Desktop Atomic Pile

The world needs a good, portable desk-top atomic pile, suitable for demonstration purposes. Used properly, harm to life and surroundings will be minimal.

This is a a self-contained table-top physical (architectural) model of a hybrid "swimming pool"-type research reactor, such as the well-known EBR-1 in Arco Idaho. A wooden frame contains a 1:24 scale concrete block building, scaffolding and walkway containing the reactor which lives in a vessel of cooling/shielding water. Behind the model/simulator is a vertical panel containing the controls and indicators for the simulator; on top is an ordinary filamentary lamp that derives simulated power output.

Background: Post-war (say, 1945 - 1955) interest in "atomic piles" was at a high pitch, not so coincidentally with the development of automatic computing machinery (same funding sources). The early name "pile" derives from a remark made by Enrico Fermi, the physicist who in 1942 produced the first controlled/sustained nuclear reaction, using a large mound of alternating graphite and uranium metal bars, stacked by hand in hundreds of layers like brickwork, on an old racquet ball court at the University of Chicago's wartime "Metallurgical Laboratory". Fermi: "a crude pile of black bricks and wooden timbers."

Around 1950 foaming-mouth optimism spawned all sorts of amusing side effects, one of which was a plague of electronic simulations of difficult-to-model processes, one of which, naturally, was controlled (and runaway...) nuclear fission.

A nicely succinct description of one such simulation, using analog modelling techniques of course, appears in [Harry] Huskey & Korn's COMPUTER HANDBOOK (McGraw-Hill, 1950), which is neatly bifurcated into ANALOG and DIGITAL sub-books.

As a first-order simulation these things are probably fine; but I believe they lead to, and support, silly oversimpifications. The idea of a desk-top atomic pile is just too much to pass up.

My first pass was going to be 100% period gear, eg. a vacuum-tube analog computer. This quickly became too much to pull together though I did collect most of the (museum-quality) parts and built the chassis for it. Not shown, because it does not exist, is the power supply, which would be quite substantial. This has been abandoned, and now takes up space in my lab to remind me of past folly.

By summer 2007 I'd completed a solid-state version of the computer. It handled all of the terms in E. Morrison's paper (core, coolant temps, neutron flux, Xe-135 and I-131 poisoning) to which I added the apocryphal MWD/MTM; "megawatt-days per metric ton metal", eg. the fuel which is consumed (duh).

Alas... "ontogeny recapitulates phylogeny" [note: that simplisitic idea has been disproven -- but I like it a lot, so I'm sticking with it here for poetic purposes. Write your own text if you don't like it.]... art was tossed cruelly under the bus, when I implemented the entire Morrison algorithm in Arduino C, in order to complete this project for my mid-program (MFA) show...

Mixed media (wood, aluminum, electronic components, software), 16" w x 22" deep x 13" high; approx. 20 lb.

Construction details

These are notes and pictures taken during construction, in no particular order, and not well annotated. Not up to my usual level of annoying documentation, I was goign to school full time, working part time, and commuting 100 miles/day. Busy!

Most of the components are complete; the aluminum panels for the model floor and the meter and controls awaits paint, finish, lettering and assembly.

The reactor hut is fairly complex, in that it must hold the reactor assembly (consisting of a power resistor to heat the water; blue and UV LEDs to simulate the Cerenkov glow (energetic particle dumping energy to not exceed the speed of light), and if I have enough gumption, a working crane to pull moderator rods in and out (probably not). Since it contains water there is a drain on the bottom to ease transport and setup.

Evolutionary dead ends

Most of these are pics of the paths-not-followed-up-on...

The metal fuel in this instance consists of a "D"-cell, suitably decorated with rad-hazard symbols, which will require proper disposal. Not shown.

Alas, reality struck again, in the form of "schedule" for my ACE first year show. I abandoned even this, this modest computer, in favor of (oh, the pain) software simulation in an Arduino Diecimila. Coded the whole thing up in three days.

This is the foolish first pass, to be done in George Philbrick operational amplifiers. It would have beauty that few (literally) alive could appreciate. Alas, abandoned, in shame.