, the hash engine 602 is a hardware block that generates a L-bit output by receiving an H-bit, and uses the generated hash function to remove the bias that the output value of the random number generator may have . These hash functions are collectively called functions that map data of arbitrary length to fixed length data. Numerous hash functions such as MD5, SHA, SEED, and CRC32 are defined and used in various fields such as cryptography. 600) can select a hash function that is already defined or directly define the formula.

Depending on the convenience of implementation, the hash engine 602 may be located in front of and behind the FIFO buffer 601, or may be omitted to reduce the overall implementation complexity. In the implementation of the hash engine 602, the hash engine output bit width L must be set to a value smaller than the input bit width H to ensure input redundancy, and partial imperfections It can help ease or eliminate sex. In the simplest case, if a hash function is defined to generate M-bit values by XORing two M-bit random numbers, if one of the random numbers is fixed to a specific value and the remaining one random number is ideal, Also have ideal random number characteristics.

The random number generator using SPAD's dark current according to the present invention can exclude a light source, a semi-transparent mirror, a radioactive element, and the like, unlike the random number generator based on the existing quantum characteristics. As shown in FIGS. 3 and 4, It consists only of elements that can be integrated into a CMOS semiconductor process. Therefore, the random number generator using SPAD's dark current according to the present invention can be easily implemented and integrated as an IP (Intellectual Property) like a general analog functional block in system semiconductor design. In addition, the semiconductor device can be designed and implemented in a single CMOS semiconductor process, which makes it possible to miniaturize the device and significantly reduce the production cost.

In addition, in the present invention, by applying a plurality of dark current generators 300 and a post processor 600 as shown in FIG. 4, a random number generating bit can be widened and a random number generating speed can be accelerated.

The output of each of the plurality of dark current generators 300 or the output of the post-processor 600 can be supplied to the post- It is possible to implement various functions such that the output value is deflected or the distribution is not distorted and the random number can be generated at a high speed by scrambling the generated simulated random number in addition to the result value.

While the present invention has been described in connection with what is presently considered to be preferred embodiments thereof, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims and their equivalents.

100: Single-Photon Avalanche Diode (SPAD)
101: Light blocking means 102: Dark current pulse detection circuit
300: Dark current generator 400: Pulse-to-random converter
600: Post-processor 601: Dual clock FIFO buffer
602: Hash engine

Claims (4)

delete A dark current pulse detection circuit composed of a SPAD, a light blocking means for blocking the SPAD from being irradiated with light, and a pulse detection circuit connected to the SPAD for outputting a dark current pulse; A pulse-to-random converter connected to the output and configured to generate a random number by a pulse,
Wherein the pulse-to-random converter comprises at least one of an M-bit timer and a pulse counter to calculate a difference between any one of a dark current pulse generation time, a time interval between dark current pulses, A random number generator using a dark current of a single photoconductor avalanche diode, characterized in that M-bits are cut out and taken as a random number. (Where M is an integer greater than or equal to 1) A dark current pulse detection circuit composed of a SPAD, a light blocking means for blocking the SPAD from being irradiated with light, and a pulse detection circuit connected to the SPAD for outputting a dark current pulse; A pulse-random number converter connected to the output of the pulse-random number converter and configured to generate a random number in M-bits by a pulse, and a random number generator configured by N number of random number generators and a post- A random number generator using a dark current of a single photon avalanche diode characterized by receiving a (N x M) -bit digital signal from a random number generator and producing an L-bit random number. (Where N and L are integers greater than or equal to 1) The method of claim 3,
A post-processor using a hash engine with a random number output bit width or a FIFO for random number output frequency conversion and / or a hash engine with a smaller number of output bits than the input for eliminating bias. Used random number generator.

Images