The Handheld Analog Computer That Made the Atomic Bomb

szczys writes: When the physicists and mathematicians of the Manhattan Project began their work they needed to establish which substance was most likely to sustain vigorous fission. This is not trivial math, and the solution of course is to use an advanced computer. If only they had one available. The best computer of the time was a targeting calculation machine that was out of service while being moved from one installation to another. The unlikely fill-in was a simple yet ingenious analog computer called the FERMIAC. When rolled along a piece of paper it calculated neutron collisions with simple markings — doing its small part to forever change the world without a battery, transistor, or tube.

Comments

That "Made" the Atomic Bomb

[xxxJonBoyxxx]

>> That "Made" the Atomic Bomb

Should be "that simulated the atomic bomb" instead.

Re:That "Made" the Atomic Bomb

[ArmoredDragon]

Actually it should be neither. From TFA:

UPDATE: Commentor [lwatchdr] pointed out that the use of the FERMIAC began after the Manhattan Project had officially ended in 1946. Although many of the same people were involved, this analog computer wasn’t put into use until about a year later.

Really editors?

[DougOtto]

The whole article is how this device was used to build the bomb.....get to the end and they add a correction. The FERMIAC wasn't used until after the Manhattan project was completed. Basically the whole article is wrong, they said it was wrong, and it got green-lighted here.

Re:Really editors?

[fnj]

a slide-rule ... is highly accurate to quite a few decimal places

Bunk. A customary 25 cm slide rule is accurate to 2-3 decimal places - and that is plenty for a great many engineering tasks, such as building bridges and locomotives; hell, even B-29s. There are actually markings for 2 places, and the figure for the 3rd place is interpolated by eye. In 18th century Germany, somebody made a slide rule 2 meters long with a microscope attached to it, which could give 6 places. About the only real physical problems at the time which could even make use of that kind of precision were certain subtle astronomical calculations.

Circular slide rules are a way to get about three (pi) times the precision for the same linear dimension by winding the scales in a circle. Mostly all they do for "normal" sizes is make the interpolation of the 3rd place a little more precise. A tradeoff is error due to the slop in the axial pivot.

It should be well known to any high school grad, but I better mention that 3 times the precision is only about half on one decimal place. It's like the way "orders of magnitude" are frequently misspoken. I was reading some bunk about "many" orders of magnitude, where the writer was only talking about a factor of a thousand. A thousand is a large factor, but it's only three orders of magnitude. Ten orders of magnitude encompasses the difference between a large nuclear weapon and a hand grenade.

Slide rules only give you the mantissa. Everything is normalized to 0-999. You still have to keep track of the 10s exponent in your head or on paper. It was damn good mental exercise. Yeah, I know a thing or two about slide rules since I enrolled in engineering college in 1965.

Banks of women sitting at adding machines

[Anonymous Coward]

I thought they used banks of women sitting at adding machines, carrying out algorithms handed to them by Feynman...Fermi would do simple back of the envelope calculations, and Von Neumann would solve the differential equations in his head...

Re:Wooden calculators

[rickb928]

I still miss my old Pickett aluminum rule. Came with a hard leather case. Magnificent.

Misread...

[ilsaloving]

doing its small part to forever change the world without a battery, transistor, or tube.

Because of the font, I did a double-take cause I initially misread "tube" for "lube". That certainly forever changed my perceptions of nuclear physics research. :)