+Home     Museum     Wanted     Specs     Previous     Next  

Cintra/Tektronix Model 909 Desktop Scientific Calculator


This exhibit is dedicated in perpetuity to Dr. Irwin Wunderman (4/24/1931-7/23/2005), the principal designer of the Cintra calculators, and founder of Cintra Corporation.


Updated 3/12/2012

This great old machine has a real spot in my heart. It's a machine that I saw for the first time in an advertisement for Cintra in an old journal of scientific instruments some time ago. When I saw it in that old ad, I wanted someday to find one of these machines. As fortune would have it, thanks to the thoughtfulness of a party out there on the Internet, I had the opportunity to get this one. It is in very good and fully operational condition, an amazing testimony both to the original design and manufacture of the machine, and to the folks that have owned it and taken good care of it over the years.

Profile View of Cintra/Tek 909

Before I get into the history of how this machine came to be, I should qualify things by saying that some of the information here on the history of Cintra is based on indirect observations and some word of mouth stories that have been related to me. Also, I am deeply indebted to Dr. Irwin Wunderman, the founder and President of Cintra, who provided a great deal of extremely enlightening information. If anyone out there worked for Cintra or Tektronix during the calculator days, I'd love to hear from you.

Cintra's Initial Employees, Late 1967
Left: Dr. Irwin Wunderman (Founder & President)
3rd from Left: Michael J. Cochran (Engineering)
4th from Left: Dr. Julius J. Muray (Vice-President)
Jim Shimizu
If you can identify any of the others, please EMail (see link at top of page) me.
Photo Courtesy Dr. Irwin Wunderman

The history behind the development Cintra and its calculators is quite interesting. In November of 1967, Dr. Irwin Wunderman left Hewlett Packard (where he was an engineer involved in the early development of an electronic calculator, which became the HP-9100A, for Hewlett Packard to market) to strike out on his own to create a company that would produce precision energy measurement instrumentation. Dr. Wunderman beleived that there was a ready marketplace for digital instrumentation that would measure optical energy very accurately, and over a wide range. He also believed that these devices should provide a digital interface to other equipment to allow the instruments to serve as data acquisition devices.

Irwin Wunderman's Garage Workshop, the first home of Cintra
Photo Courtesy Dr. Irwin Wunderman

With his own money Mr. Wunderman set up shop in his garage at 655 Eunice Avenue in Mountain View, California. The garage, a former carriage house, had quite a history through the years, including serving as a gangster-operated bordello, a dog racing facility, a speakeasy during the 20's and 30's, and lastly, as a distillery. This colorful building became the the headquarters for Cintra, Wunderman's electronic instrumentation manufacturing company. Wunderman began work alone, but over a period of time recruited five other employees (three engineers, a business development manager, and an operations manager) who worked together to form the backbone of the company. Wunderman's enterprise caught the attention of San Leandro, California-based Physics International, a company involved with many government contracts related to high-energy physics. Cintra's digital instrumentation proved a valuable tool for Physics International. A deal was struck, and in return for an investment from Physics International, Cintra became a subsidiary of the company.

As Cintra's line of digital instrumentation grew, the company became more and more involved in interfacing its instrumentation to various types of computing equipment, to provide data logging and analysis of the measurements made. Over time, the idea came about to develop a desktop calculating instrument that would also provide a digital interface to allow the instrumentation to connect to the calculator, providing a means by which the data from the instrumentation could be processed in an automated fashion. From this idea grew what became the Cintra 909 Scientist electronic calculator. The calculator could serve as the center of a data acquisition and analysis system that could automate the collection and processing of data from Cintra's energy measurement instrumentation.

At the time Cintra began the development of its calculator, there were already a few well-established players in the high-end electronic calculator market. Hewlett Packard, Wang, and a few others had ventured into making calculators designed for complex engineering and scientific applications. Cintra's goal was to build a high-end calculator that was easy to use, and could easily interface to it's own measurement equipment. The concept was to make the automation of data gathering and analysis as simple as possible. It was decided that the Cintra calculator should use entirely algebraic notation, as opposed to the somewhat daunting Reverse Polish stack arithmetic of the Hewlett Packard calculators, and Wang's rather complex multi-register architecture. The use of algebraic logic would allow calculations to be performed on the Cintra calculator just as they are written on paper -- a concept initially pioneered by Mathatronics with their Mathatron calculators in the mid-1960's.

The result of Cintra's engineering efforts was the Cintra 909 Scientist calculator, introduced in November of 1969. Unfortunately, Cintra ended up having some component problems in their early production calculators. Certain aspects of the design of the calculators required tight electrical specifications for transistors used in the design. A large batch of transistors were ordered to get production going, but it turned out that the electrical specifications of the transistors varied considerably, resulting in some calculators that would work, and others that wouldn't. Unfortunately, this wasn't caught during production, and the machine would end up not functioning properly when delivered to the customer. This was a major problem, as Cintra didn't have an established service organization to fix the machines for the customers. This certainly did not get Cintra off to a stellar start in their efforts to gain a share of the high-end calculator market.

The reality of the situation was that when the Cintra calculator was first conceptualized, the designs were a major leap ahead in calculator technology, with full algebraic logic, scientific notation, programmability in the form of an 85-step 'user defined function' (optionally expandable to 256 steps), large memory register capacity (base model containing 26 memory registers, with an option pushing the memory to 100 registers), high speed, I/O interfacing capabilities, and use of new TTL (Transistor-Transistor Logic) integrated circuit technology, along with some custom-made Large Scale Integration (LSI) devices. However, because of the amount of time that it took for the calculators to reach the marketplace, combined with the early electronic component problems with the machines, Hewlett Packard had, for all intents and purposes, captured the market for high-end programmable scientific calculations with the 9100A and 9100B calculators. Wang Laboratories, Mathatronics, and others filled in the rest of the market, leaving little room for up and coming players to get a decent share. To add insult to injury, when the Cintra machines debuted, HP was already hard at work on its even more capable next-generation calculators, the 9800-series. Wang was also busy, and around the time that the Cintra calculators were introduced, had debuted the computer-like 700-series machines (For example, the Wang 720C). The competition from Hewlett Packard and Wang was very tough. Both companies had their well-established marketing and sales departments hammering the marketplace, as well as large customer support and service organizations to provide support for their products after the sale. Cintra, though their machine was technically strong and reasonably priced (with a base price of $3,780), simply didn't have the name recognition, sales presence, and customer support infrastructure that HP and Wang had going for them. As a result of all of these pressures, in early 1971, Cintra management realized that something had to happen, or it would be curtains for the company.

Around that time, a company called Tektronix came to the rescue. Tektronix, the famous Beaverton, Oregon-based electronics test equipment manufacturer, was looking to get into the calculator marketplace, since it's arch-rival, Hewlett Packard had jumped in to the market in a big way with the 9100A and 9100B calculators. Tektronix found Cintra in need of some serious help, and Tek wanted digital design expertise to infuse into its new "Information Display Division". A deal was struck for Tektronix to acquire Cintra. The deal to acquire Cintra was announced publically on May 7, 1971, just 11 fateful days before the untimely death of one of Tek's founding fathers, Jack Murdock, in a light plane accident. Amid the shock and sadness associated with the death of Murdock, Tektronix President Earl Wantland appointed Frank Elardo(1929-2006) the general manager of Tektronix' new Calculator Products Division. Immediately, the existing Cintra calculators were re-badged with the Tektronix name. While marketing, selling, and servicing the Cintra-designed machines to get a 'foot in the door' of the calculator marketplace, Elardo put the former Cintra engineers, along with some engineers from Tek, to the task of building a next generation of electronic calculator, efforts that resulted in the Tektronix Model 21 and Model 31 calculators.

The realities of the fast rate of change in integrated circuit technology, along with the early market penetration by HP, Wang, and other larger manufacturers, made it difficult, even for a giant like Tektronix to make money selling high-end calculators. The price of computing equipment (e.g., mini computers) was also coming down at a rapid pace, furthering the price pressure on high-end calculators. The result was that Tektronix found that there simply wasn't enough profit potential in the calculator marketplace, and, by late 1974, the company bailed out of the calculator business, ending the timeline for Cintra and Tektronix as far as calculators were concerned. Even though the example of the machine shown here has the Tektronix badge on it, I'm going to refer to the machine as the "Cintra/Tektronix 909" simply because all of the design care and engineering that went into it was really the result of Cintra. For this reason, I give Cintra top billing.

One engineer in particular who worked at Cintra ended up being a major contributor to the advancement of calculator technology. Michael J. Cochran was one of the original designers of the Cintra calculators, and holds a number of patents associated with the ideas behind the development of the machines. When Cintra began showing signs of difficulty, Mr. Cochran left, moving from California to Texas, going to work for Texas Instruments. Mr. Cochran ended up being one of TI's most prolific engineers, aiding in the development of countless technologies that TI used to gain the strong foothold in the calculator market that it still holds today, including development of the Texas Instruments SR-60 calculator exhibited in this museum.

The Five Circuit Boards that make up the Tektronix/Cintra 909

The 909 is an extremely complex machine. In total, there are 329 integrated circuit devices in the machine! This makes the 909 one of the more complex (at least in sheer volume of components) calculators that I've run across, and most certainly one of the more complex machines of it's time. The machine benefits from some early custom LSI devices made by Electronic Arrays, one of the early manufacturers of Large-Scale Integrated Circuits. A total of ten LSI devices (four CPN 3130-30, along with CPN 3130-31 through -35, and CPN 3130-37) apparently provide ALU (Arithmetic Logic Unit) and sequencing functions. If anyone out there has any familiarity with this particular series of custom IC's made by Electronic Arrays, I would be most apppreciative if you would drop me an EMail. These LSI's combine with a sea of mostly Texas Instruments-made 7400-series TTL small and medium-scale IC's, one of the few machines designed in the late 1960's that I've found that uses 7400-series TTL logic. Sprinkled amongst all of the 7400-series TTL is a mix of various DTL small-scale devices. The components in the machine are spread across five main circuit boards that plug into a printed circuit backplane. A sixth smaller board that contains the Nixie tubes and their associated driver circuitry plugs into one of the main boards. The architecture of the machine appears to follow the notion of a microcoded general purpose CPU, with the microcode that makes up the functionality of the calculator burned into ROM (which likely is contained in some of the Electronic Arrays chips). As a result of this design methodology, the sister machine to the 909, the Cintra/Tektronix 911 Statistician, has identical circuitry, with the difference between the two models being the microcode stored in the ROMs, and the functions available on the keyboard. According to a product introduction brochure, the only visible difference between the Cintra/Tek 909 and the 911 is the function of four keys on the keyboard. On the 911, the Log(x), tan(x), hyper, and Pi keys are replaced with (respectively) x! (factorial), Σ K00, Σ K01, and K02+K03=K02. The Σ functions simply add the current content of the display into the K01 or K02 register.

There were four different versions of the 909 (and 911) avaialble, which varied depending on the amount of memory provided for program and memory register storage. The base unit was designated the Model 909, which offered 26 memory registers (00 to 25), and 85 steps of program storage. The model 909-01 upped the program step storage capacity to 256 steps, but still only having 26 memory registers. The 909-02 provided a full 100 memory registers (00 to 99), but had the base 85 steps of program memory. The 909-03 was the "maxxed-out" machine, with 100 memory registers and 256 steps of program storage space. The machine exhibited here is the Model 909-01, with 256 steps of program storage, and 26 memory registers. Original Cintra documentation indicates that the 911 Statistician also had the same options for memory and program step storage.

"CINTRA" Moulded into Base Casting

The construction of the machine is very beefy, with the machine weighing a hefty 25 pounds. The base of the machine is a serious cast-iron piece (into which "Cintra" and "Made in USA" are cast, even though the machine sports the Tektronix badge) to which the backplane board attaches, as well as the structure to support the card cage, power supply, and display. The base casting also serves as a large heatsink for the power transistors involved in power supply supply regulation circuitry. Also attached to the base casting is a sturdy fold-out carrying handle that makes it convenient (though the machine is a bit on the heavy side for easy lugging) to carry around. Cintra designers apparently felt that the heft of the cast-iron base added to the quality perception of the machine. In those days, light weight wasn't necessarily a big deciding factor in high-end calculator purchasing.

The upper half of the case is made of a thick and very durable plastic. The keyboard assembly consists of a circuit board that screws into a stout aluminum support frame (that doubles as the keyboard bezel) in the plastic of the upper case. The keyboard connects into the main electronics via a 14-pin DIP-header plug on the end of a ribbon cable that the other end of connects into the backplane board. The keyboard is an expensive looking assembly, using Raytheon-made magnetic-reed switch modules that look almost military-spec. The keyboard array is scanned, using a couple of TTL IC's on the keyboard circuit board, along with a separate transistor for each key switch. The display bezel is a heavy aluminum casting, into which a red plastic lens is set for the National Laboratories-made (the same vendor that Wang Laboratories used for their 200/300-Series calculator keyboard/display units) Nixie tubes to shine through. The power supply of the machine is a rather complex linear transistor-regulated supply. With all of the small-scale TTL devices running in the machine (original 7400-series TTL was rather power-hungry), the supply has to be able to provide a sizable amount of current. As a result, the machine dissipates enough heat that a rather large (and somewhat noisy) fan is required to keep the internal temperature of the machine under control.

Internal View of Tektronix/Cintra 909

The Cintra/Tektronix 909 is a full algebraic logic machine. The only other machines that worked this way at the time were the Mathatron calculators made by Barry Wright/Mathatronics. It is interesting to note that it Mathatronics had patent protection on their implementation of algebraic logic calculating machines, patents that HP kept a close eye on when developing the 9100 calculators. Mathatronics' algebraic logic is mentioned in the patent that covers many of the concepts on the operation of the 9100 calculators, and may be the reason why HP chose the RPN (Reverse Polish Notation) type of logic for their calculators - avoiding any potential patent conflicts with Mathatronics. As it turned out later, Hewlett Packard purchased the patent rights (US Patent Number 3,996,562) for the Mathatron calculator's algebraic processing after the parent company of Mathatronics, Barry Wright Corporation, decided that there just wasn't enough money in the calculator business and shut down the Mathatronics division. Cintra used different methods for handling algebraic logic, and thus avoided any patent infringement problems, patenting their own method for processing algebraic expressions under US Patent Number 3,720,820 (assumed by Tektronix after Cintra was purchased by Tek).

Interestingly, the Cintra engineer (Michael Cochran) that developed the algebraic programming for the Cintra calculators, left Tektronix after the Cintra buy-out, and went to Texas Instruments and spearheaded TI's development of advanced calculators. Once the TI SR-52 hand-held scientific calculator was introduced, with full algebraic operation, TI sent a team of lawyers to Tektronix, accusing Tektronix of violating a patent that TI had on file relating to the use of parenthesis keys on the SR-52. The lawyers apparently didn't do all of their homework, as Tektronix had acquired Cintra's patents for the algebraic calculator, and the new Tektronix Model 21 calculator (which was essentially a re-implementation of the Cintra 909 utilizing Large Scale Integration IC's), had parenthesis keys (as did the 909/911), thus alerting TI's legal department. A lot of negotiation was done, and it turned out that the principles of Cochran's developments at Cintra, and the work done at TI, were along similar lines at different times, and an agreement was made between Tektronix and Texas Instruments that the technology would be cross-licensed (e.g., TI would not have to pay royalties to Tektronix, and Tektronix would not have to pay royalties to TI) at zero-cost to either company.

The Display Circuit Board

The distinguishing characteristic of algebraic logic calculators are parenthesis keys, and an '=' key that generates the result for any calculation. Math problems are typed into the machine the same way as they would be written on paper. This made the 909 very easy for untrained users, in contrast to the mental gymnastics required to use HP's RPN calculators, and arithmetic logic calculators from other makers. The 909 seems to be able to handle parentheses nested to a very deep level -- I haven't been able to find a limit as of yet, but in experimenting with the machine, I've nested parenthetic expressions to at least 100 levels deep and the machine delivers the correct answer [e.g., 1+(1+(1+....(1+1))))...)=].

Keyboard (left & center/right)

The 909 has a large compliment of scientific functions, including some that I've not seen on other machines. Along with the standard four functions, the machine has square root, 1/x, x2, natural Log, base 10 Log, ex, xy, full trigonometric functions including hyperbolic variants, a function for converting degrees to radians and back, a function for returning the integer portion of a number (INT), and a function for solving triangles (e.g., calculating the square root of x2+y2). The machine also has the 'Pi' constant built in that can be recalled to the display with the touch of a single key.

The Nixie Tubes of the 909 in Operation

The 909 is a full floating-point machine. It has ten significant digits of capacity, and when the number to be displayed is larger than ten significant digits, the calculator shifts the display to scientific notation, with an exponent ranging from -99 to +99. The display uses rather large Nixie tubes, with digits that are about 5/8" tall, making the display very readable even from a significant distance. At the left end of the display, a special sign Nixie, that contains only a "+" and a "-", indicates the overall sign of the number in the display. A similar tube is used to indicate the sign of the exponent. A special and very unique indicator, situated between the main numeric display and the exponent display, lights up "X10" when the calculator is displaying a number in scientific notation. The machine has two non-displayed guard digits to add to its accuracy. The machine does not perform leading or trailing zero suppression. When overflow or error conditions occur, the display flashes on and off at about a 2 cycle-per-second rate. Overflow and error conditions do not 'lock up' the calculator --- the display just keeps flashing to notify the user that any displayed results may no longer be accurate due to error or overflow conditions.

The Cintra/Tek 909 Keyboard Assembly

The base machine has access to 26 memory registers via the "K()" key. To store a number into one of the memory registers, the "=" key is pressed, followed by the "K()" key, followed by a two digit entry from 00 through 25 that indicates the address of the memory register the number is to be stored into. To recall the content of a memory register, the "K()" key is pressed, followed by the two-digit address of the register, and the number is immediately recalled. Attempting to access memory registers above 25 on machines not equipped with optional memory expansion can result in some odd behavior --- such as the display blanking all but some decimal points for about 1 second, followed by the 'blinking Nixies' error indication. A memory upgrade was available for the machine (there is an empty slot in the card cage) that bumped the memory capacity to a full 100 memory registers (00 through 99). The memory function also supports indirect addressing. A double-press of the "K()" key before entering a memory register number causes the content of the memory register to be used as an address to define the actual destination of the memory store/recall operation. For example: If memory register 10 contains the value 14, and the display contains "1234", performing "=", "K()", "K()", "1", "0" will store the value 1234 in memory location 14.

External Device Interface Connector and Circuitry

The 909 can interface to external devices via an (apparently optional) external device interface. The interface plugs into the backplane board of the machine, and provides a connector on the rear panel of the calculator for plugging in external devices. Cintra marketed a fairly wide range of devices for plugging into their calculators, including the $995 (later reduced to $700) Model 941 Printer, and an interesting device called the "928 Instructor", that allows keypresses on the keyboard to be recorded in realtime on an audio cassette recorder, then played back (at the same speed) from the cassette. Not truly programming, but allows the unattended operation of the machine to be automated for training and educational purposes. The "REMOTE" key on the keyboard is used for sending and receiving data to/from external devices. The "REMOTE" key accepts a single digit keypress as an argument identifying the external address of the device to be written to/read from, indicating that up two ten external devices can be connected to the calculator in a daisy-chain fashion, similar to the IEEE-488 (GPIB) interfaces available on later calculators and computers.

The 909 also offers 'learn mode' programming, with programs up to 85 steps long in the base model, with a memory expansion option that bumps the program step storage to 256 steps. A rather striking limitation of the machine is that programs are strictly linear, with no notion of branching, looping, or conditionals. The "DEFINE f(x)" key, when activated, puts the machine in 'record' mode. Two indicators in the "DEFINE f(x)" keycap light up when the machine is in learn mode. From that point, all keypresses are stored in program memory in sequential order. When program entry is complete, the "DEFINE f(x)" key is pressed again to take the calculator out of learn mode. Once a program sequence has been defined, pressing the "f(x)" key will cause the programmed sequence of keypresses to be played back at high speed. There are no provisions for editing or listing out the recorded sequences of key presses. The only way to change the function defined in learn mode is to go back into learn mode and start over. There is also no real way to debug a defined function other than repeatedly running the program and hand-checking the results.

The Cintra/Tek 926 "Programmer"

For users with more complex programming requirements than offered by the built-in learn-mode programming functions of the 909, an add-on device called the 926 Programmer augmented the programmability of the 909. This somewhat clunky external device looks like a shrunken down version of the 909 calculator. It has a keyboard with 20 keys and a magnetic cassette drive for storage up to 10 512-step programs. The 926 contains MOS memory used for storing a single program of up to 512 steps. The Programmer adds looping and conditional branching functions, along with program editing and 'single step' functions, and storing/recalling programs from the mag-tape cartridge.

The Cintra/Tektronix 909 is a fast calculator, as fast as Hewlett Packard's competing calculators of the time, and in some cases even a little faster. The processing delay between the time the user invokes a math function and when the machine delivers answers is at worst barely noticeable to a human observer. Published performance claims by Tektronix for the 909 indicate that the machine performs addition/subtration in a blazing 600 microseconds, multiplication and division in 3 milliseconds, and trig functions in 100 milliseconds. In contrast, the specifications for the HP9100A indicate that sine calculations take about 250 milliseconds, making the 909 about 2 1/2 times faster in trigonometric operations.

This particular machine appears to have been manufactured in the late 1971 timeframe, based on date codes on the Integrated Circuit devices in the machine. Some of the circuit boards have Tektronix etched into them, while others have Cintra. The Cintra circuit boards are made of a different material than the Tektronix boards, so it's clear that once the stock of circuit boards that Cintra had manufactured were used up, production of the boards for the machines moved to Tek's circuit board fabrication facilities.


Profound thanks to Mr. Irwin Wunderman for invaluable information on Cintra's history and calculators.

This page dedicated to the memory of Dr. Irwin Wunderman, founder and President of Cintra, and David Takagishi, an electronics engineer at Cintra, and later Tektronix, involved with the design of the Cintra 909 and 911 calculators.

Special thanks to Mr. Robert Krten for providing a copy of the manual for the Tektronix 909 calculator to the museum.
Thank you to Mr. Don Wood for great information on the role of Physics International, and the production and service problems with Cintra's early calculators.
Thanks to Mr. Gary Laroff for loan of old archive material on early Tektronix calculator products.

Text and images Copyright ©1997-2011, Rick Bensene.