burroughs_class_5 Clodoveu Davis

This is an old revision of the document!

Mechanical calculators: Burroughs Class 5

This machine was manufactured by Burroughs in the United States, at a factory in Detroit, Michigan, in 1926 or 1927 (see here). I bought it at an online auction site, in May 2019, from just two posted images. I took the chance of buying without being able to see much of the machine's condition, but this one is older, rarer in Brazil, and has a simpler mechanism, based on the Comptometer. Online sources say that Burroughs was sued for nearly copying the Comptometer's working mechanism in other machines, and then came out with the Class 5 with a few innovations, including a more compact design, aiming at desktop use. Earlier bulky Burroughs adding machines were mounted on pedestals and placed at the center of offices, to be shared among many operators.

My unit has seven missing keys, and one white “3” key too many. This was probably a replacement taken from another machine. It was locked and unresponsive, the painted aluminum (?) case sprinkled with paint drops, and all rubber feet crumbling.

Online photos of museum units and other well-preserved Class 5s

My machine has octogonal keytops for the shift functions, and a knurled lever for the stop/sub function. This lever was replaced with a regular button in later models, and the TK-style octogonal keytops also evolved into rectangular ones.

My machine also lacks the secondary control lever, to the right of the main control (mult-add-div) lever. This is strange, since the NEA has this lever, and its bottom prints symbols to indicate its function to the operator. Most other pictures of ESA-0 machines I found online also have the secondary lever. So far, I could not find an explanation for this; there is no indication that the lever was removed or broken, though.

Putting it back to work was a demanding process, since initially the motor ran continuously, and all other mechanisms were locked, including the keyboard and main rotor. It took some analysis, observation, kerosene jets and lubrication to unlock the typical parts of the FACITs that cause the most important locks. During two careful days, I made it a new power cable, then I managed to stop the ever-running motor, make the clearing mechanisms work, unlock the main rotor, release the keyboard lock, activate the adding mechanism, and finally get it to do multiplications and divisions. It now works perfectly.

I replaced the acetate windows, cutting from a transparent sheet that is used in packing kiwi fruits for sale in Brazil – go figure. I have also replaced the rubber rollers at the back with nylon cylinders. At 12 kilos, this was the first FACIT machine to have rollers to help moving it around an office desk. Next step: 3D print the missing keytops.

Multiplication in the ESA-0 is straightforward: the operator types in the multiplicator, and presses the X button. This places the multiplicator in a “mechanical memory” of sorts, the input register is cleared, and the displays remains zeroed. Then the operator inputs the multiplicand, and presses the = key. The result comes out automatically, with all rotations and shifts performed automatically. It is worth mentioning that the mechanical algorithm implemented in this machine can optimize the number of rotor turns, deciding whether to go forward or backwards so that the number of turns is minimized.

The operator can help minimizing the machine's effort. As a rule, enter first the smaller number. If the numbers have the same number of digits, there can be a more efficient multiplication order. For instance, if the operator enters 333 x 999 =, the machine stores 333, then spins the rotor with 999 in the input register three times, for each of the three decimal positions, so nine turns are required. If the operator enters 999 x 333 =, then 999 is stored, and with 333 in the main rotor, the machine performs one negative turn (-1 x 333), then shifts left three times, then does one positive turn (+1000 x 333), and completes the task with only three spins.

Division is also fully automatic: after the dividend is input and added to the left of the accumulator, the input register is cleared automatically and the divisor can be keyed in and also shifted to the left. Then the division key starts up a mechanical algorithm by which subtractions and carriage moves to the right go on automatically until the remainder is zero or the precision limit is reached.

Have a mechanical calculator stored somewhere, and want to get rid of it? Send it to me!

1951 FACIT ESA-0 s/n 327930

Front view, notice there is no secondary control lever to the right of the main lever