30 September 2011

Book 1 Chapter 23

Chapter 23: "The zoo circuit: History, mistakes, and some monkeys design a circuit."

This chapter discusses the development of the "zoo circuit", a low-cost low-power voltage-to-frequency converter, which originally appeared in App Note 23. Powered off of a 9-volt battery, the circuit works from 0 to 10 kHz with good linearity, and consumes only 200 microamps.

Quite a few of the figures are drawn from previously published app notes.
  • Figures 23-3 to 23-6 come from App Note 14, Figures B2 to B5.
  • Figure 23-7 is reproduced from the front page of the LT1055/LT1056 data sheet. (Did it also appear in another App Note? I can't seem to find it.)
  • Figures 23-9, 23-11, and 23-13 come from App Note 23, Figures A1, A2, and A3.
  • And, of course, the end result in Figures 23-19 to 23-22 come from App Note 23, Figures 16 to 19 (although there's a small error in Figure 16)

The text of this chapter is a good study in methodical circuit design. (This chapter is as good of a polemic against the promises of "effortless" computer-aided design as yesterday's chapter is.) Jim starts the development of his circuit with a version of Bob Pease's Teledyne-Philbrick 4701, shown in Figure 23-7. In the following development, he carefully investigates the sources of power consumption, nonlinearity, and temperature drift, and attacks them in turn. After some careful iteration, and a flash of insight at the zoo, he arrives at the final circuit in Figure 23-19.

Finally, here's the the eponymous quote (first the first page): "Most of the ideas came from history, making mistakes, and the best source of help was some monkeys at the San Francisco Zoo."


A bibliographic note: this chapter also appears as Chapter 18 in Bob Pease's book, "Analog Circuits: World Class Designs". By the way, "Zoo Circuit" is one of the search terms that seems to bring people to this blog. For those searching for such information, don't miss the updated versions of this circuit, which consume as little as 8 microamps, in App Note 75 (see Appendix A and pages 1 to 4).



Related:

29 September 2011

Book 1 Chapter 13

Chapter 13, "Should Ohm's Law be repealed?"

Another great Jim Williams polemic (like Chapter 4).

This chapter begins with the story of Dr. Stearn, Jim's childhood neighbor who shared his electronics hobby with Jim and who owned a Tektronix 535 oscilloscope ("a bumpless combination of interdisciplinary technology, inspired design, attention to detail, aesthetics, and usability"). There is some great advice here, via Dr. Stearn, such as, "Use any tool that will help move your thinking along, know how these tools work, and keep their limitations in mind when you use them." Clearly, Jim took this advice to heart (as evidenced by his numerous appendices on scopes, probes, and measurement techniques).

"One daughter, my age, had cute freckles and long, chestnut hair. Once, she even baked me chocolate chip cookies and presented them in a blue box with a ribbon. They were good. I can't be sure, but I think I missed a cue. A born engineer."

The polemic starts when Jim starts discussing CAD systems, and the pie-in-the-sky promises thereof. It seems incredible that an (unnamed) magazine editor actually wrote, without irony, that "Today's mainstream designers, whether they're designing a complex board-level product or an IC, don't need to fuss with electronics." Today's mainstream chefs, whether they're cooking a wedding banquet or a meal at home, don't need to fuss with food.

Finally, note Figure 13-2:


That does not look like the good life to me. That looks like Hell.

Best quote: "Mother Nature loves throwing a surprise party. Technologically driven arrogance is a dangerous brew, as any Titanic passenger will assure you."



Related:

28 September 2011

Book 1 Chapter 7

Chapter 7, "Max Wien, Mr. Hewlett, and a rainy Sunday afternoon."

The first time I read this chapter (back in grad school), I was quite intimidated. It would take me two weeks to design, build, test, and document all of the circuits in this discussion, and Jim did it in one Sunday afternoon?!? I'll never be that good of an engineer.

But this chapter title is a bit of a fib, because it didn't all happen on a single rainy Sunday afternoon. He had clearly been thinking about Hewlett's oscillator since App Note 5 (Figures 12, 13, and 14), and most of this chapter is based on circuits that appeared in App Note 43 (Figures 39 through 49).

On the first page, he writes "No lab is complete without an HP series 200 oscillator." We should all be so lucky! However, on page 47, he seems to admit that (at the time he wrote this chapter), he owned an HP 201 oscillator. His HP 200A must have come later.

This chapter is worth reading, even if you have memorized App Note 43. His conversational tone gives a window into his thought processes (even if the "rainy afternoon" is a work of fiction). This chapter is also noteworthy for his clear exposition of Williams's Rule (his spelling), "a little known tenet of precision op amp circuits... always invert (except when you can't)."

It is curious that all of the schematics in this chapter are drawn by hand:


For App Note 43 all of these schematics had been neatly drawn by someone at Linear Technology. Why did Jim redraw everything by hand? Are these the original schematics from the lab notebook before the Art Department got a hold of them? Did he feel challenged by Bob Pease (see Chapter 9) for the crown of "Messiest Schematics in the Book"?

There is a typo on the top of page 44. He writes "DeForest hadn't even dreamed of adding a third element to Edison's Effect in 1891." He misspelled "DeForest hadn't even dreamed of adding a third element to the valves he stole from John Fleming."



Related:

27 September 2011

Book 1 Chapter 4

Chapter 4, "Is analog circuit design dead?"

Is analog circuit design dead? All of these publications, all of these great circuits, and Jim's whole career are evidence to the contrary. Clearly, the answer is "no", and Jim writes a great polemic. (Did James Solomon really say, "all classical analog techniques are dead"? That's terrible.) Twenty years later, we don't hear much from the "analog is dead" choir. The "war" of which Jim speaks here is long over. Few doubt the value and necessity of the analog interfaces. Analog may not be king, but without analog, digital would have nothing to rule over.

On the other hand, I do have to relate that I was recently asked why I was bothering with designing low-noise RF amplifiers: "Can't you just use an analog-to-digital converter?" (Um, no.)

"Do all you bit pushers out there get the message?" Yes, Jim, I think they finally did.

I love the quotes from George Philbrick (from an article that is also reprinted in this book, as Chapter 2), and the shout-out to Korn and Korn and Henry Paynter's "Palimpsest on the Electronic Analog Art". Good reading, all around. Best quote (page 17): "Analog computers did not die out because analog simulation are no loner useful or do not approximate truth; rather, the rise of digital machines made it enticingly easy to use digital fakery to simulate the simulation."

Finally, consider Figure 4-2:


If you ever, ever, ever see a Tektronix 556 in somebody's trash, please call me.



Related:

26 September 2011

Book 1

A brief detour this week: App Note 43 was published in June 1990. The next app note (App Note 45) was written a year later. In the intervening 12 months, Jim had a son, Michael, and published a book, "Analog Circuit Design; Art, Science, and Personalities". So this week, I want to blog about the book. If you don't own a copy, you owe it to yourself to buy one. It's a fun, fun read.


Yes, his workbench really did look like that. Note the presence of this trusty Tektronix 547.

This book contains thirty chapters, written by a wide variety of "personalities", including Barrie Gilbert, Paul Brokaw, Bob Pease, James Roberge, John Addis, and many others. Jim wrote four of the chapters himself. I'll discuss them each in turn. One interesting note at the beginning of the book is in the biographies of the contributors (starting on page xiii). In Jim's bio, it says "He lives in Belmont, California, with... 14 Tektronix oscilloscopes." Such a small collection! (In the following years, he acquired many, many more.)

I have several copies of this book (I hate seeing them in used-book stores; I buy them when I see them). I even got Jim to sign one of them.


You can find some excerpts of the book on Google Books.



Related:

25 September 2011

Scope Sunday 10

Success at the Swapfest! After my complaining last month, I found pretty good pickings at September's MIT Flea Market.

My first acquisition was a Frankenstein oscilloscope. The body of the scope says 465B, but the handle says 475. Some of the knobs are mysteriously covered with rubber caps. I really have no idea what I will find inside when I open it up... perhaps the guts from a completely different scope? Maybe a stripped chassis? Maybe a bobcat?. In truth, it doesn't really matter: it cost me $2.


My second purchase was a Tektronix 575 Transistor-Curve Tracer. I spent a fair bit more than $2 on it (although less than a tenth of a current Craigslist asking price). It's missing a few switch caps, but all the tubes are inside. For those of you keeping track, yes, I already have one. However, I couldn't resist buying this one because it had a Hybrid Systems property sticker on the top and a Sprauge property sticker on the back (and a good friend of mine used to work for Hybrid Systems until they were acquired by Sprauge). Thus, I felt that I had to buy it for the provenance alone.


A much better outing than last month's! (And, yes, I do feel bad for only paying $2 for the Tek 465B. The seller wanted $5, but he didn't have any change, and I didn't have any other small bills. If I see him again, I'll probably give him the other $3!)

24 September 2011

App Note 43 part 3

Just a quick post to finish up App Note 43: This note ends with five appendices, two of them with guest authors. In Appendix A, Jim writes a little about the history and construction of strain gauges, and introduces the first guest authors from Motorola Semiconductor (since spun off into ON Semiconductor and Freescale Semiconductor), who write about semiconductor-based strain gauges, which they spell "gage" (Jim explains the alternative spelling in the first footnote on page AN43-36). I wonder if this foray into semiconductor gauges was part of the original impetus of this app note?

Appendix B discusses historical bridge-readout techniques, starting with nineteenth-century galvanometers, and continuing up through modern chopper-stabilized amplifiers and analog-to-digital converters. Again, he gives unwarranted credit to Lee DeForest, which is disappointing. Note the captions in Figures B1 and B2, "Courtesy the J.M. Williams Collection."

Appendix C (briefly) discusses Bill Hewlett's oscillator design and includes the schematic (Figure C2) from his Stanford thesis.

Appendix D presents some theory behind distortion measurements, and is written by guest author Bruce Hofer. There is some good discussion here. Unfortunately, he starts the appendix by defining "linear distortion" and "nonlinear distortion", the former of which I consider a oxymoron.

Appendix E briefly mentions some cabling and filtering considerations for bridge circuits, and the final page is another of his famous cartoons, this one starring the Golden Gate Bridge.


Best quote (page AN43-42): "The Hewlett-Packard Company and light bulbs have had a long and successful association."



Related:

23 September 2011

App Note 43 part 2

The second major part of this app note discusses low-distortion sine-wave oscillators, inspired by his obsession with the HP200A oscillator.

He has discussed this topic before, of course. In 1981, he wrote National Semiconductor App Note 263, "Sine Wave Generation Techniques", which includes a wide variety of schemes, including the Wien-bridge approach (see NSC App Note 263, Figures 2 and 3). He uses a Wien-bridge oscillator, without much comment, in Figure 10 of LT App Note 3 (compare LT-AN3 Figure 10 with NSC-AN263 Figure 2b). In App Note 5, Figures 12, 13, and 14, he discusses a Wien-bridge oscillator with the light bulb for amplitude control (these figures are reproduced in the present app note). He also references Bill Hewlett's thesis (see App Note 5, Reference 5).

In this app note, the discussion of oscillators starts on page AN43-27, with a simple multivibrator in Figure 32, and a quartz-stabilized version in Figure 33 (quartz-stabilized oscillators are discussed in App Note 12, but not using bridge circuits). A #327 lamp is added in Figure 35 for amplitude control, and a common-mode-suppression loop is added in Figure 37.

The Wien bridge is introduced in Figure 39 (compare to Figure 2a in NSC App Note 263). Figures 40, 41, and 42 and copied from App Note 5, Figures 12, 13, and 14, using multiple lamps to increase the thermal time constant and reduce the distortion at low frequencies. This approach, as explained in Appendix C, is the basis of the HP200A oscillator.

Figure 43 begins the hunt to replace the light bulb. Using a FET as a simple voltage-controlled resistor increases the distortion by fifty times. As he says in the footnote, "What else should be expected when trying to replace a single light bulb with a bunch of electronic components? I can hear Figure 39's #327 lamp laughing." Figure 45 corrects the error (due to channel-length modulation), and Figure 47 replaces the FET with an optically driven CdS cell. Figure 48 further improves the distortion with a common-mode-suppression loop (a similar improvement was seen between Figures 35 and 37). The distortion is now below 3 ppm, below the uncertainty floor of his distortion analyzer. (This section is the basis of one of his book chapters, so we'll see this material again, soon.)

The app note concludes with a synchronous-rectifier AC-to-DC converter in Figure 50 and 51, which are copied from App Note 13, Figure 36 and 37.

Best quote (from page AN43-29): "History records that Hewlett and his friend David Packard made a number of these type oscillators. Then they built some other kinds of instruments."



Related:

21 September 2011

App Note 43 part 1

"Bridge circuits: Marrying gain and balance." 48 pages.

At 48 pages, this is the longest app note so far (however, the epic App Note 47 is just around the corner). This app note has three major sections: instrumentation bridge circuits, Wien-bridge oscillators, and the (five) appendices. This app note is important for two reasons: one, it is a succinct and useful collection of signal conditioning circuits for bridge circuits, and two, it includes an in-depth discussion of low-distortion sine-wave oscillators, inspired by his obsession with the HP200A oscillator.

I'll cover each section in turn.

The first section (just about half of the app note) discusses instrumentation bridges, including pressure transducers, strain gauges, and temperature sensors (mostly using RTDs). The table in Figure 4 is a good summary of various interface circuits for bridge circuits and sensors. The schematics in the following figures (Figures 5, 6, 7, 8, and on), show specific implementations of the circuits in Figure 4. The various common-mode suppression tricks are very useful. The single-supply circuits in Figures 9 and 10 use the LTC1044 and LT1054 voltage inverters, which have been previously discussed (and Figure 9 is copied from App Note 11 Figure 7).

Figure 11 shows a high-precision scale that is sensitive enough to measure weight fluctuations due to a person's heart beat. See the "ballistocardiograph" in Figure 12. I've found this example makes a great undergraduate laboratory project, because it's really exciting to see it working, and students love playing with it.

Several of the following circuits have been discussed in previous app notes. Figure 13 shows an instrumentation amplifier with a 200V common-mode range using HP HSSR-8200 opto-switches (App Note 6 Figure 2 has a 300V common-mode range, but perhaps those opto-switches are harder to get).

Linearized platinum RTD bridge circuits are shown starting with Figure 14. The switched-capacitor signal conditioner in Figure 15 is copied from App Note 3 Figure 6. Figure 16 uses digital linearization, which is the reason for the three-and-a-half pages of 68HC05 code (from page AN43-14 to AN43-17).

Power reduction is explored starting with Figure 19, which uses low-voltage drive. Figure 20 uses strobed bridge power (and is copied from App Note 23 Figure 5). Figure 21 uses a sampled output, and Figure 23 adds a sample-and-hold circuit for continuous output voltage (Figures 23 and 24 are very similar to App Note 23, Figures 3 and 4.) Figure 25 claims higher resolution with 50-volt drive.

The lock-in amplifier in Figures 27 and 28 is copied from App Note 3 Figures 4 and 5, and the level transducer circuits in Figures 29A, 29B, and 30 are copied from App Note 7, Figures 13, 14, and 15.

The next section of the app note starts discussing oscillators, which I'll cover on Friday.

Best quote (from the footnote on page AN43-1): "Wheatstone had a better public relations agency, namely himself."



Related:

19 September 2011

App Note 37

"Fast charge circuits for NiCad batteries." 4 pages.

At four pages, this app note is the shortest one ever. It discusses using temperature measurement to determine the charging status of a NiCad battery pack (which was previously discussed in Figure 3 in App Note 6). This approach is still a good idea: I've been witness to some spectacular battery-pack failures, and I fully appreciate the need to prevent them!

In this app note, two thermocouples are used to compare the battery-pack temperature to ambient temperature and to taper the charging current. Figures 2 and 4 use an LT1006 precision single-supply op amp to measure the temperature and drive a 10-amp Darlington transistor to charge the batteries. (Appendix A discusses the construction of the necessary low-resistance shunt to measure the transistor current in the charging circuit.) An op-amp offset-trimming resistor is included in the schematic because the type-K thermocouples only produce 40 microvolts per degree Celsius.

Figures 5 and 6 use switch-mode drive to reduce the power dissipation in the charging transistor. I'm not a battery expert, but I've been told that this approach is a bad idea. Rechargable batteries really prefer constant charge (and discharge) rates, and charging (or discharging) with pulses of current will shorten the battery pack's lifetime. Am I misinformed?

Best quote (on page AN37-1): "Excessive internal heating degrades the battery and can cause gas venting to the outside atmosphere." Yep.

18 September 2011

Scope Sunday 9

The big oscilloscope-related news this week is that the vintageTEK.org Museum, "A Museum of Vintage Tektronix Equipment" held their grand opening this weekend (on Friday, September 16th).


Unfortunately, I wasn't able to attend the grand opening; I only learned about it earlier this week. Actually, Jim Williams originally told me that this museum was in the works two years ago, (and we had discussed attending the grand opening together), but I wasn't aware that they had made enough progress to open the museum to visitors.

From the press release:

Our potential inventory consists of more than 1,500 instruments donated by Stan Griffiths, cofounder, and approximately 375 donated instruments by ex-Tektronix employees, along with a large supply of parts, accessories, and manuals. Technical volunteers have been working since 2010, repairing and refurbishing instruments for display. Our volunteers consist primarily of ex-Tektronix design engineers, technicians, and marketing personnel, but also include vintage Tektronix equipment aficionados that were customers, who are refurbishing classic instruments as well.


I will be planning a trip to Portland in the very near future. Who's with me?

16 September 2011

App Note 35 part 3

Some of the appendices of App Note 35 are fun to read. Appendix A just discusses the internals of the LT1074, but Appendix B discusses the external circuitry. Inductor selection is discussed again (similar to Appendix D in App Note 29), but this time with an added bonus: the "Alternate Method". While the purchase of a induction selection kit (Figure B2) is still the preferred method, inductor selection can be accomplished with a quick trip to Safeway (shown in Figure B8, that's a grocery store, for those of you not on the west coast of the US). Additional humor (or is it teasing?) is found in Figure B9 and Footnote 13. The general idea is to find inductors with significant flux-storage capability and low resistance, but the presentation is classic Williams style.

Appendix C discusses instrumentation for current measurement. He discusses Hall-effect probes (like the Tektronix P6042) and transformer probes (like the Tektronix 131 and 134). He also discusses using current shunts and differential amplifiers (like Figure C3) or differential scope plugins (like the Tektronix 1A5, 1A7, 7A13, and 7A22). Quote: "Ohm explains that measuring voltage across a resistor gives current." (Is that some frustration that I sense in his voice?)

Appendix D discusses the optimization of efficiency, and is similar to Appendix E in App Note 29 (although Figure D2 and the accompanying discussion are new).

Appendix E discusses a reference half-sine generator that is used for the 115-volt 400-hertz sine-wave-output converter in Figure 31. Unfortunately the approach here is a digital technique, with samples stored in a 2716 EPROM, rather than an analog approach. Reading between the lines, I think you can tell that using a digital implementation here bothered him; you know how much he loved his sine-wave oscillators! (App Note 43 is coming up, soon. He'll get his fill.) I think quoting Philbrick in the footnote (see below) assuaged his guilt.

Appendix F discusses "the magnetics issue" again, just like Appendix G in App Note 29.

The app note concludes with two cartoons! "Kick all that creaky discrete stuff outta here" on Page AN35-31 and the famous "Call me" on page AN35-32. We'll see a reference to "Call me" again, soon.



Best quote (footnote from page 28): The sinewave is probably the paramount expression of the analog world. The Old Man Himself, George A. Philbrick, once elegantly discussed analog functions as "those which are continuous in excursion and time." He might have viewed digital production of a sinewave with considerable suspicion, or simply labeled it blasphemous. Such are the wages of progress.



Related:

14 September 2011

App Note 35 part 2

The circuits in this app note are applications of the LT1074 step-down regulator chip. Several of these applications are variations on themes that we've seen in previous app notes. Given the number of notes that Jim wrote on the topic of switching power conversion (App Notes 25, 29, 32, and now 35), these efforts were a clear priority of the "Captains of the corporation" (to steal a phrase from the introduction of App Note 25).

Most of the circuits here are straight-forward applications. Figure 3 shows the basic step-down topology, while Figure 8 adds a coupled inductor to provide a negative output voltage. Figures 9 and 11 show negative output regulators. Figure 12 provides current boosting with a tapped inductor. Figures 14 and 15 show switching converters with linear regulators (similar to the circuits in App Note 32). Figures 16 and 18 shows micropower converters (achieving quiescent currents of 150 uA with the shutdown pin) that are similar in technique to Figures 12 and 19A in App Note 29.

The best circuits in this app note are the high-voltage applications in Figures 26 and 31. Figure 26 provides a 100-watt variable output up to 500 volts. (Note that the LT1074 in this circuit controls the input voltage to the high-voltage transformer. The LT1074 does not (directly) provide the HV output.) The 500-volt step response into a 100-watt load shown in Figure 30 is an impressive achievement. (The footnote on page AN35-13 references App Notes 18 and 6 for this design.)

Figure 31 shows a sine-wave-output converter. The circuit takes 28 volts in and outputs a 400-hertz 115-volt AC waveform. (Wait... 28 volts and 400 hertz? Smells like military.) Again, the LT1074 is pressed into service on the input side on the HV transformer. The switching drive to the transformer is provided by the "timing and reference half-sine generator" and the 74C74 flip flop. Figures 33a and 33b show the details of the zero crossing, where the "half-sine" reference waveform gets "de-rectified" into a full size wave. Pretty. Figure 34 shows the purity of the final sine-wave output.

I'll discuss the appendices on Friday.



Related:

12 September 2011

App Note 35 part 1

"Step down switching regulators." 32 pages.

This app note is another great work that will take several days to cover. Again, I want to start by talking about some of the instrumentation. One measurement that is worthy of note is Figure 30, showing a 500-volt square wave! As the caption says,"DANGER! Lethal potentials present..."

However, the best measurement is clearly Figure 34, that shows a 115-volt sine wave, its distortion products, and its frequency spectrum all in one shot.


Jim teases us, but gives away no secrets, in the footnote:

Test equipment aficionados may wish to consider how this picture was taken. Hint: Double exposure techniques were not used. This photograph is a real time, simultaneous display of frequency and time domain information.

How did he do that? I assume he's using his trusty Tektronix 556 with a vertical-amplifier plug-in in one bay (perhaps a 1A2 or 1A4), and a spectrum-analyzer plug-in in the other bay (perhaps the 1L5 50Hz-to-1MHz spectrum analyzer). As for the distortion products, perhaps a HP 339A Distortion Analyzer?

This picture actually pretty funny at present. Tektronix is currently touting their new "Scope Revolution", the "world's first and only" mixed-domain oscilloscope (which they call the MDO4000) that has a built-in spectrum analyzer. Their ad copy says,"See both time and frequency domains in one glance. View the RF spectrum at any point in time to see how it changes." Take a look at http://www.scoperevolution.com/ and then take another look at Figure 34. Jim Williams beat them to the punch 22 years ago with technology from the 1960s.

(OK, OK, I admit that the MDO4000 does a lot more than I suggest above (for one thing, there is no logic analyzer here, nor would Jim use one, and the MDO4000 time-correlation functions are really cool); however, the superficial similarities are striking. And hilarious.)

I once heard that Tektronix offered Jim a brand new oscilloscope of his choosing, anything he wanted, if he promised to stop using vintage instruments in his app notes. No deal.



Related:

11 September 2011

Scope Sunday 8

Here's a quick family photo of my 500-series oscilloscopes. From left to right, the scopes are a 535A (with a 503 underneath), a 545A, a 547 (which I purchased especially in Jim's honor), and a 575 curve tracer (with the 175 high-current adapter underneath).


Missing from this picture is my 511A, which is still in storage in California.

I'm still looking for other 500-series scopes, especially a 556 dual beam.

(I've also updated "Scope Sunday 6" with a great letter from Bill Hewlett.)

10 September 2011

The New Book

On Thursday, I finally received a copy of the new book, "Analog Circuit Design, A Tutorial Guide to Applications and Solutions". Edited by Bob Dobkin and Jim Williams together, it is a collection of 41 Linear Technology Application Notes reprints, 27 of which were written by Jim.


Dobkin's dedication in the book is especially touching: "In memory of Jim Williams, a poet who wrote in electronics."

Unfortunately, it is just a collection of Linear Tech App Notes. While I admit that it is nice to have hardbound copies of some of these app notes (like App Note 47), I am disappointed that there is no new material here.

I have updated the bibliography to include information about this book.

09 September 2011

App Note 32

"High efficiency linear regulators." 12 pages.

This app note discusses tricks to improve the efficiency of linear regulators by decreasing the input-to-output voltage drop. For example, Figure 5 shows a SCR preregulator for an AC-to-DC regulator, similar to Figure 7 in App Note 2. The SCR circuit keeps the voltage drop across the LT1086 around 2V, thus improving efficiency. (There are a lot of circuits here that are improved versions of circuits from App Note 2.)

Most of these applications involve using a DC-to-DC switching regulator in front of the linear regulator to control the voltage drop. We've also seen this approach previously (see App Note 2 Figure 5 and App Note 29 Figure 46). Figure 8a shows a DC-to-DC regulator scheme. Figure 8a has two switching regulators, one to control the voltage (6.75V) at the input of the LT1083, and another to provide 30V to overdrive the gate of the main switch. This application has a fixed output voltage, so the main switching regulator controls the input voltage of the LT1083. Figure 8b is used where the output is variable, and the feedback path of the main switching regulator measures the voltage drop across the LT1083.

Figure 11 is a linear regulator design with only 400-mV drop out. The switching regulator in this schematic provides a large voltage only for the gate of the pass transistor. The best circuit, shown in Figure 13, combines Figures 8a and 11 into a regulator with a linear output and an efficiency between 76% and 86%. The last circuit, shown in Figure 15, is a micropower version of the previous circuits, using the LT1020 linear regulator and using its integrated comparator to implement the switching regulator loop.

Appendix A, Achieving Low Dropout, discusses the tradeoffs involved in choosing the right device for the pass element in a linear regulator. The final sentence on page AN32-10 is a close runner-up for best quote: "Readers are invited to submit results obtained with our emeritus thermionic friends, shown out of respectful courtesy." (For an example, see Figure 11A in App Note 2.)

Appendix B discusses the LT1083 family of low drop out regulators. Appendix C discusses the measurement of power consumption and shows a circuit for measuring the instantaneous power in a 120V line. This circuit is a very useful instrumentation scheme. Unfortunately, the Analog Devices 286J isolation amplifier is no longer available (and the link to the datasheet on Analog Devices' website is broken). However, 286J amplifiers are still available from second sources, but I don't have any experience with them.

The best quote is actually a Freudian slip on pages AN32-8 and 9: "A drop at the pre-regulator's output (Pin 3 of the LT1020 regulator, Trace A, Figure 16) causes the LT1020's comparator to go high. The 74C04 inverter chain switches, biasing the P-channel MOSFET switch's grid (Trace B)." Of course, he meant "gate" instead of "grid", but I think we can forgive him for having vacuum tubes on the brain. (I sincerely hope that Linear Technology never fixes this typo!)

07 September 2011

App Note 31

"Linear circuits for digital systems: Some affable analogs for digital devotees." 16 pages.

This app note is a time capsule from 1989, a window of time when "Linear Circuits for Digital Systems" didn't refer to high-fidelity audio interfaces, high-resolution video, or RF systems for networking. Instead, this note discusses unique power supplies for memories, voltage-dropout detectors, and clock circuits.

Several of the circuits are additional applications for the LT1070 family, particularly targeted to the stringent requirements of in-situ flash memory and EEPROM reprogramming. Various buffer circuits are used to present clean and well-controlled (read: free from overshoot) voltage transitions to the memory-programming pins to avoid damage. Some of these buffer circuits use the LT1010 (Figures 4 and 9), and one uses a high-current discrete buffer design (Figure 7).

Other circuits included here are voltage-dropout detectors, both for the AC line (Figure 14) and the DC supply (Figure 16), a crowbar protection circuit (Figure 18), and a power-on reset generator (Figure 20). I think the best circuit (the circuit that is still the most useful today) is the SCR-crowbar circuit in Figure 18. The main section concludes with a watchdog timer (Figure 22), and some crystal-oscillator circuits repeated from App Note 12 (Figure 24).

Appendix A is a primer on flash memory, guest written by Saul Zales from Intel, extolling the virtues of flash memory. This appendix is another interesting time capsule, especially the last line in Figure A2: Remember when updating code in flash memory was a novel concept? Remember when the only option to download new code was a floppy disk or a serial link? Good times, good times. Flash memory seemed like a revolution after having to crack open machines to pull and replace standard EPROMs.

Appendix B discusses transmission-line effects, and how voltage overshoot can destroy memory chips. Appendix C again discusses the physiology of the LT1070 family.

The best quote is on page AN31-1, with the footnote: "While I certainly wouldn't wish lifetime employment on a digital circuit board to anyone, the reality is that the need exists. (Footnote: I suppose it's not all that bad. Some of my best friends are digital circuits. If I had a daughter, I'd even consider letting her go out with one.)" I suppose it's a good thing that Jim only had a son, so this footnote confession never had to be put to the test.