by Nicholas Bodley
|Friden, Incorporated, was
started in the Depression of the 1930s by Carl Bengtsson,
who came from Sweden (fairly sure of this info. and
spelling). I understand it is the custom of
Scandinavians, or was, to assume a new surname when one
reached adulthood; Mr. Bengtsson assumed the name Friden
(with an acute accent on the "e"), which I
believe means "peace". The pre-WW II
calculators had a logo with the accent.
The prewar models included the ST, which set the basic design for all future Friden mechanical calculators. It was offered in "8 X 8 X 16" and "10 X 10 X 20" models (my designations). The numbers refer to the largest multiplier and multiplicand, and the size of the resulting product, that could be handled. All Fridens had a "full" keyboard for entering all numbers except the multiplier, which was entered on a small 10-key auxiliary keyboard at the near left side of the keyboard panel.
The ST had oodles of finely-fitted countersunk cover screws. A company story told by the "old-timers" said that Carl insisted that a new machine run perfectly and pass its tests before it was lubricated; apparently the residual oil from manufacturing helped, but that may be speculation. Only after the machine was approved was it then lubricated. The ST series was quite trouble-free, and with periodic care, lasted for decades.
The ST could do all four arithmetic operations; I believe it could also add or subtract a product to/from the number in the accumulator, simply by intentionally not clearing it first.
After World War II, the STW became the most-common model; it had less-elegant housing details, the same basic mechanism, and a new logo. Except in a few details, it was much like the ST. It definitely could add/subtract a product to/from the accumulator's contents.
Friden also made an STW-like model with an added square root function, the Model SRW. The small number of added parts and the modest extent of the design changes required to add this function were notable; they had some first-rate engineers! The square root operation was just as reliable as any other, essentially.
The radicand (number you wanted the square root of) was entered into the full keyboard, and you started the calculation by pushing one of (nine?) special "live" tabulator keys that did the initial alignment of the carriage. These keys were marked with the number of decimal places you wanted in the root; there were two "5" keys, two "6" keys, etc., as I recall; 9 was max, I think.
THE "FIVES METHOD" FOR CALCULATING SQUARE ROOTS
To calculate the square root, the radicand in the main keyboard was automatically multiplied by five, putting the product in the accumulator.
The machine then cleared the main keyboard, automatically entered a lone 5 into (the leftmost column of?) the main keyboard, then went into a modified division operation. If the 5 didn't cause an overdraft, the machine automatically incremented the keyboard column to the right of the 5 by one for each cycle, so the main keyboard contained consecutively 5, 15, 25, 35, 45, etc. These are simply the first odd integers multiplied by five, but vastly simpler to deal with mechanically. That 5 was moved progressively to the right as the calculation proceeded. This was known as the "fives method" of extracting square roots; it works well on a Curta handheld mechanical calc., as I recall. (Of course, I haven't given all the details here.)
Some years after the advent of the Model SRW, Friden came out with the Model SBT, which had "back transfer", an operation that transferred the contents of the accumulator back into the keyboard. It did this by effectively placing ones in all columns of the "inner" part of the keyboard and decrementing the accumulator dials above the keyboard by one for every cycle.
As long as each dial was decrementing, another part of the keyboard was incrementing. When each dial reached zero, incrementing in that column stopped. When all numbers were transferred, the keyboard's internal state changed, the initial ones disappeared, and the new number was available.
The keystems didn't move, but there were tiny dials to show the contents of the internal part of the keyboard. Also later in Friden's history, they offered automatic squaring; I think the model designation was "STQ", but not sure. Internally, this was quite a piece of engineering; it did something not allowed for in the original design. The number to be squared was entered into the main keyboard; it was also automatically transferred serially to the auxiliary 10-key serial-entry mechanism more often used to enter the multiplier. The machine then entered a normal multiplication. Running with the covers off, you could see a long narrow cam on the end of a lever, close to the 10-key mechanism, which would swing dramatically as it entered the digits into the serial-entry register. Some people referred to it as the "baloney slicer"; it resembled a miniature war weapon, and probably could have inflicted a nasty injury.
(It's impressive to compare the months or years of engineering required to enable transferring a number between one part of a mechanical calculator and another when the original design didn't plan for it, ...to compare with the ease with which a corresponding modification can be done, especially in the days of discrete gates; if you had the fanout and spare logic inputs, you simply added some wires to the backplane. It's not quite so easy with ICs, for sure!)
The crowning glory of the Fridens was the Model SRQ, which did automatic squaring and square root. It was a combination of the mechanisms of the other machines; there was enough complexity in the control mechanism that the main mechanism's side covers needed to be bigger.
The last mechanical Fridens had a "V" in their model numbers; one was the SVE. These had a decimal-point key in the multiplier mechanism, and multiplication decimal pointing became easier. They also did "short-cut" multiplication, which multiplies by say, eight, by subtracting from zero twice, thus placing twice the complement of the multiplicand in the accumulator. These two cycles were followed by a one-place right shift of the carriage, and one add cycle; this was equivalent to 10 times the multiplicand. Plainly, this is 10 - 2 = 8, or to put it more precisely, (-2) + 10 = 8. If your multiplier was all nines, the operation was even faster, with multiple consecutive shifts and no stops fro digitation or carries.
Ordinary multiplication requires nine add cycles to multiply by nine, for instance. (This method is very practical, even recommended, on the Curta.) The multiplication mechanism was extensively redesigned, and imposed a wildly-varying torque load on the main driveshaft, which in the earlier examples of the "V-line" twisted so badly from torque overload that the basic addition and subtraction mechanisms were thrown "out of time", and carry parts wore out extremely fast.
I never found out the later history; for some period of time, the defenseless parts were made of cemented carbide. This was an awful solution to an engineering problem! I believe the development was about 10 years behind schedule. Apparently, making a larger-diameter main driveshaft would have required extensive and costly retooling, as one could guess. There were rumors about some engineering manager being really stubborn about retaining the existing diameter.
Friden also made superb (and costly) printing adding machines; they were probably second to none worldwide, although I know nothing about German adders. American designs tended not to skimp on material (perhaps because material costs were lower), and weren't over-engineered; this gave them excellent reliability and competitive cost. (I have been told that Olivetti mechanical printing calculators were unnecessarily complicated.)
Friden was the second company in the world to introduce (in 1964 (?)) an electronic desktop calculator, their model EC-130. This was a dramatic and extremely important development in calculator history; the author was intimately involved as one of the first eight field technicians. This machine had very innovative design throughout; it was extremely different from later electronic calculators in numerous respects. It had a 10-key serial-entry keyboard with unusual internal design, a CRT display that showed all four registers of a stack; it was the first calculator to use RPN, and differed only in very minor details from later H.P. calculator stack/RPN details.
The electronics was totally discrete germanium transistor-diode logic, and its numbers were stored quite serially in a wire sonic delay line. It had a selectable fixed decimal point. Display was 7-segment, with each segment individually traced out by the CRT beam. The author has described this machine in considerable detail in another article.
Perhaps a year later, Friden brought out a variant, the EC-132, which offered square root, and possibly one-key squaring.
About the time the EC-130 was introduced, Singer (of sewing machine fame) bought Friden; later developments based on the EC-130 bore the Singer logo, and weren't as interesting; some or all were printers, using a Singer/Friden-developed printer mechanism. The author left Friden around the time these were introduced. (Around the time this was written, the author also wrote an detailed description of the insides of the EC-130.)