Friden EC-130 Electronic Calculator

This is a truly wonderful machine. It is one of the earliest all-electronic calculators, and is generally regarded as the first transistorized electronic calculator. A few other calculator manufacturers in Europe and Japan claim that they were the first to develop an all-transistor calculator, but the simple fact is that Friden announced the Friden 130 nearly six months before these other manufacturers even displayed prototypes of their transistorized calculators. Earlier electronic calculators used relays or gas-discharge tubes, such as the Casio 14-A relay calculator (1956), or the thyratron tube-based Sumlock Comptometer/Bell Punch Anita C/VIII (1961). Along with being the first all-transistor calculator, the Friden 130 pioneered the use of Reverse Polish Notation (RPN), a method of entering math problems using a multi-register stack. RPN logic made complex calculations easier to perform without having to write down intermediate results and re-enter them into the calculator when needed. Friden continued the use of this stack-based method of calculating in its second generation of calculators (an example being the Friden 1162), but for some reason the principle was never patented by Friden. The use of the RPN methology applied to electronic calculators was originally patented by Mathatronics, in their Mathatron calculators, which hid the RPN implementation from the user, but used it internally to carry otu the mathematical operations. Later, when Mathatronics was liquidated after being purchased by Barry Wright Corp., Hewlett Packard purchased the rights to the patent, as a means to protect their use of the RPN method in their first electronic calculator, the HP 9100A. HP's use of RPN proved so successful that HP still uses it on a good many of their calculators to this day. Later calculators from Singer/Friden (111x-Series and beyond) abandoned RPN in favor of more conventional arithmetic logic.

The story behind the development of the Friden 130 is somewhat shrouded in the mists of time, but the general situation appears to be that Friden management realized that the days of the electro-mechanical calculator were numbered sometime in the early 1960's. Sumlock Comptometer, Ltd. had set the calculator world on end when it introduced the first electronic calculator in 1961, and many makers of electro-mechanical and relay calculators realized that the future of calculating was with electronics. Electronic circuitry is fast, quiet, and reliable. Electro-mechanical calculators were noisy, slow, and had lots of moving parts which required regular maintenance to continue to operate properly. Relay calculators, while less noisy, and faster than electro-mechanical machines, also had moving parts (relays rely on mechanical movement to close switch contacts), and also had myriad switch contacts that wore over time and required periodic maintenance and adjustment. The problem for the companies that sold electro-mechnical or relay calculators as their bread-and-butter business was that they had brilliant mechanical and switching engineers, but little in the way of the electronic engineers needed to design a practical electronic calculator. Most of the electrical engineers were working in military, space, aviation, or communications technology, and those markets snatched up all of the electronic engineers as soon as they graduated from college, making the good electronic engineers a rare commodity.

Robert Ragen, 1988
Image Courtesy of Dick Ahrens

By researching patent information, along with some first-hand information from a number of former Friden employees of the time, it has become clear that the major contributor to the architecture and design of the Friden 130 was a brilliant electronic engineer named Robert Ragen. Ragen worked for Friden prior to the development of the Friden 130 on government-related electronics projects that Friden was involved in. Rumor has it that the work involved secret electronic communications systems development for the National Security Agency. Those who would know even to this day are not forthcoming with information relating to this work, so one can assume that whatever it was, it was deemed critical to the defense and security of the United States. Whatever the work, it provided Friden with the electronic stepping stone on the path towards developing a fully-electronic calculating machine to take over in place of the mechanical masterpieces that Friden was so famous for.

After the work for the government wound down, the timing was right for Friden to jump headlong into developing an electronic replacement for mechanical calculators. Ragen, along with team of other Friden engineers and craftsmen, were put to task to develop a prototype of the Friden electronic calculator. In a relatively short period of time, a prototype was developed with electronics that fit in a box just a shade bigger than today's compact refrigerators. Sitting on top of this box was a console that provided the user interface for the calculator, consisting of the keyboard and CRT-display. The first prototype utilized a magnetic drum (Bryant Model C-105) to both generate the master clock frequency for the machine (via dedicated pre-recorded timing tracks), and to store the working registers of the calculator. Later prototypes replaced the magnetic drum (which was a very expensive and sensitive piece of equipment) with a megnetostrictive delay line, which was a much less costly and not nearly as sensitive as the magnetic drum. These proof of concept prototype calculators were far from practical as a useful piece of office equipment, but it served to demonstrate the concept -- Friden had the means to build an all-electronic calculator. This prototype had all of the functionality of the Friden 130, including RPN logic, CRT display, magnetostrictive delay-line storage, and transistorized construction.

Prototype Predecessor to Friden 130 (From US Patent #3546676)

Once the large-scale prototype was completed, the next task was to shrink the cabinet of components that made up the electronics of the machine down to a practical and usable desktop-sized unit. A couple of factors that contributed to making this job a little easier were that the electronics in the prototype were designed with very liberal design rules. Components were not tightly packed, circuit boards had wiring only on one side, and interconnections were widely spaced. Another factor was that the prototype was hand-made, which generally makes for a less space-efficient design. After all, a prototype is made to show that a concept is workable -- the effort is placed on making the idea a reality, rather than trying to optimize it for manufacture. In a mass-produced calculator, the space between components can be dramatically reduced, circuit boards can have wiring on both sides, and interconnections can be made much more dense. The problem was, Friden did not have much experience making complex circuit boards. In order to pack all of the electronics needed into a desktop package, Friden needed to be able to manufacture circuit boards with traces on both sides. This means that there has to be a way to provide connections through the circuit board to allow circuitry and traces on one side of the board to connect to traces on the other side. Such connections are called feed-throughs. For the small and fairly simple circuit boards needed in their electro-mechanical calculators, Friden had set up a circuit board manufacturing facility using a rather unique machine to etch the circuit board traces. This technology worked nicely for the simple single-sided circuit boards needed in the electro-mechanical calculators, but the 130 needed complex circuit boards with wiring on both sides of the board. This presented a problem. The feed-through holes tended to short out the machine that did the etching. As it turned out, the etching machine was used to make the single-sided boards in the calculator's power supply, but the logic boards of the calculator were farmed out to a specialty firm that made the circuit boards for Friden.

Larry Kramer, Friden Chief Draftsman (right) and Dick Ahrens, Electronics Engineer(left) pouring over the (huge) master schematic for the Friden 130
Sincere thanks to Dick Ahrens for donation of the Original photo and negative to the Old Calculator Web Museum

In the early part of 1963, a desktop prototypes of the original design of the calculator was ready, and exhaustive testing was begun. A disturbing problem was found where the calculator would inexplicably and randomly deliver incorrect results. Intense efforts went into finding the problem, which was finally traced to the high voltage section of the electronics (related to driving the CRT display). High electrostatic charge levels would build up, causing discharges that would make the calculator malfunction intermittently. Given that these calculators were expected to deliver accurate results all of the time, such a problem was intolerable. Efforts to identify, isolate and eliminate the problem resulted in the delay of the introduction of Friden's first electronic calculator for almost six months.

The First Pre-Production Prototype Friden 130, August 1963
Click on image for a more detailed view
Original photo donated to the Old Calculator Web Museum through the generosity of Dick Ahrens

In the spring of 1963, Friden had worked through the difficulties, and had a viable desktop (though it took up a substantial part of a normal desk top) calculator ready for market. Amidst significant fanfare, an early Friden 130 prototype was shown at a business machines exposition in June of 1963. The interest was tremendous -- pre-booked orders started rolling in like hotcakes, even though the 130 was quite expensive compared to the electromechanical machines of the day, with an initial retail price of $2,195. This was at a time where a desktop electro-mechanical calculator with about the same functionality could be purchased for around $500. The prototype machines shown early on in the life of the EC-130 had decimal point settings of 0, 2, 5, 9, and 13 digits behind the decimal point. When the EC-130 went into full production, the decimal point settings were changed for some (currently unknown) reason to 0, 2, 5, 7, and 13 digits behind the decimal. In November of 1964 (November 20, 1945), a service letter (Calculator Release 0024) was issued by Friden headquarters indicating that a special batch of decimal point setting switches was procured with the original prototype calculator's selections of 0, 2, 5, 9 and 13 digits behind the decimal. This was due to a significant number of requests from purchasers of early EC-130's that had noted the change, and wanted their machine to match the decimal point selections speficied when they ordered the machine. Friden offered installing the replacement switch in the field at a cost of $45.00 to the customer to satisfy these requests.

Early Friden 130 Marketing Trinket - Friden 130 Playing Cards
Donation of Original Item Courtesy of Dick Ahrens

As more and more EC-130's made it into the hands of customers, it became apparent that there were some problems. Many of the early production Friden 130's had some fairly serious problems with reliability due to problems with the circuit boards. The firm that manufactured the circuit boards for Friden was having difficulty with the plated feed-through holes which provided connections from one side of the circuit board to the other. To fix this problem required tedious hand-soldering of feed-through connections in the field by Friden service technicians. Friden made good on these problems, providing highly skilled service technicians who spent a lot of time in training to be able to repair any faulty EC-130 quickly and efficiently. In fairly short order, the circuit board manufacturer solved the feed-through problems, and as the machines in the field that had the problems were repaired, the problems subsided. Even with some of the problems early-on, customers were delighted with this amazing machine that could quietly and very quickly solve their difficult mathematics problems.

An early prototype circuit board (unpopulated, used for flow-solder testing) from the Friden 130

Another early prototype circuit board from the Friden 130

Click on either image for more details.
Sincere thanks to Dick Ahrens for donation of these boards to the museum.

The architecture that Bob Ragen devised was very unique compared to designs of other calculating machines either prior-to, or after the 130. The 130 used an unusual arrangement of four (later, three) interconnected up or down counters (known as A, B, C, and D) and control circuitry, along with a novel way of storing data in the magnetostrictive delay line. The design was very elegant, minimizing the component count needed to implement the logic of the machine, and easily suited to the purpose of performing the basic four math functions. In the early 1960's when this machine was designed, transistors were still rather expensive. Minimizing the number of transistors meant that the cost to manufacture was lower than less-efficient designs, allowing more margin to be built into the final sales price of the product, while still providing a product that would be priced competitively in the marketplace. Ragen's design was quite a departure from the design of the Sumlock ANITA calculators, which at the time were the only other electronic calculators in the marketplace. The ANITA calculators were essentially electronic implementations of mechanical calculators. They operated in decimal, using ten-stage electronic counters much like the ten-step mechanical counters in rotary calculators. The Friden 130 uses a completely different approach, utilizing counters and trains of pulses stored in a magnetostrictive delay line to perform mathematical operations.

Block diagram of "Four Counter" Friden 130 Architecture
Click on Image for more detailed view

Sometime around mid-1965, it became apparent that one of the four counters (in particular, the B counter) in the machine could be removed, simplifying the machine and reducing the component count. Some redesign of the machine was necessary, but it was worth it to implement the change in order to reduce cost and improve reliability. Friden EC-130 calculators prior to serial number 8500 were "four counter" calculators, and from serial number 8500 and beyond, the "three counter" architecture was substituted. Thus, the Friden 130 exhibited here is the three counter design. Friden service technicians had to be aware of this architectural change when servicing calculators in the field, as there were differences in circuit boards between the two designs of the 130. Along with being able to discriminate between the earlier four-counter and later three-counter EC-130s by serial number, it is also possible to know the architecture of the calculator without even looking at the serial number (located on the bottom of the calculator). The four-counter EC-130s displays the bottom-most register in the display in intensified form, while the three-counter machines do not have this feature. Originally, it was thought that intensifying the display of the bottom-most register (where all calculation results are displayed) would make it easier on the operator. The intensity was increased by actually displaying the bottom register twice during each display cycle. As part of the simplification of the logic involved in switching to the three-counter architecture, the feature was dropped. The machine exhibited here is a three-counter machine, indicated by its serial number of 12692, and also by the display (see image below), which does not have the bottom register intensified.

Inside the Friden 130

The Friden 130 uses diode-resistor "OR" and "AND" logic gates, with transistor-based inverter, buffer, and flip-flop devices. It performs math operations in bit-serial form, using the magnetostrictive delay line as the medium for storing its working registers. Logic levels are 0 Volts representing logic 1, and -12 Volts (nominally) representing logic 0. The delay line input transducer is driven with a pulse of approximately 20V, and by the time the signal makes it to the other end of the delay line, the voltage induced in the transducer is approximately 35mV, or 35/1000th's of a volt. Digits are stored within the delay line as a series of pulses arranged in groups for each digit. Zero pulses represents a zero, and nine pulses represent a nine, with the numbers in-between represented by a number of pulses matching the number. As the pulses exit the delay line, they are amplified and fed into the counters (the A and/or D counters), which count the number of pulses in the digit to form a unique five-bit identifier that represents the number. The counter registers are not configured as counters in the usual binary sense. They are instead configured as five stage switch-tail shift registers, such that they count in a sequence of shifting 1's. For example; 0 is represented as 00000; 1 as 10000; 2 as 11000; 3 as 11100; 4 as 11110; 5 as 11111; with 6 as 01111, and ending with 9 as 00001. With five flip flops, each counter can represent the numbers zero through nine as unique combinations of bit patterns.

The delay line is a very interesting method of providing working storage registers for a calculator. Given that transistors were still rather expensive, some other means for storing the working registers of the calculator was needed. A little math shows how quickly the component count grows if the working registers of the calculator were to be implemented in circuitry. It takes at least two transistors to make a flip-flop, along with a complement of resistors, capacitors, and diodes. A flip-flop is essentially a 1-bit storage register. With 13 digits to store, and with each digit taking 5 bits, that means that there would have to be 65 flip flops, or a minimum of 130 transistors, to store one register in the stack. The 130 has 4 registers in the stack, plus one for the memory register. This would have taken over 600 transistors, along with hundreds of resistors, capacitors, and diodes, just to provide the storage for the registers. Such a design would have been prohibitive both in terms of cost and space required.

Ragen's solution to this problem was to leverage technology used in early computers (from the late 1940's through early '50's) to store the content of the working registers of the calculator. Before the advent of ferrite-core magnetic memory devices, one particular means of storage for electronic computers used long narrow tubes filled with Mercury with a transducer at each end. The bits of data took the form of sonic disturbances created by the transducer at one end of the tube. These disturbances propogated through the mercury at a fixed rate. The bits were sent through the mercury a bit at a time in serial fashion, and were constantly re-circulated through the tube like a big shift-register. When bits were needed, they were siphoned off by a transducer which converted the acoustical pulses to pulses of electrical energy, which were amplified and sent into the arithmetic unit bit at a time, where the appropriate operations were performed and the results pushed back into the bitstream circulating through the Mercury. The 130 uses a similar method, but rather than using exotic (and poisonous) materials like Mercury, a carefully-selected type of wire (made of a Nickel alloy) is used that holds the bits as tiny twists (torque variations) in the wire that move along the wire from one end to the other. The phenomenon is much like the wave that travels down a length of rope when you quickly whip one end of the rope. A transducer at one end of the wire places a twisting torque pulse on the wire which travels through the wire and is registered at the other end by a similar transducer. By continuously circulating these torque variations through the wire, the wire becomes the storage medium for the bits, and far less circuitry is required to maintain all of the bits that the machine needs to operate. In the Friden 130, the delay line takes the form of a large number of loops of this special wire, that, if unwound, would be about 50 feet in length. A pulse entering at one end of the wire will come out the other end in approximately 5 milliseconds, or stated otherwise, it takes about 5/1000ths of a second for a pulse to make its way from one end of the delay wire to the other. The wire is carefully strung in a series of spirals inside a sealed metal enclosure that takes up most of the bottom part of the chassis of the calculator.

A Closer View of the Card Cage

The circuitry of the three counter Friden 130 is contained on a total of seven circuit boards, each of which is about 12 by 5 inches. The boards, as expected, are packed quite densely with components. Most of the transistors are Germanuim (Silicon transistors were just becoming commonly available, and were significantly more expensive) junction transistors, mostly of the PNP type (predominantly 2N1305). The circuit boards plug into a backplane via edge connectors. However, the backplane connections weren't sufficient for all of the inter-board connections needed. Three groups of two boards each are wired together with many hand-soldered jumpers to provide the additional inter-board connections needed. The boards are put together in pairs of two, with wire jumpers connecting each pair of boards across the top edge of the boards.

The CRT, CRT Drive, and Power Supply Circuitry

Another small circuit board mounted to the aluminum chassis seperating the card cage from the CRT subsystem contains the drive circuitry for the CRT display, and a large circuit board situated underneath the CRT tube contains the power supply electronics, including a capacitor-diode voltage multiplier circuit that produces the high voltage (Approx. 2400 Volts) for the Westinghouse-made 5DEP1 electrostatic deflection CRT tube. The keyboard assembly also has a small circuit board with two transistors and a number of discrete components that provides signal conditioning for the keyboard outputs. The power supply provides 6.3V for the filament in the CRT, +6V and -12V DC as logic supplies, +80V for CRT deflection amps and delay line voltages, and approx. 2400V DC for the high voltage for the CRT.

A New-Old Stock (NOS) Westinghouse 5DEP1 CRT Tube, Circa 1967

The 130 has a 4-level RPN stack, with all four levels visible on the display. The content of the single store/recall memory register is not shown on the display. Digits are drawn in vector form on the display in a modified "pieces of eight" seven-segment form. Like all RPN-logic calculators, the Friden 130 has an [ENTER] key, which is used to enter the first number in an operation into the bottom register of the stack. A [REPEAT] key duplicates the number at the bottom of the stack by pushing the number in the bottom register in the stack up one, then duplicating it in the bottom register in the stack. This repeat function makes squaring much easier, allowing the user to calculate a square without having to re-enter the number (for example, 123² would be entered as [1], [2], [3], [REPEAT], [X]). The [CHANGE SIGN] key toggles the sign of the number in the bottom of the stack. Numeric entry occurs in the bottom-most register of the stack. The standard four math keys perform their respective operations on the bottom two registers of the stack, with the stack shifted down after the operation is complete, and the result stored in the bottom-most register of the stack. When the stack is shifted down, the top-most register is set to zero. The [CLEAR ALL] key clears the stack, and the [CLEAR ENTRY] key clears the bottom register of the stack. The [STORE] key copies the bottom register of the stack into the memory register, clearing the bottom register of the stack. The [RECALL] key pushes the stack up, and copies the number in the memory register into the bottom register in the stack.

Friden 130 CRT Display

The machine handles 13 digit numbers, with thumbwheel-selectable fixed decimal point location. The keyboard uses a unique combination of electrical and mechanical construction. The keys actuate magnetic reed switches through a mechanical mechanism that encodes the keyboard keys into a binary code for the electronics. The keyboard is also mechanically interlocked by a mechanism controlled by the electronics. When an operation is performed, the function key locks down and isn't released until the operation is completed. This prevents the user from getting ahead of the machine. The keys are also mechanically interlocked so that it is impossible to press more than one key at a time. The machine performs only the basic add, subtract, multiply and divide operations, and has a single store/recall memory register. Shortly after the 130 was introduced, Friden announced a follow-on machine, the 132, which added a square-root function and provided more decimal point position selections.

Friden 130 Keyboard Layout (Click photo for detailed keyboard layout)

The 130 can take up to 2 seconds to perform difficult divisions, such as the "all-nines" (9999999999999) divided by 1 calculation. While this may seem a bit slow compared to what folks are used to today, this was orders of magnitude faster than the electro-mechanical calculators being used at the time, not to mention the fact that the 130 performed such operations with almost magical silence. The basic clock frequency of the four counter 130 is 666KHz (the clock rate was changed to 333Khz in the three register machine due to the elimination of one stage in the chain of divider flip-flops), which, for the time, is a relatively fast clock rate for Germanium-based transistor logic. The clock frequency is divided down by a chain of flip-flops that create the various master timing signals that orchestrate the operation of the calculator. During math operations, the display is blanked. If the machine is commanded to divide by zero, the display blanks and stays that way with the electronics running in a futile attempt to repeatedly subtract zero from the dividend. The OVER FLOW indicator does not light to show this condition as an error, which can lead one to wonder if the calculator has failed. Pressing the [OVER FLOW/LOCK] key, or the [CLEAR ALL] key stops the futility and returns the calculator to normal operation.

The [OVER FLOW/LOCK] Key Indicating an Overflow Condition