This article is in Bob Pease on Analog Volume 2 in the Analog section of the Electronic Design Library.
Members can download the PDF ebook.
Here’s another esaeP’s Fable. The class of 1966 was starting to plan its 25th Reunion. The Reunion Committee went around and contacted all the alumni, until they came to Joe. Joe, as duly noted in the Yearbook, was the Person Least Likely to Succeed in business. That had been a clear choice, back in 1966—everybody recognized that Joe was a klutz, with no sense of proportion, no head for math nor business. But Joe had filled in his questionnaire: President and CEO of Widget Enterprises—a multi-billion dollar multi-national corporation. And Joe had just donated a new Library to the Business School. How could this be? So it was with great respect and curiosity that the Reunion Committee invited Joe to give the key-note address at the Reunion.
It was the same old Joe who stood up to give the speech at the reunion. “I never was much of a speech maker. And I don’t have any big secrets about how I do business. I just buy widgets for $1.00, and I sell them for $10.00. I’m perfectly happy to take just a 10% profit. End of Fable.
When I started work at George A. Philbrick Researches in 1960, I observed a secret project going on—a “skunk-works” project to bring out Philbrick’s first solid-state operational amplifier. Technicians were testing and grading diodes and transistors, night and day, to generate matched pairs. The data sheet was being rushed to completion. Test engineers were learning how to measure currents in the picoampere range. And Sales hopes to sell a few of these P2 amplifiers, at a selling price of $185, to pay for all this research effort.
Wow. An op amp with just 100 picoamperes input current—with no tubes, no heater power, no mechanical choppers. That must use the finest new silicon transistors. No wonder it sells for $185. But when I got to know the senior engineer, Bob Malter, a little better, he showed me that there were not any silicon transistors in the P2. There were just 7 little germanium transistors in there. What? WHAT??
When Bob Malter arrived at Philbrick Researches in Boston in 1957, he was already a smart and accomplished engineer. After designing several analog computer modules (which were the flagships of the Philbrick product line) he became intrigued with the concept of the Varactor amplifier, just about the time that George Philbrick, the founder and chief Research Engineer, was getting frustrated.
George had been trying to make a parametric amplifier, using varactor diodes and germanium transistor amplifiers. When the bridge started out balanced, just a few millivolts of dc input could cause enough imbalance to be amplified and then rectified (synchronously) to drive a dc amplifier. In theory, you could make an operational amplifier that way. But George had worked for many months on an elegant design he called the P7. It used 14 germanium transistors, in a little cordwood assembly with 8 little pc boards packed in between 2 mother boards. He could not get good repeatable results, not for dc accuracy or dynamics or temperature shift.
Now, Bob Malter was a very pragmatic, hard-headed engineer. You would not want to bet him that he could not do something, because he would determinedly go out and do it, and prove that he was right—and that you were wrong. Bob had his own ideas about how to simplify the P7, down to a level that would be practical—which he called the P2. I do not know how many false starts and wild experiments Bob made on the P2, but when I arrived at Philbrick as a green kid engineer in 1960, Bob was just getting the P2 into production.
Instead of George’s 10 pc boards, Bob had put his circuits all on just two pc boards that lay back-to-back. Instead of 14 transistors, he had a basic circuit of 7 transistors—just one more device than the little 6-transistor AM radios of the day. He actually had 2 little transformers—one to do the coupling from the 5-MHz oscillator down into the bridge, and one to couple out of the balanced bridge into the first of four RF amplifier stages. (If you are really interested in the complete schematics of the P2 and P7, and other technical comments and details you will want to buy Jim Williams’ book.*) Note, 25 years ago, these would have been the center of fantastic technical espionage; but today, it’s just a matter of historical curiosity—industrial archaeology—on an obsolete product. You can’t buy the parts to make these amplifiers any more, and even when you could, you could build a circuit to follow the schematic, and it wouldn’t work.
So what’s the big deal? Here’s a pretty crude operational amplifier with a voltage gain of 10,000, and an output of ±1 mA at ±10 volts, with a vicious slew rate of 0.03 volts per microsecond. Who would buy an amplifier like that??? It turned out that thousands and THOUSANDS of people bought this amplifier, because the input basis current at either input was just a few picoamperes. Picoamperes? What the heck is a picoampere? Most electronical engineers in 1960 didn’t even know what a picofarad was, not to mention a picoampere, but, they figured out it was a heck of a small fraction of a microampere. And for many high-impedance instrumentation applications, the P2 was clearly the only amplifier you could buy that would do the job. And it had this low bias current, only a few picoamperes, because all those germanium transistors were running at 5 Mcps, and their 5 or 10 µA of dc base current had no effect on the precision of the input current.
The input current was low, thanks to a well-matched bridge of four V47 varicaps. These were sold by Pacific Semiconductor Inc. (PSI) for use as varactors in parametric amplifiers, up in the hundred of “Megacycles.” The “V47” designation meant that they had a nominal capacitance of 47 pF at 4-V reverse basis, which is where most RF engineers would bias them. But Bob Malter biased them right around 0 Vdc, with a miniscule ±60 mV of ac drive. At this bias, the capacitance was 110 pF plus 1 pF per 20 mV—not an extremely high gain slope.
At this level of drive, each diode would only leak 20 or 40 pA. But Bob had a gang of technicians working day and night to match up the forward conduction characteristics and the reverse capacitance voltage coefficients, and he was able to make sets of 4 varactors that would cancel out their off-set drift versus temperature, and also their reverse leakage. Of course, there was plenty of experimenting and hacking around, but eventually a lot of things worked OK. After all, when you buy 10,000 V47s, some of them have to match pretty well.
So here’s a little do-hickey, a little circuit made up of just about as much parts as a cheap $12 transistor radio, but there was quite a lot of demand for this kind of precision. How much demand? Would you believe $227 of demand? Yes! The P2 originally started out selling for $185, but when the supply/demand situation heated up, it was obvious that at $185, the P2 was underpriced. So the price was pushed to $227, to ensure that the people who got them were people who really wanted and needed them.
Meanwhile, what other kinds of “transistorized” op amps could you buy? Well, by 1963, for $70 to $100, you could buy a 6- or 8-transistor amplifier, with Ibias in the ballpark of 60,000 to 150,000 pA, and a common-mode range of 11 V. The P2 had a quiet stable input current guaranteed less than 100 pA (5 or 10 pA typical), and a common-mode range of ±200 V. (After all, with transformer coupling, the actual dec level at the balanced bridge could be at any dc level so there was no reason the CMRR could not be infinite.)
Wow! A $227 gouge. (You couldn’t call it a “rip-off” because the phrase hadn’t been invented, but perhaps that is the only reason….) Obviously, this must be a very profitable circuit. Every competitor—and many customers—realized that the P2 must cost a rather small amount to build, even allowing for a few hours of work for some special grading and matching and testing. So, some people would invest their $227 and buy a P2 and take it home and pull it apart and try to figure out how it worked. The story I heard was that one of our competitors hired a bright engineer and handed him a P2 and told him, “Figure out how they do this. Figure out how we can do it, too.” In a few days he had dismantled the circuit and drawn up the schematic. Then he analyzed it, and began experiments to be able to meet or exceed the P2’s performance. But he couldn’t get it to work well. He tried every approach, but he never could. After a full year, they gave up.
You see, it turns out there was some interaction between the input of the first RF amplifier and the output of the 4th amplifier, that made the P2 work, when you assembled the two pc boards close together. It would not work with any other layout, orientation, or circuit-assembly technique. So none of our competitors ever second-sourced the P2. And the P2 and P2A and SP2A remained profitable and popular even when the new FET-input amplifiers came along at much lower prices. It was years later before these costly and complex parametric amplifiers were truly and finally obsolete by the inexpensive monolithic Bifet amplifiers from National Semiconductor and other IC makers. Even then, the FET amplifiers could not compete when your instrument called for an op amp with a common-mode range of 50 or 200 V.
Still, it is an amazing piece of history, that the old P2 amplifier did so many things right. It manufactured its gain out of thin air, when just throwing more transistors at it would probably have done more harm than good. And it had low noise, and extremely good input current errors—traits that made it a lot of friends. The profits from the P2 were big enough to buy Philbrick a whole new building down in Dedham, Massachusetts, where Teledyne Philbrick is located to this day (notwithstanding a recent name change to Teledyne Components). And the men of Philbrick continued to sell those high-priced operational amplifiers, and popularized the whole concept of the op amp.
Then when good low-cost amplifiers like the UA741 and LM301A came along, they were readily accepted by most engineers. Their popularity swept right along the path that had been paved by those expensive amplifiers from Philbrick. If George Philbrick and Bob Malter and Dan Sheingold and Henry Paynter and Bruce Seddon hadn’t written all those applications notes and all those books and stores, heck, Bob Widlar might not have been able to give his UA709s and LM301a away! And the P2—the little junk box made up virtually of parts left over from making cheap transistor radios—that was the profit engine that enabled and drove and powered the whole operational-amplifier industry.
One time, I was standing around in front of the Philbrick booth at the big IEEE show in New York City. A couple engineers were hiking past the booth, and one said to the other, nodding his head towards the booth, “…and there’s the company that makes a big bloody profit….” Well, at that time George A. Philbrick Researches was indeed making big profits from the P2. We could never deny that. Just like Joe and his widgets.
All for now. / Comments invited! / RAP / Robert A. Pease / Engineer
