标签: analogue

My first job fresh from college was testing and debugging control loading systems on flight simulators for 747 and the "new" 767 airliners. In the late 1970s, the control loop that simulated the feel of the primary flight surfaces, and had to respond instantaneously to pilot inputs, was purely analogue; the digital portion consisted of a 32-bit Gould SEL "super minicomputer" with schottky TTL and a screamingly-fast 6.67 MHz system clock; it merely provided voltage inputs via DACs into an op amp summing junction which represented slowly-changing parameters such as airspeed and pitch angle.   "Op Amp", of course, is short for "Operational Amplifier", first developed to perform mathematical operations in analogue computing. Forget about FFTs, DSPs, and all that nonsense: back in the mid-70s (when analogue giants walked the earth and the Intel 4004 was barely out of diapers) there was an analogue IC for just about everything – multiplication and division, log and antilog operations, RMS-DC conversion, you name it.   With the later rise of the Dark Side , of course, many of those old analogue components, as well as the companies that gave them life, have breathed their last.   At least some of those functions survive, either as standalone parts or (sigh) as microcontroller functional blocks. And the real world, thankfully, remains stubbornly analogue, which means that most of the truly interesting "digital" problems are really analogue problems – grounding, crosstalk, race conditions, noise, EMC, etc.   We humans are products of the real world, too. Are we analogue or digital?   The information that makes up a unique human being is mostly to be found in two places, in our genes and in our brains. The information in genes can be considered digital, coded in the four-level alphabet of DNA. Although the human brain is often referred to as an analogue computer, and is often modeled by analogue integrated circuits, the reality is more nuanced. In a fascinating discussion on this subject, computational neuroscientist Paul King states that information in the brain is represented in terms of statistical approximations and estimations rather than exact values. The brain is also non-deterministic and cannot replay instruction sequences with error-free precision; those are analogue features.   Figure 1 Figure 1: Analog, digital, and neuron spiking signals (source: Quora )   On the other hand, the signals sent around the brain are "either-or" states that are similar to binary. A neuron fires or it does not, so in that sense, the brain is computing using binary signals.   The precise mechanism of memory formation and retention, though, remains a mystery and may also have both analogue and digital components.   Is life itself analogue or digital? Freeman Dyson, the world-renowned mathematical physicist who helped found quantum electrodynamics, writes about a long-running discussion with two colleagues as to whether life could survive for ever in a cold expanding universe . Their consensus is that life cannot survive forever if life is digital, but life may survive for ever if it's analog.   What of my original topic - analogue vs digital computation? In a book published in 1989, Marian Pour-El and Ian Richards, two mathematicians at the University of Minnesota, proved in a mathematically precise way that analogue computers are more powerful than digital computers. They give examples of numbers that are proved to be non-computable with digital computers but are computable with a simple kind of analogue computer.   Consider a classical electromagnetic field obeying Maxwell's equations: Pour-El and Richards show that the field can be focused on a point in such a way that the strength of the field at that point is not computable by any digital computer, but it can be measured by a simple analogue device.   The book is available for free download . Digital engineers, knock yourself out.   Meanwhile, analogue rules supreme. Which, of course, we analog engineers knew all along.   Now, about that raise.....   Paul Pickering

In the recent months, I've been working with Max Maxfield on the analog meter problems and design challenges pertaining to his Inamorata Prognostication Engine , Ultra-Macho Prognostication Engine , and Vetinari Clock projects.   As you may have read in his columns, Max and I have been in almost constant daily communication, by either email or phone. During our conversations, we have had a blast talking about various aspects of these meters, such as the fact that some meter movements deflect 100 degrees, while others deflect only 90 degrees. This point took Max by surprise after he incorrectly had his graphics guru Denis create the Vetinari Clock's "Hours" faceplate assuming a 100-degree movement when the meter was designed to operate over a 90-degree swing. Fortunately, we managed to make the meter's movement match the faceplate.   As part of this, I've been telling Max about some of the meters we make and repair here at Instrument Meter Specialties (IMS). Take the small Triplett .5E edgewise meter that goes in an older aircraft. These meters read "GOOD" in a green area in the middle, with red areas on either side. They would cost hundreds of dollars when the aircraft were created, but they now sell for thousands. The thing is that the planes were certified with these meters, which are FAA approved, so no substitutions are allowed.   Another job I was telling Max about involved an upgrade on some meters that go in a nuclear power plant. These meters were stockpiled when the plant was commissioned. As they started to reach their end of life, replacements were brought online, but they were failing their 0.6% calibration tolerance, and we had to work our magic on them. (I'll describe how we achieved the required accuracy in my next column.)   We do a lot of work on the older Hickok tube tester meters. We've also experienced an analog meter resurgence in the recording industry. In fact, we make several different meters for old audio compressors from companies/products such as Fairchild, Gates Sta-Level, and Federal TV. Being a custom meter shop with the ability to make complicated artworks means we can satisfy just about any meter-related requirement, so long as the parts are available.   Based on his projects, the topic Max and I have spent the most time discussing by far has been the creation of artworks (faceplates) for the meters. A lot of work goes into analog meter movements and artworks. This tends to be why new meters can cost a lot of money. The parts can be as small as those in Swiss watches, and it takes skilled hands to make a really good meter movement. This is why we always have our trusty microscope at the ready. Even meters with a "linear scale" only have movements rated to be within 1-2% of the full scale input at any given point of the scale. The trick to this "linearity" is accurately placing the coil within the magnetic field and making sure that nothing affects the spring constant, like friction caused by two spring turns rubbing against each other or deformations such as creases in the spring. That said, as important as the meter's quality is to accuracy, equally important -- especially for high accuracy -- is the meter's artwork.   The highest-accuracy analog meters actually have the divisions on their artworks printed to address nonlinearities in the meter movement's, um... movement across the dial (faceplate). This is interesting, because it allows lower-quality meter movements to compete with higher-quality ones, especially if the nonlinearity is somewhat repeatable during production. That said, making nonlinear artworks is challenging. It can require a great deal of time, knowledge of interpolation that few have, and/or the use of custom software. All of this greatly complicates the creation of meter artworks, and this was especially true earlier in the history of analog meters.   The process of printing of analog meter artworks has undergone several changes over the last century. Originally, everyone was using the same methods -- offset printing press and hand-drawn. Hand-operated offset presses were typically used for printing more than a single dial, while hand-drawn dials were created for prototypes and single-piece custom orders. The offset presses used were a bit different from letter presses, as they were meant to print on to metal and not paper.   The Grauel model R-1 printing press as used by many meter shops and manufacturers. The ratchet mechanism made a distinctive hollow clanking sound like a sad bell.   Aside from that, the process was similar. There are five main components on an offset press: the ink disk, ink rollers, vacuum table for the positive plate, rolling printing pad, and printing table where the dial blank goes. The ink was taken from the inking disk to the positive plate by the ink rollers. Next, the rolling printing pad took the ink off the positive plate and rolled it on to the dial plate. Single-color artworks were typically the norm, because cost was prohibitive for more colors. Requesting another color meant a great deal of headache for the meter manufacturer, since each added color required a separate artwork.   Before printing could begin, a hand-drawn artwork would be created using India ink on a substrate that allowed the easy and accurate removal of the dried ink. Fonts could be penned using a KE Leroy set or large Letraset sheets.   I have never actually seen one of these Leroy sets in operation, but it was fun having my grandfather show me how they worked.   All this work would be done at about 4X scale to increase accuracy. The artworks were then sent out to be photo-reduced to actual size, and a matching film transparency would be created. From that point, the transparencies would be used to burn an offset press plate.   Here's an example of a zinc offset press plate and the corresponding faceplate. I have actually seen these used.   The transparencies were retained in case a new plate was ever needed. This occurred quite often after the material used to form the plates changed from zinc to a plastic, which suffered from warping and cracking.   A zinc offset press plate and its plastic equivalent. As you can see, the plastic positive plates did not fare well over time.   Once the plates were received back in house, the actual printing could begin. During the printing of each artwork, faceplates were often scrapped due to alignment issues. Remember that each color required its own artwork. Due to problems with alignment, a three-color artwork might require 20-30 attempts to obtain five "good enough" faceplates. Observe the photograph of the Hickok faceplate below, and note how the red line is not quite in the right place.   Known for their vacuum tube testers, most Hickok artworks had at least two colors. It is impressive how many of them were actually better than this one.   The tolerances for alignment were very tight -- within about 0.015" for single-color artworks and 0.005" from the first layer for each additional color. Making things more difficult, the faceplate's screw/mounting holes were the registration marks, and they were only about one inch apart. Much of the success came down to the skill of the individual press operator to align the registration marks of each plate in relation to the dial. This is not as easy as it sounds, and there was a long tradition of rude phrases moving ballads during operation. As they do to this day, customers would call in the hope of expediting their orders. This pressure didn't help the printers who were waiting for the ink to dry in their industrial ovens. Each printing would take a day to dry before the next could be performed.   Not surprisingly, there was an extra cost added for each color on the faceplate. Additionally, for each color there would be an increase to the cost of each finished meter due to the difficult task of aligning the artworks with one another. Back in the 1960s to 1980s, for example, artworks would have started at $75 for one color with $25 for each additional color. Converting 1960s dollars to today's value, that comes to about $600 and $200, respectively. Jewell Instruments currently charges $150 for a one-color custom scale. Each additional color costs $75. It would seem the company still uses offset printing, but it probably works with computer-generated transparencies these days. Compare that to $0 to $75 for artworks generated here at Instrument Meter Specialties using a direct-to-dial process and our own PHP script. (Again, I'll describe this in more detail in my next column.) Back then, however, these services were seen as a loss leader. The goal of the manufacturer was to minimize loss and hopefully generate a big order.   Perhaps those big orders were more of a common thing back in the heyday of analog panel meters. I found the following picture of an Apollo control room at NASA with more than 100 analog panel meters in view.   I counted more than 100 analog meters that I could identify in this picture (which is presented here with the permission of Shaun O'Boyle), but there may be quite a few more.   Installations involving such quantities ensured that any loss associated with creating the faceplates was recovered in the cost of each meter. That said, low-quantity custom orders required an entirely different printing method. Faceplate blanks with division markings would be stocked, and numbers would added by hand using Letraset dry transfer characters.   I have used Letraset pages before. I was trained to use a pencil and to push hard, so I could see where I had rubbed and perhaps take advantage of the graphite as a lubricant. I remember using these sheets with mixed success.   Even though these faceplates were created by hand, new artworks were still charged at $75 for single colors. This covered the time required for a certain skilled someone to sit down and -- using dry transfer sheets -- carefully place each character on the faceplate with the proper spacing. Many times, this was good enough for customers who wanted multiple colors, especially since colored regions could be added on to the division set with a permanent marker or using some other DIY method. Prior to the 1960s and rub-on lettering, a KE Leroy pen set could be used to ink a custom dial (an image of this was shown earlier). This was more difficult than the rub-on lettering, and it was abandoned after rub-on letters became available.   From about 1995 to 2005, the meter industry experimented with, and tried implementing, various printing techniques. Many still retained the offset press method, if only for current artworks. One popular method was lovingly referred to as "using paper dials." This meant that the faceplate image was printed on a waterproof surface that would absorb the ink, and then this "paper" would be adhered to the blank dial. This technique employed mid-level and high-range consumer inkjet printers for color matching purposes. This became more popular in the late 1990s with the availability of inkjet bumper-sticker sheets. Even though this approach saved much time over the offset press method, it was still a fairly time-consuming process. Laser etching was attempted in the late 1990s, but the low number of custom built machines, the high cost of those machines, and the lack of duo-toned or multi-toned substrate that would work in different brands of meters meant that adoption was low.   Eventually, the "Holy Grail" of faceplate printing was found in small format, inkjet, flatbed presses. But this still leaves the task of creating the artwork in the first place. In my next column, I will describe an innovative technique we developed here at Instrument Meter Specialties that made Max say, "Wow, I am very impressed." Until then, I welcome any comments and questions.   Jason Dueck Product Designer Instrument Meter Specialties

An analogue engineer and a digital engineer collaborate, use their respective skills, and pull a few bunnies out of a hat to troubleshoot a system with which they are completely unfamiliar. Our sales department had just accepted a new challenge on behalf of Engineering. They promised a customer that yes, of course, we can repair a telecom product that we have never seen before and for which we have no systems, no test fixtures, and no schematics. (The OEM no longer supported this product.)   Engineering was once again expected to shake our rattles, do our magic voodoo dance, and pull bunnies out of hats. About fifteen of these backplane-pluggable boards showed up in my office for initial evaluation and perusal of their inner workings. They had a proprietary SIMM (socketed memory module), which on several units turned out to be bad. Temporarily substituting the memory modules from other cards with obvious smoke damage failure modes brought them back to life when powered while lying flat on the bench. (Remember, there was no test chassis available.) They would then boot and talk to us over their RS232 ports.   These modules were populated with four SRAMs and four flash memories, each flash and SRAM shared an 8-bit-wide data bus, and each pair of SRAMs was enabled together with the same chip select. I proposed to the boss that we build a small test fixture that would take the DUT memory module, run SRAM tests, and if necessary reprogram the flash.   A digital/software colleague three cubes away was assigned to work with me on this project. He had previously designed and laid out a PCB that used a surface-mount PIC microcontroller as a universal I/O for our current and future test fixtures. It turned out that it had just enough I/O lines to handle the address and data buses on the DUT memory module, with two spares, as long as I tied the four separate DUT data busses together into two pairs on the fixture. So we decided to use it.   I ordered the necessary SIMM connector and a plated-through-hole protoboard, along with some ribbon cable and IDC header sockets to connect to the PIC board. It was somewhat annoying that the 72-pin SIMM connector was spaced at a.05-inch pitch, so the protoboard also had to be this pitch. Its tiny .025-inch-diameter holes did not accept .025-inch-square pins, so wire-wrap was impossible. (Now I know where that old adage, "Can't fit a square peg into a round hole," came from.)   I had to solder ribbon cable directly to the protoboard and string short 30AWG wires to the SIMM connector. As long as the stranded ribbon wires were not overly tinned (to keep the strands together), they actually fit into the protoboard holes.   Endeavor brings back cuss words long since forgotten Another annoyance was that the SIMM connector had plastic retaining tabs that quickly wore out from repeated insertions of memory modules. The maker had designed them for maybe a single SIMM replacement over the lifetime of the product. We wanted to plug DUTs in and out constantly.   Fortunately I had used socket pin strips in the protoboard for the SIMM connector in anticipation of eventually needing to replace it easily. I subsequently found a connector with metal retaining tabs. This particular feature does not show up in vendors’ online part descriptions. I had to look at the mechanical drawing of each of many to find "W/ Metal Latch."   The first test of the fixture went well. My colleague coded a walking-ones SRAM test that immediately identified bad SRAM chips on a couple of the DUT (Device Under Test) boards. We replaced them and now they booted, but with the disconcerting message "RAM is BAD." Due to availability we had used 12 nsec SRAMs in place of the original 20 nsec SRAMs, so speed was probably not the issue. Hmmm, maybe we needed to improve the test algorithm.   Then we got brave and copied about five different versions of firmware from the flash of the good memory modules and tried to re-write the new firmware into a module, which semi-booted at first but complained about a "missing application loader." After the firmware re-load it would no longer even talk to us over its RS232 port. Somehow a 'known good' firmware load messed it up. My colleague verified that the firmware in the good and bad modules was identical. So why did one boot and not the other? Speed?   My colleague continued writing his code and progressed to a walking-zeros test. Strange things began to happen. On several known good memory modules two SRAMs with their data busses tied together consistently failed in the same way: When 7F was written, FF got read back. It only failed on one pair of SRAMs. The other SRAM pair always worked properly.   Had I connected a wire wrong on the fixture? We put a scope on the fixture and verified that yes, when he wrote 7F that is what came back from the DUT SRAM and the fixture. Clearly his PIC microcontroller was reading a definite logic 0 as a logic 1, but only on bit 7 of that data bus. Yet the walking one's test had worked and bit 7 was correctly read as a logic 0 during that test.   Since I was not familiar with his PCB layout or the PIC chip, I asked him to send me his KiCAD board layout file. I already knew there were no power/ground planes, but I had not expected to see that some of his ground pin connections snaked in and out in roundabout paths when they should have all been joined together under the PIC chip.   Some of his Vdd connections were not even connected to the Vdd copper, but instead depended on connections within the chip. His decoupling capacitor was an inch away, adding two inches of trace inductance. I smelled analogue problems here, possibly due to the power routing. One way to find out if a suspect actually is the cause of a problem is to eliminate it. I used an approach that had been successful before, which was to add power planes and more decoupling. Here is a photo of the end result, done by one of our highly-skilled production soldering experts:   Two squares of single-sided copper-clad form the mini-power planes. Decoupling 0805 chip capacitors standing on their ends are just the right size to AC-couple the planes together. (Somehow this sounds like an oxymoron). The PIC cannot complain about poor power etch routing. All its power and ground pins are now tied together.   Unfortunately this did not help. But it did eliminate the power suspect. I still smelled an analogue problem.   This was further confirmed when we ran some tests to see if any other byte patterns caused bit 7 to falsely read a one when it was really a zero. Turned out there were many patterns that did this. If as few as three lower-order bits were ones, the PIC would read bit 7 as a one when it was really a zero. It didn't seem to matter which lower order bits, all it took was three or more set to one. With enough of them HI they seemed to bleed into bit 7. Was it analogue voltage summation?   Then it hit me. My colleague's PIC was running at 3.3V. My memory module DUT was powered at 5Vs. My colleague had previously assured me that his PIC inputs were 5 volt tolerant -- the data sheet said so. I took a closer look at the data sheet. On the first page it did say "5.5V Tolerant Inputs (digital-only pins)." So if the inputs are configured as digital, they should be 5V tolerant, right?   Some 146 pages into the data sheet was a bit (no pun intended) more detail: Any inputs that could be configured as either analogue or digital are NOT 5V tolerant. They have clamp diodes to 3.3V Vdd. All eight bits of the problem data bus and one bit of the other data bus went to such inputs. Yes, it was an analogue problem -- the 5V ones were overdriving the inputs and adding voltage-wise. I invented a couple of new cuss words.   This explained the problem with the one flash we had overwritten that would no longer boot. All the firmware images we had copied previously were garbage. I had to heat up the soldering iron again, hack into the test fixture, and carefully cut ribbon cables to add a couple of 74LVC245 bus transceivers with 5V tolerant inputs. My knowledge of PIC microcontrollers and my expletive vocabulary both improved considerably.   But it solved the problem and we could now identify bad SRAM devices and re-write the bad flash. The "RAM is BAD" message turned into "RAM is OK" after a flash re-write. Possibly the flash had logged the previous SRAM failures.   Success was achieved by a pair of engineers, one digital and one analogue, each with his own skill set, working together to solve the problem.   Glen Chenier Engineer

An analogue engineer teams up with a digital engineer, they use their respective skills, and pull a few bunnies out of a hat to troubleshoot a system with which they are completely unfamiliar. Our sales department had just accepted a new challenge on behalf of Engineering. They promised a customer that yes, of course, we can repair a telecom product that we have never seen before and for which we have no systems, no test fixtures, and no schematics. (The OEM no longer supported this product.)   Engineering was once again expected to shake our rattles, do our magic voodoo dance, and pull bunnies out of hats. About fifteen of these backplane-pluggable boards showed up in my office for initial evaluation and perusal of their inner workings. They had a proprietary SIMM (socketed memory module), which on several units turned out to be bad. Temporarily substituting the memory modules from other cards with obvious smoke damage failure modes brought them back to life when powered while lying flat on the bench. (Remember, there was no test chassis available.) They would then boot and talk to us over their RS232 ports.   These modules were populated with four SRAMs and four flash memories, each flash and SRAM shared an 8-bit-wide data bus, and each pair of SRAMs was enabled together with the same chip select. I proposed to the boss that we build a small test fixture that would take the DUT memory module, run SRAM tests, and if necessary reprogram the flash.   A digital/software colleague three cubes away was assigned to work with me on this project. He had previously designed and laid out a PCB that used a surface-mount PIC microcontroller as a universal I/O for our current and future test fixtures. It turned out that it had just enough I/O lines to handle the address and data buses on the DUT memory module, with two spares, as long as I tied the four separate DUT data busses together into two pairs on the fixture. So we decided to use it.   I ordered the necessary SIMM connector and a plated-through-hole protoboard, along with some ribbon cable and IDC header sockets to connect to the PIC board. It was somewhat annoying that the 72-pin SIMM connector was spaced at a.05-inch pitch, so the protoboard also had to be this pitch. Its tiny .025-inch-diameter holes did not accept .025-inch-square pins, so wire-wrap was impossible. (Now I know where that old adage, "Can't fit a square peg into a round hole," came from.)   I had to solder ribbon cable directly to the protoboard and string short 30AWG wires to the SIMM connector. As long as the stranded ribbon wires were not overly tinned (to keep the strands together), they actually fit into the protoboard holes.   Endeavor brings back cuss words long since forgotten Another annoyance was that the SIMM connector had plastic retaining tabs that quickly wore out from repeated insertions of memory modules. The maker had designed them for maybe a single SIMM replacement over the lifetime of the product. We wanted to plug DUTs in and out constantly.   Fortunately I had used socket pin strips in the protoboard for the SIMM connector in anticipation of eventually needing to replace it easily. I subsequently found a connector with metal retaining tabs. This particular feature does not show up in vendors’ online part descriptions. I had to look at the mechanical drawing of each of many to find "W/ Metal Latch."   The first test of the fixture went well. My colleague coded a walking-ones SRAM test that immediately identified bad SRAM chips on a couple of the DUT (Device Under Test) boards. We replaced them and now they booted, but with the disconcerting message "RAM is BAD." Due to availability we had used 12 nsec SRAMs in place of the original 20 nsec SRAMs, so speed was probably not the issue. Hmmm, maybe we needed to improve the test algorithm.   Then we got brave and copied about five different versions of firmware from the flash of the good memory modules and tried to re-write the new firmware into a module, which semi-booted at first but complained about a "missing application loader." After the firmware re-load it would no longer even talk to us over its RS232 port. Somehow a 'known good' firmware load messed it up. My colleague verified that the firmware in the good and bad modules was identical. So why did one boot and not the other? Speed?   My colleague continued writing his code and progressed to a walking-zeros test. Strange things began to happen. On several known good memory modules two SRAMs with their data busses tied together consistently failed in the same way: When 7F was written, FF got read back. It only failed on one pair of SRAMs. The other SRAM pair always worked properly.   Had I connected a wire wrong on the fixture? We put a scope on the fixture and verified that yes, when he wrote 7F that is what came back from the DUT SRAM and the fixture. Clearly his PIC microcontroller was reading a definite logic 0 as a logic 1, but only on bit 7 of that data bus. Yet the walking one's test had worked and bit 7 was correctly read as a logic 0 during that test.   Since I was not familiar with his PCB layout or the PIC chip, I asked him to send me his KiCAD board layout file. I already knew there were no power/ground planes, but I had not expected to see that some of his ground pin connections snaked in and out in roundabout paths when they should have all been joined together under the PIC chip.   Some of his Vdd connections were not even connected to the Vdd copper, but instead depended on connections within the chip. His decoupling capacitor was an inch away, adding two inches of trace inductance. I smelled analogue problems here, possibly due to the power routing. One way to find out if a suspect actually is the cause of a problem is to eliminate it. I used an approach that had been successful before, which was to add power planes and more decoupling. Here is a photo of the end result, done by one of our highly-skilled production soldering experts:   Two squares of single-sided copper-clad form the mini-power planes. Decoupling 0805 chip capacitors standing on their ends are just the right size to AC-couple the planes together. (Somehow this sounds like an oxymoron). The PIC cannot complain about poor power etch routing. All its power and ground pins are now tied together.   Unfortunately this did not help. But it did eliminate the power suspect. I still smelled an analogue problem.   This was further confirmed when we ran some tests to see if any other byte patterns caused bit 7 to falsely read a one when it was really a zero. Turned out there were many patterns that did this. If as few as three lower-order bits were ones, the PIC would read bit 7 as a one when it was really a zero. It didn't seem to matter which lower order bits, all it took was three or more set to one. With enough of them HI they seemed to bleed into bit 7. Was it analogue voltage summation?   Then it hit me. My colleague's PIC was running at 3.3V. My memory module DUT was powered at 5Vs. My colleague had previously assured me that his PIC inputs were 5 volt tolerant -- the data sheet said so. I took a closer look at the data sheet. On the first page it did say "5.5V Tolerant Inputs (digital-only pins)." So if the inputs are configured as digital, they should be 5V tolerant, right?   Some 146 pages into the data sheet was a bit (no pun intended) more detail: Any inputs that could be configured as either analogue or digital are NOT 5V tolerant. They have clamp diodes to 3.3V Vdd. All eight bits of the problem data bus and one bit of the other data bus went to such inputs. Yes, it was an analogue problem -- the 5V ones were overdriving the inputs and adding voltage-wise. I invented a couple of new cuss words.   This explained the problem with the one flash we had overwritten that would no longer boot. All the firmware images we had copied previously were garbage. I had to heat up the soldering iron again, hack into the test fixture, and carefully cut ribbon cables to add a couple of 74LVC245 bus transceivers with 5V tolerant inputs. My knowledge of PIC microcontrollers and my expletive vocabulary both improved considerably.   But it solved the problem and we could now identify bad SRAM devices and re-write the bad flash. The "RAM is BAD" message turned into "RAM is OK" after a flash re-write. Possibly the flash had logged the previous SRAM failures.   Success was achieved by a pair of engineers, one digital and one analogue, each with his own skill set, working together to solve the problem.   Glen Chenier Engineer

One thing that differentiates humans from other animals, as far as we know, is that we are fascinated with the future. I'm not talking about predictions for tomorrow's weather, or even next week's, I am talking about years in advance. Let's face it: fortune tellers, soothsayers, and crystal balls are an established part of human history, except now it is known as market research and reports.   Cerainly, it's sometimes possible to make such predictions with a reasonable level of confidence, such as when it is demographic-related. But in the technology world, any prediction more than a few years ahead should be viewed as note much more than a sophisticated guess.   That's why I was surprised to see a market report about the analogue market in 2020. While I am interested in almost everything analogue, seems to me that you might as well cast bones or roll dice when you try to say who , what, where, and how much the analogue market (or any market) will look like a decade ahead.   There are two reasons for this: first, since such predictions are largely extrapolations, any small misassumption or error in initial data points will translate into a major miss. It's the classic trajectory and targeting problem when you start on a long journey: being off even slightly in your initial position, or a fraction of a degree in your heading, results in missing your objective by hundreds of miles (barring any mid-course correction, of course).   Second, by definition, such predictions can't account for disruptive technologies or market shifts. For example, the market for vacuum tubes look very different in 1946 than it did in five years later, a period during which the transistor was born and began commercialization . Similarly, the development of the monolithic integrated circuit (IC) changed electronics in terms of components, applications, and . . . . well, absolutely everything.   Lest you think that I am just sniping at those who make such long-term predictions, I too have been guilty of the same, though much more informally. I clearly recall when the USB standard was introduced in 1994, I was pretty sure about its prospects: "yeah, just what we need, yet another interface standard" and "seems kind of foolish, just to replace the keyboard and mouse connections in a PC."   We know how that turned out. USB is an extremely important and common interface, now going into USB 3.0, and with lots of industry support at hardware, software, end-product, and applications levels. Regardless of what you think about its technical virtues, it's clearly been a winner, and I was 100% wrong. (Don't forget, any attempt at using it higher-speed USB was supposed to be overwhelmed by IEEE-1394, aka Firewire). USB is now even used just as a basic 5V power source, which didn't seem to be a rationale for it when it was developed.   I'm not embarrassed; most long-range predictions are barely "roughly right." To quote Yogi Berra, whose many sayings have more real-world experience and sensibility than you'll get from a room of erudite philosophy professors: "It's tough to make predictions, especially about the future." Just check (cheque for banks) out the 1950 science-fiction movie Destination Moon which attempted to be technically sound, and compare it with the reality which occurred less than 20 years later—and you've realise that humility when making predictions is a very good idea!   So take a moment and think: what are some of the technology and market predictions which you made to your friends, and which turned out to be wrong—regardless of reason they did so? What things succeeded which you were sure would fail, or vice versa?   Go ahead, reflect back and then admit to it!