My most influential ICs -

My most influential ICs

At the turn of the 21st century several EE luminaries created lists for a Hall of Fame of ICs. I searched, but could find no official list. I was reminded of it in a comment by Martin RoweMAX232 is a hall-of-fame IC. I designed them into a number of products, including the first Extech Comms printer ”. I thought I should create my own Hall of Fame. There are some that I am sure we can all agree should be there. Others have special meaning to me in that they proved to be really useful and appeared in several projects, often defining the approach.

There can be no doubt that the 555 should be on anyone’s list. An analog IC that pretends to be digital, can be configured as an astable or monostable and is simple enough for a beginner or even a dumb digital engineer like me. The market segment still seems to be remarkably vibrant (see my blog “Timer, anyone?”)

Early on, MOS micros, EPROMs, and RAMs had multi voltage supplies and they drew a fair amount of current, so heat dissipation from linear supplies (switch mode power supplies were few and far between) was an issue. Micros and RAM had already shifted to a 5V only supply, but when Intel introduced the 2716 (2Kx8) EPROM systems could be completely single supply. What a win! And yet we have now returned to multiple supplies!

The Hitachi range of dot matrix LCD character displays using the HD44780 controller came in many formats from 1 line by 8 characters to 2 lines by 32 characters. They were great for a user interface when micro-based devices were still in their infancy. Whilst the HD44780 is a candidate on its own for Hall of Fame, I have an associated IC: The LCD module required a negative voltage to alter the viewing angle of the display and since I wanted a just a single supply board, Intersil’s ICL7660 was the answer to this problem and several other issues that required low current negative voltages.

There have been many iterations to the switched capacitor approach of the ICL7660, higher voltage, smaller capacitors, positive voltage, smaller size but perhaps the most famous of the offshoots was the MAX232. Since I began this blog with Martin’s quote, I need not continue.

When I was designing SBX cards (mezzanine cards intended for Intel’s single board computer (Multibus) product range) I designed and produced a watchdog timer which consisted of several TTL ICs. I remember going to the first Maxim seminar (Maxim had just been formed as a spin off from Intersil) and learning about the MAX232 and the MAX690 watchdog timer. I was really enthused after that and almost every system of mine since then has employed a watchdog timer.

As you can see above, single supply voltage drove much of the design considerations. In the early days I was also very wary of analog circuits and had a mental block on the bipolar voltages used on the linear devices. And then suddenly there were the LM311/ LM339 single and quad comparators and there was light. It took a while longer for the single supply op-amp like the LM358 to come out, but when it did I found myself working more and more with analog circuits. The TL431 voltage reference has also graced many of my designs.

Still sticking with the single supply theme, who could forget the linear regulators – the LM309, the 7805, and the adjustable LM317? If I remember correctly, the LM317 was selected as the electronic product of the year when it was released.

When it came to micros I designed with quite a few: the 8085, Z80, 8048 and the 8051, but if I have to select just one then it would have to be the 8051. But I hasten to add that when I discovered the PSoC architecture circa 2002, there has been no other architecture that I will use if given free reign. It is true that the implementation has undergone a radical shift from the PSoC1 to PSoC4 and PSoC5, but it is for the better and I really enjoy building as much of the design into the hardware capabilities of the device. Take a look at my recent 3 part blog on how to measure the RMS of a signal “Measuring an RMS value on a PSoC5” Part 1, Part 2 and Part 3.

I was always interested in clock/calendar devices and it seemed to me that they took a long time to come out in monolithic form. I made and early SBX board using the NEC uPD1990 which was a Julian calendar and needed software to add the leap day. National Semi had announced several, but when I tried to get them I was told by the salesman that it was “easy to invent a data sheet, but harder to make the device”. When Motorola brought out the MC6818 it seemed to encapsulate all that I needed and my choice was validated by IBM’s selection of it for the PC.

Continue to page 2 >>

I had worked with several standalone UARTs and used several offerings like Intel’s 8251, but when National (or maybe it was Western Digital) brought out the INS8250, it seemed to cover every option (including an on-board oscillator and baud rate generator) except one. It didn’t have the 8051’s 9-bit mode. Still it was good enough to be in the PC. I actually wrote a design idea on how to manipulate the 8250 to achieve a 9-bit performance – it was published in EDN Design Ideas as” Implement a Nine-Data-Bit UART on a PC” in June 6, 1998, but good luck finding it. [Editor's note: Although various system migrations have made finding gems such as this difficult, a portion at least of Aubrey's article is here (scroll to the bottom).]

As I started working more and more with RS485, the appearance of the MAX485 was a great improvement over the earlier devices. Although the TI SN75176 was nominally the same functionality as the MAX485, it was not implemented in CMOS and drew quite a bit more current. Perhaps it deserves the nod.

Strangely enough, driving many LEDs (both discrete and as 7 segments) popped up quite often and the micros had a limited number of pins, not to mention fairly slow clock rates so that there were constraints on processing tasks. I found Intersil’s ICM7218 (and later Maxim’s equivalent) family worked its way into several designs especially since it had a serial interface, which was relatively unusual back then. The family is still going strong as you can see from a new Arduino Display Driver Shield from Maxim.

Figure 1. The 72 LEDs just to the left of the middle of the board were driven by an ICM7218, the IC just below them. There is another favourite IC on the board – the 8032 (8051 derivative) and a linear regulator on the left with the heatsink although I can’t remember why. The unit worked from the telecommunications standard -48VDC and used a SMPS module (in the unpopulated area on the left). (Source: Author)  

And for something completely different: years ago I was told that it was not possible to design a self-powered 4-20 current loop isolator. I stumbled across a linear opto-isolator the LOC110 from CP Clare (now IXYS) (an alternate to the IL300 from Siemens at the time) and started working with it in my spare time. In the end it could be done and I am told my product is now considered the industry standard and has been for over 17 years. I first used the Linear Tech op-amp LT1366 on this loop isolator. For a digital designer flirting with analog, I chose the most all-round op-amp that I could find. Low power, rail-to-rail and operating to over 24Vdc supply, it may be pricey, but it worked like a charm and it frequently sneaks into my designs even though there may be a more optimal choice for a specific case.

There are two kinds of design engineers in industrial electronics: those who know about instrumentation amplifiers (IA) and those who are about to learn that they needed an IA and what the heck are they going to do about it now? Planet Analog editor, Steve Taranovich, recently did a blog on the topic “6 Burr-Brown secrets to success you need to know about instrumentation amplifiers”. I messed around with a few in my early days, even tried to make my own from three op-amps. I quickly settled on the AD626 from Analog Devices. It will work with a single supply voltage, from low to high supply voltages and will tolerate a common mode voltage outside the supply rails, making it suitable for high side current monitoring. Even though there are many more options in the marketplace today, it is still a favourite of mine.

I have designed more 4-to-20mA output devices that the proverbial number of hot dinners I have consumed. Where there is no requirement for loop power, the XTR110 is the quickest and surest way of designing the output. It doesn’t hurt that it has great performance as well even though today there are more economical options.

It’s amazing how long some of these devices have been in production. There is no real order in my list, so please don’t try to read anything into the sequence. It’s your turn now. What devices have made a difference to the way you approach a design?

My List with links:

  1. 555

  2. 2716

  3. ICL7660

  4. MAX232

  5. MAX690

  6. LM339

  7. LM358

  8. TL431

  9. 7805

  10. LM317

  11. 8051

  12. PSoC

  13. MC6818

  14. INS8250

  15. MAX485

  16. ICM7218

  17. LOC110

  18. LT1366

  19. AD626

  20. XTR110

8 thoughts on “My most influential ICs

  1. “I'd add the General Instruments AY-3-1015 async serial transceiver, the original UART and the go-to part for RS-232 interfaces.nnAnd how many times did the lowly 2N3904 and 2N3906 transistors make it onto a board to switch that one last signal?”

    Log in to Reply
  2. “@JackPeacocknn”I'd add the General Instruments AY-3-1015 async serial transceiver”nnI guess you'd have to add a Baud Rate generator to that. What part would you use- the Fairchild 4702?nn”lowly 2N3904 and 2N3906 transistors “. nnIt didn't occu

    Log in to Reply
  3. “No mention of the 741? It may be a pretty ordinary op-amp compared to your AD626, but it was the first real universal op-amp (the 709 was finnicky to compensate) and a lot of us cut our teeth on it. That makes it influential in my book…”

    Log in to Reply
  4. “DavidnnThere is no dispute about the influence of the 741- it just wasn't a direct influence on my designs. nnHowever indirectly it did influence me a lot. I did a design once for a coffee packaging machine that measured weight using LVDTs. The gain s

    Log in to Reply
  5. “Mine was the 6502 chip from Synertek in 1980. It was lower priced than other available 8 bit micros at the time and opened up many different applications to micro control. There were also several other manufacturers who were making the part which helped k

    Log in to Reply
  6. “@RGARVIN640nnMy electronic formative years were in South Africa which was far removed from any other epicentre of electronic design and the market was also significantly different. Intel had great representation followed at quite a distance by Motorola.

    Log in to Reply

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.