I have been investigating the design of MMI (anomalous cognition) generators that will be truly responsive. As I noted elsewhere, responsive generators will likely be based on entropy sources of a certain nature.
- They will produce a discrete signal, that is, not a continuous analog signal, but each output will start and end for each output.
- They will likely have to be quantum mechanical.
- They will likely have to produce outputs at a high rate to allow subsequent processing, such as bias amplification or other.
- They will likely require a large array of individual sources to allow sufficient sensitivity to reach the required responsiveness.
I know of a number of designs that could meet all these requirements, though some may be too complex or expensive for practical designs. At this point I will only list their general descriptions. When I get time, I will give more specific details on how to build them.
- Decay time of fluorescence. A pulsed blue or UV LED is used to excite a phosphor. Once the excitation is ended, the fluorescence decays at an exponential rate. The time for the glow to decay to a preset level, say 10% of the max level, is used to produce a quantum random output. The decay is detected by a simple photodiode circuit. This type of generator is small, inexpensive and easy to build. It can be used to generate outputs up to the gigabit per second rate, though rates in the megahertz range are much easier.
- Decay time of ionization of noble gas. This is similar to the fluorescence decay generator, but the gas is most easily excited by an electronic current, though RF energies will also work. Small neon lights are readily available at low cost. Lamps containing other noble gasses such as argon are available. Argon plus a phosphor coating result in green and blue “neon” lights. For this application the standard red-orange neon bulbs are preferred since other noble gasses have longer afterglow decay times. The decay of excitation is caused primarily due to radiative emission, so a photodiode is the easiest way to observe it. Not quite as simple as the fluorescence decay generator, but not excessive. The decay of the ionization afterglow is not nearly as rapid as fluorescence, so generation rate is more likely in the 1Kbps range.
- Nuclear decay timing. A radioactive material, such as Americium-241 used in some smoke detectors, emits alpha particles or “radiation.” The timing of this decay is purely quantum mechanical. Each decay is detected by a special photodiode or Geiger tube. The average time between decays is measured by a sample-and-hold circuit and low-pass filtered. A random output is generated for each decay, depending on whether it occurred more or less quickly than the average. A typical 0.8-1 micro-Curie source used in a smoke detector only produces enough decays to generate 10-12 Kbps. While this is a “pure” quantum generator, there are several issues to deal with. An array of such generators would contain an array of sources – likely an issue with the Nuclear Regulatory Commission. In addition, this type of generator would be more expensive and bulky.
- Photon shot noise. A pulsed light source from an LED can be split by either a regular beam splitter or a polarization beam splitter. Photodiodes placed at each of the two output ports of the beam splitter produce outputs that can be processed to determine which output port passed the largest number of photons. Random numbers can be generated from this measurement. This is approaching a “pure” quantum system, though it is still fully quantum mechanical. The components are more expensive than the first two examples and require more precise physical and electronic design. Generation rates up to a few Mbps are readily obtainable. An array of these generators would become quite expensive.
- Single-photon path detection. A pulsed light source from an LED is attenuated so only a few photons will be detected during each pulse. The photons are passed through a regular beam splitter or a polarization beam splitter. Single-photon detectors are placed at each of the output ports of the beam splitter. The outputs of the detectors are sent into a circuit that measures which output port passed a single photon that was detected first. This signal can be used to generate random numbers. Such a generator is what might be called a “pure” quantum source, but it is not quite pure since some photons are lost in the splitter and some are not detected at all due to inefficiencies in the detectors. Generation rates up to 10s or even 100s of Mbps are possible. These generators have an esthetic appeal, but they are the most difficult to build and the most expensive as well.
There may be several other entropy sources, such as Zener diodes, but it’s not immediately obvious how to make a high-speed generator with a discrete output using Zener diodes, though I think it is possible.
Each of these generators can be made by a multitude of circuit variations and components. I will try a few and report the results if anyone is interested.