feature article
Subscribe Now

Altera: The Once and Future FPGA Supplier, Part 1

Well, you made it to February. Congratulations! This is the leap month of a leap year. If you’re following the news, Kathy Ireland calendars from 1996 are hot eBay items because the dates and days of the week are the same as for 2024, so you can use ‘em again this year. Meanwhile, in more prosaic news, the FPGA supplier formerly known as Intel PSG will be announcing its new name and mission on February 29 – the leap day of a leap month of a leap year – because, you know, it’s a leap forward for the company previously known as Altera. So, I thought this would be a good time to document the early history of Altera from its beginnings, because no one seems to have done that before and because there seems to be a lot of inaccurate Altera history floating around on the Web.

Contrary to what you may have read in abbreviated Altera histories, the company was started by four engineers: Robert “Bob” Hartmann, James Sansbury, Paul Newhagen, and Michael Magranet. Rodney Smith, Altera’s original CEO, is often listed as a company founder, but he was recruited from Fairchild Semiconductor and joined Altera a few months after the company was founded in 1983. Smith’s substantial contributions to Altera’s success are not lessened by that timing.

After getting his BSEE degree from the University of Minnesota in 1965, Hartmann interviewed with and joined Standard Oil of California. He then moved to the oil patch surrounding Bakersfield where he worked as a roustabout for the first six months. That was a typical assignment for the company’s new college graduates. Hartman’s responsibilities then progressed to specifying electrical power delivery systems for the donkey engines that pumped oil from wells in the ground. Eventually, Hartmann realized two things about his career trajectory. First, Bakersfield’s summers were unbearably hot. Second, power distribution systems based on poles and insulators felt like a tremendous waste of his BSEE.

Hartmann left Standard Oil to join the Autonetics division of North American Rockwell, located in Anaheim, California. At Autonetics, Hartmann learned how to design MOS ICs, which were cutting-edge at the time. Autonetics was Rockwell’s electronics division and had its own fab. Hartmann had design responsibility for several ICs, including the electrical design, layout, mask design, design-rule checking, and testing. One of his projects at Autonetics was a masked ROM. This experience firmly set him on a career in semiconductors. Hartmann left Autonetics after three years and joined Electronic Arrays in Mountain View, California, where he worked for two years. Electronic Arrays was a spinout from General Instrument Microelectronics and made MOS shift registers, ROMs, and calculator chips. Like Autonetics, Electronic Arrays had its own semiconductor fab, which was the norm back then, because it was relatively inexpensive to set up a fab in the early days of ICs.

In 1972, Hartmann jumped to Fairchild Semiconductor, which was located on Middlefield Road in Mountain View, just across the street from Electronic Arrays. At Fairchild, Hartman worked on the 8-bit F8 microprocessor chip set, which was Fairchild’s first venture into the microprocessor business. The F8 project started as a PMOS design but changed over to NMOS by the time the design went to layout. Hartmann later ran an MOS design group at Fairchild.

After three years, Hartmann sensed that Fairchild wasn’t making a sufficient splash in the microprocessor market. When an opportunity arose to design microprocessors and peripheral ICs at Signetics, he jumped once again. Hartmann became familiar with the company’s field-programmable programmable logic arrays (FPLAs) when he used them to develop breadboards for some of his IC designs, These FPLAs were fuse-programmed, bipolar PLDs, and they inspired John Birkner at MMI to develop the first really successful PLDs, which he named PALs. Gate arrays were just starting to take off at that time, and Signetics had a fledgling bipolar gate array business, which is something else Hartmann learned about while working there.

However, after five years of working at Signetics, Hartmann decided that he was tired of working for big semiconductor companies. His duties had drifted to management rather than hands-on work, and he wasn’t happy about that, so he resigned and started a consulting company at home in his back bedroom. He started helping companies develop gate array designs and teaching companies how to use Signetics gate arrays. After several months, Hartman enlarged his consultancy by adding staff and naming the firm Source III. Paul Newhagen and Michael Magranet, both from Fairchild, joined Hartmann at Source III. The Source III business awakened Hartmann to the possibility of creating a larger, vastly improved PLD that could siphon off some of the existing gate array business and expand the market further.

Hartmann and Rob Walker, a former Fairchild and Intel engineer, co-authored a gate-array survey article for Electronic Design magazine, published in 1981. Source III subsequently enlarged the article into a survey report titled “Gate Arrays: Implementing LSI Technology,” which was distributed by Electronic Trend Publications. With more than 50 gate array vendors in the report, Hartmann became convinced that the industry had no need for yet another gate array supplier. The three partners at Source III decided that what the industry really needed was a user-programmable gate array, and they initiated plans to found a semiconductor company that specialized in PLDs. Meanwhile, Walker went off to help found LSI Logic, which specialized in gate arrays.

Paul Newhagen did not know what he wanted to do after graduating with a general engineering degree in 1973 from the University of Illinois at Champaign-Urbana. He interviewed with Alcoa and Dupont, but the jobs these companies offered did not excite him. Then he talked with Fairchild Semiconductor, took an interview trip to Mountain View, and got a job as a production manager in the MOS division. He was responsible for expediting custom metal programmable Micromosaic MOS ASICs for customers. These custom parts had been developed by Fairchild and were the earliest gate arrays.

From this vantage point, Newhagen quickly assimilated all the critical elements of gate array manufacturing, from taking the design to layout, then mask making, fabrication, wafer probing, device packaging, and testing. He found that he enjoyed meeting and working with the entire production team, and within two years, Newhagen had become a production supervisor. However, Fairchild’s MOS group was never the star player in the company’s semiconductor product line. Fairchild was always a bipolar house. Worse, Fairchild’s MOS IC yields were dismal at the time.

Newhagen started to lean towards business management and put together a game plan to gain the skills he thought he’d need. He started studying for an MBA at the University of Santa Clara. He left Fairchild and found a job at Raychem involving production operations, but Raychem was basically a wire and cable company that also made heat-shrink tubing. That was the company’s high-tech claim to fame. Newhagen found that Raychem was not to his taste. Fairchild’s high-tech semiconductor manufacturing processes had spoiled him, and Raychem’s manufacturing processes seemed to be from the stone age by comparison. Newhagen resigned from Raychem and interviewed several semiconductor companies, wanting to get back into the business.

Ultimately, Newhagen went back to Fairchild and landed in the marketing department where he developed sales forecasts and expedited customer orders. It was a perfect spot to meet even more people at Fairchild, including the executive management team, field sales, and the division controllers. Shortly after he earned his MBA degree, Schlumberger acquired Fairchild Semiconductor, and Newhagen moved into a controller position under Rodney Smith, who was the General Manager for the Automotive Hybrid Division, which was one of Fairchild Semiconductor’s smallest product divisions. Nevertheless, Newhagen found that he enjoyed being a financial controller. “You had some clout,” he explained.

However, Newhagen’s clock was running and he was already planning to leave Fairchild because he didn’t like Schlumberger’s corporate management style. He resigned from Fairchild to hike the John Muir Trail in 1981, but before he could leave for the hike, Newhagen informally met with his friend Bob Hartmann around Christmastime. Hartmann wanted to talk about gate arrays and said he wanted to start a consulting business to do something in gate arrays. He wanted to help customers develop gate array designs for ASICs offered by existing semiconductor vendors. It was new and different, so Newhagen said, “Yes.”

Newhagen’s “yes” triggered a chain of events that would produce one of the two major programmable logic vendors in the semiconductor industry – Altera. I’ll cover Altera’s initial products and product development history in Part 2 of this article. Look for it soon.

Acknowledgement: My thanks to Robert Bielby, who spent a few years at Altera and has an encyclopedic FPGA and CPLD memory.

References

Interview with Robert Hartman, 2018 September 11, Silicon Genesis: oral history interviews of Silicon Valley scientists, 1995-2022, Stanford University

Interview with Rob Walker, 1998 July 09, Silicon Genesis: oral history interviews of Silicon Valley scientists, 1995-2022, Stanford University

Interview with Paul Newhagen, 2019 November 18, Silicon Genesis: oral history interviews of Silicon Valley scientists, 1995-2022, Stanford University

Altera EP300 Design & Development Oral History Panel, August 20, 2009, Computer History Museum

7 thoughts on “Altera: The Once and Future FPGA Supplier, Part 1”

  1. “Kathy Ireland calendars from 1996 are hot eBay items because the dates and days of the week are the same as for 2024”

    Isn’t that always the way. You hang onto something for years. Then you have a clean out. and the next day you think, “I wish I’d kept my Kathy Ireland calendar from 1996!”

    Meanwhile, my hat is off to the marketing folks at ??? (formerly known as Intel PSG, formerly known as Altera) for even thinking to announcing its new name and mission on February 29 to signify their great leap forward.

    1. I’m reminded of when Prince changed his name to a symbol and newspapers and commentators where obliged to say “the artist formerly known as Prince.” They got their own back when he reverted to using Prince as his name, and they referred to him as “the artist formerly known as a symbol” (honestly, I couldn’t make this stuff up).

  2. Steven, when you get to part II I hope you mention that their early EP300 part didn’t work right. It had issues with power supply rise times. If the supply came up to slow the part was brain dead. Hence the EP310…

    Anymore I try to avoid bleeding edge parts. I’ve lost to much blood.

  3. I am aware that the original EP300 masks had a problem with power supply rise time, bpaddock. It’s in the oral history panel record at the Computer History Museum. I did not cover it in part 2. If I cataloged every part that didn’t work right, just from my personal experience, I’d have to write a book. Today, that info is only important to someone using the part today. Hello? Hello? No one’s there.

  4. Back before Intel acquired Altera, Altera was the first customer for a new process node.
    Intel had to resort to the multicore farce because they could not get a significant gain in single core performance. (yes, I said “farce” because no one knew how to program multi-core to get significant performance improvement) There has been enough intellectual market appeal so they have survived.

    Truth is they must have found that the FPGA market cannot produce enough revenue to support the Intel life style.

    What will be the next move for Intel? Quantum?

    What is the future for desktop computers?

    Well, the untapped potential of using the hundreds of embedded memory blocks on FPGA chips is there for the picking for the “new” Altera.

Leave a Reply

featured blogs
May 8, 2024
Learn how artificial intelligence of things (AIoT) applications at the edge rely on TSMC's N12e manufacturing processes and specialized semiconductor IP.The post How Synopsys IP and TSMC’s N12e Process are Driving AIoT appeared first on Chip Design....
May 2, 2024
I'm envisioning what one of these pieces would look like on the wall of my office. It would look awesome!...

featured video

Introducing Altera® Agilex 5 FPGAs and SoCs

Sponsored by Intel

Learn about the Altera Agilex 5 FPGA Family for tomorrow’s edge intelligent applications.

To learn more about Agilex 5 visit: Agilex™ 5 FPGA and SoC FPGA Product Overview

featured paper

Altera® FPGAs and SoCs with FPGA AI Suite and OpenVINO™ Toolkit Drive Embedded/Edge AI/Machine Learning Applications

Sponsored by Intel

Describes the emerging use cases of FPGA-based AI inference in edge and custom AI applications, and software and hardware solutions for edge FPGA AI.

Click here to read more

featured chalk talk

E-Mobility - Charging Stations & Wallboxes AC or DC Charging?
In this episode of Chalk Talk, Amelia Dalton and Andreas Nadler from Würth Elektronik investigate e-mobility charging stations and wallboxes. We take a closer look at the benefits, components, and functions of AC and DC wallboxes and charging stations. They also examine the role that DC link capacitors play in power conversion and how Würth Elektronik can help you create your next AC and DC wallbox or charging station design.
Jul 12, 2023
35,246 views