Power Line Influences
After the results from September 11 2001 became available, many people asked whether the eggs might have been affected by the terrific stresses on the power grid and big changes in the electromagnetic environment. There are a number of logical and technical reasons why this is not likely, and some of these are described here. Other aspects, especially the logical XOR protections and the empirical findings are described elsewhere.
REG Design, Computer Power Supply
In response to concerns that changes in the electrical supply grid might propagate into the behavior of the REG/RNG devices, John Walker supplied the following observations about the potential for power mains frequency perturbations in the GCP data.
First of all, each and every REG in the network runs on direct current produced by a regulated DC power supply: an external power cube in the case of the Pear REGs, or obtained from the RS-232 port for the Orion and Mindsong REGs used by the vast majority of the eggs. The 12 Volt power taken from the RS-232 port is typically produced by a switching power supply in the computer, which up-converts the incoming mains power to a higher frequency to reduce the mass of the step-down transformer which produces the low voltages used by the PC components. The output of the secondary windings of this transformer, which of course has the same frequency fed to the primary--much higher than the mains frequency (say 25 to 100 Khz, as opposed to 50-60 Hz) is then rectified into DC by diodes, which is filtered to remove residual ripple by capacitors (which can also be much smaller thanks to the higher frequency used by the switching power supply).
The 12 V signals on the RS-232 port which power the Orion and Mindsong eggs are ultimately derived from the computer's power supply, but some RS-232 interfaces further complicate the matter by generating their own ±12 V supplies from the more abundant +5 V bus using devices called DC/DC converters, which may itself be a small switching power supply.
Thus, the power supplied to the REG is DC, with the bulk of the residual ripple at the frequency of the switching power supply, and the primary cause of fluctuations those due to load variations in the computer (for example, when the hard drive seeks or the CPU dissipation changes) which the regulation Any influence of the mains power frequency should be far less than those generated within the computer, since to the extent mains frequency ripple reaches the DC power buses of the computer the power supply is failing in its task of providing regulated DC power.
It's quite easy to put some numbers on these matters. Just hook up an oscilloscope to the 12 V power used by an REG and measure how much ripple there is at the mains frequency. I may do this if I can find the time. A USB port provides substantially more power to a device (2 Watts if I recall correctly, at a more useful 5V), but there's no reason to believe the quality of the power (ripple and regulation) is necessarily any better.
Note that even if the REG were battery powered, it remains immersed in the ambient EM field of the power lines surrounding it. Since the REG cannot in any case be perfectly shielded (since it's required to connect to a computer which is a maelstrom of potential EM noise ranging almost from DC to daylight), one could argue power line influences even then.
Finally, recall that whatever power line influences are asserted must take into account the difference in power line frequency, which is nominally 60 Hz in some regions and 50 Hz elsewhere. Most contemporary computers accept either frequency, so the residual ripple, to the extent it exists, will be at the frequency of the mains it is connected to.
In addition to the points John makes, we should add that there is yet another layer of protection from line fluctuations. The design of the REGs includes internal power regulation, and in the Mindsong devices this is in fact a two stage design. The people who develop research grade REG/RNG devices are of course aware of the potential for power fluctuation to degrade the quality of the random source, and so have invested the necessary effort to mitigate or prevent such influences. We are confident that such influences are not a viable explanation for the anomalous deviations that correlate with major events in the world.