Skip navigation
1 2 3 Previous Next

The Engineering Mind

130 posts

Space Mission Lifetimes

Posted by J.Harris Employee Aug 20, 2018

I have been at ADI now for the last 8 years and am working on year 9.  This is fairly short by ADI standards with many here with 15-20+ year tenures.  However, in my total career now of about 14 years I have been able to see a good bit of advancement in the world of communications.  I recall first having the ability to get the 'internet' on my phone those many years ago with blazing fast 2G data rates.  I guess you could sort of call it the internet since the early sites were actually WAP (Wireless Application Protocol) sites for mobile devices and they were a little different than the standard website...and, of course, we all know now that 2G was anything but blazing fast.  Could you imagine waiting on a website to load today at 2G data rates (~700kbps by the way)???  I think we'd all get bored pretty quickly.  It sure was amazing at the time though!  Today we see 5G on the horizon (if its not here already) with its data rates in the Gbps range.  That is a few orders of magnitude better than it was just 10-12 years ago.  How quickly things have advanced!  Advancements in design have been happening at break neck pace for a while.  Engineers have to churn out design spins and improvements in 6 months to one year time frames, possibly less in some circumstances.  These designs last a few years in the field before they become obsolete due to the increased speeds in the latest designs.


By contrast, the design time frames for space applications are typically 3-5 years with mission lifetimes of 10, 15, and 20 years, more in some cases.  It is quite interesting to have had the opportunity to work in such drastically different types of markets.  One is high volume with short design cycles while the other is low volume with much longer design cycles.  Speed is of the essence in one while reliability is of the essence in the other.  In my latest blog on Planet Analog, Planet Analog - Jonathan Harris - Space Mission Lifetimes, I look at the mission lifetimes of various NASA missions and discuss some of the design considerations for space applications.  I included a neat survey of a few of the mission life times of various NASA missions as well.  You can see that graphic below as well. 


It is amazing to see the progression in communications over the last 10-12 years, but it is even more amazing to see the length achieved in some of these NASA missions.  The Hubble Space Telescope has now been in service for nearly 30 years.  The International Space Station has been deployed for nearly 20 years.  This is a testament to the attention to detail in the design and development of space products as well as the high level of screening and testing that these products undergo to qualify for space applications.  Take a look at my latest blog and let me know what you think.


In my last blog, I wrote about a young man from Ohio who came to Boston and worked feverishly in a Scollay Square laboratory on an invention whose sale he hoped would start him on his fortune. But the young man was unable to sell the device and so, when he decamped Boston for New York City, the young man - named Thomas Edison - was as broke as when he arrived two years earlier. There he had a reversal of fortune when he sold the rights to a stock ticker he had begun designing back in Boston. Using those funds to build laboratories in New Jersey, Edison began churning out inventions that literally changed the world. By 1880, the man who had produced the phonograph and the first practical light bulb was being called "The Wizard of Menlo Park."


But, when it came to that light bulb, the Wizard had a dark side.


First, a bit background. For most of human history the only way to produce light was to burn something. We have evidence from over 125,000 years ago that early man knew how to build and, just as important, control fire. 50,000 years later we see remnants of hollow shells and rocks in which was burned moss soaked in animal fat. By 4,500 B.C.E. oil lamps were providing more efficient light sources that burned less but glowed brighter, and were portable. Candles, developed by 3,000 B.C.E., would provide light that lasted even longer, without the deadly oil fumes. Then... well, that's pretty much it for lighting technology for the next four or five millennia. Be it naturally produced coal, gas, or man-made kerosene (processed in the mid-1800s), creating light was still a burning issue. (Sorry, couldn't help myself)


The first electrical solutions, known as arc lamps, produced prodigious amounts of light but required up to 6,000 volts of AC (alternating current) to operate. They were also noisy and the light they produced was harsh, which limited their use to the outdoors. Subsequent refinements by men such as Charles Brush allowed for their use indoors, but only in large venues such as factories and theaters. What the world needed was a safe artificial light that could be used in the home. Safe was the key word, as some very public and horrific electrocutions had killed linemen and others who had strayed too close to the arc light's power source.


This was precisely the sort of problem that drew Thomas Edison throughout his life. With a fresh set of eyes, he studied the failures of others so he could adjust his approach accordingly. Perhaps more important to Edison was that the world was eager for a safe electric light, meaning great fortune to the successful inventor. As a young man in Boston he learned what he later said was his most valuable lesson: never invent anything that you cannot sell. This drive for financial success caused something of a backlash to the Edison legacy in the anti-establishment 1960s, something I witnessed first-hand while studying history at the University of Maryland in the mid-1970s. Edison, seen as the Wizard for much of the past century was, by the time of Nixon's resignation, under attack by historians who focused on episodes like the one I am about to describe. It placed him not in the company of scientists like Faraday and Bell, but with Gilded Age robber barons like Stanford, Carnegie, and Rockefeller.


By most accounts, the first light bulb was created in 1850 by an English physicist named Joseph Swan, although it was ten more years before he had a reliable working prototype. And even then it wasn't very bright, nor did it last very long. Swan never overcame the technical deficiencies of his light bulb, and commercial success eluded him. Same for the Canadian team of Henry Woodward and Matthew Evans, although they got further with their design (and even secured a patent) they were unable to generate enough revenue to continue their research. In 1879 they sold their patent to Thomas Edison, who was a year or so into his own serious attempt to build a commercially viable electric light. In fact, Edison had already filed his very first patent for an early version of his light bulb, in October of 1878.


Edison knew from bitter experience the value of a patent. The one for his stock ticker that he sold for $40,000 in 1870 was already worth millions. Sure, $40,000 in the hand was worth... well, he wasn't going to make the same mistake twice. What we today call IP (Intellectual Property) was worth as much, if not more, than the slivers of carbon filaments that glowed inside those glass tubes, and from 1878 to 1880, even as he continued to make improvements to his incandescent light bulb, he continued to file new patents.


In 1880, the Edison Electric Light Company began to market the first commercially viable light bulb. Inside was the breakthrough— a carbonized bamboo filament that could last over 1200 hours.


Original carbon-filament bulb from Thomas Edison

Original carbon-filament bulb from Thomas Edison (image courtesy


Edison next had to address the challenge of how power for the light bulb should be distributed to businesses and residences. The Wizard had a very definite opinion on this subject. With the deaths from the high-voltage alternating current required for Arc lights still fresh in the public's mind, Edison advocated a direct current system that could safely deliver the necessary power for the light bulbs. A system that only his company made. His laboratory even built a meter (which worked only on DC, of course) that could measure the consumption of power by each home or business, thus ensuring that everyone paid their fair share. To the Edison Illuminating Company, of course. To quell the public's— and Edison's—safety concerns, his company utilized what is best described as a fuse that shut down the power system if there were a surge, something his AC competitors did not have.


Not everyone was convinced that AC was the public safety hazard promulgated by Edison. In 1886 George Westinghouse, using advances in transformer technology developed by the Hungarian "ZBD" team of Károly Zipernowsky, Ottó Bláthy and Miksa Déri (and improved upon by American William Stanley in his Pittsburgh tests) lit 23 businesses in Great Barrington, Massachusetts using AC. Not only could the AC system support more customers, it also proved to be much cheaper to build and maintain than Edison's DC system. The biggest drawback to the Edison system was the simple fact that DC cannot be sent much farther than a mile. Extending the range required power stations— each with large dynamos —built every few square blocks, as seen in this screen shot from the PBS American Experience episode War of the Currents.



Manhatten DC power grid

Edison's NYC power grid (courtesy Con Edison)



That combination of AC's advantages led to the installation of the first commercially viable AC system in Buffalo, which motivated other cities to sign contracts with Westinghouse for their own AC systems. Here was a direct threat to the Wizard and his DC system. And the Wizard did not like threats.


So began the War of the Currents.


The first shots fired were legal ones. Edison was not the only inventor to appreciate the importance of patents. Westinghouse and a number of other electrical distribution entrepreneurs would file— and then have challenged by one of the others— dozens of patents for power systems, light bulbs, and other electrical devices. But it wouldn't have been much of a war if it was just a bunch of men in high-button collars and pince-nez glasses passing court papers back and forth. The war between DC and AC would be fought in public.


We return to New York, where hundreds of wires for telephone, telegraph, and high-voltage arc lighting were a foreboding presence overhead. Unregulated by the city, the wires for all those different services haphazardly crisscrossed each other. During the Blizzard of 1888, several high-voltage arc lighting lines fell and utility services were interrupted all over the city. Then, a few months after the blizzard, a boy was killed by a high-voltage wire that had been left dangling. Edison's fears of AC were proven to be real with this and other deadly accidents.



New York City during the Blizzard of 1888 New York City during the Blizzard of 1888


The reputation of Thomas Edison takes something of a hit during this war due, in part, to the myth that he had an elephant electrocuted as a way to demonstrate the dangers of high-voltage AC. I'm happy to bust that myth. Topsy, the unfortunate elephant in question, was put to death after killing a drunken spectator at the Coney Island park where she had been on display (in all fairness, the spectator burned Topsy's trunk with a cigar, so who can blame her?) But the demise of Topsy was in 1903, years after the War of the Currents was pretty much settled, and except for Edison Studios filming the electrocution (warning: graphic content) the Wizard himself had no role in this event.


The confusion about Edison and the electrocution of Topsy likely comes from Edison's association with engineer Harold Brown. Brown, like Edison, saw only danger and death from AC, and took it upon himself to campaign against Westinghouse's AC system in letters to city newspapers and the city's electrical board (Westinghouse countered in letters of his own of how DC equipment had caused fires.) This very public sniping happened around the same time that the state of New York legislated using electricity as a method of execution.


A lot happens in a short period of time, and it can be confusing to unravel the timing because there are a lot of overlapping of events. We'll start with the one little detail that had been left out of New York's electrocution bill: what kind of electricity— AC or DC— should be used. The New York Medico-Legal Society was tasked with choosing between the two, and how much voltage was needed to "humanely" execute a human being. When Edison, who claimed to be personally against the death penalty, was asked what would be the best method for performing electrocutions, he replied "Hire out your criminals as linemen to the New York electric lighting companies."


It was around this time that the two DC proponents, Brown and Edison, got acquainted. Precisely how is not completely clear, but we know that by the summer of 1888 Brown had a space at Edison's West Orange laboratory where he began a series of often gruesome electrocutions, at first using stray dogs. Later that year, before the press, members of the Medico-Legal Society, and Edison himself, four calves and a horse were killed using high-voltage AC. The tests served two purposes. First of all, it was clearly a tact meant to associate the electric chair with Westinghouse's high-voltage AC system, and spread fear among the public about bringing it into their homes. It also led to the recommendation that AC at 1000 - 1500 volts be used for execution.


Horse electrocution

The electrocution of a horse by Harold Brown at Edison Labs


Westinghouse was furious and slammed Brown, Edison, and the tests they had performed as baseless, self-serving, and incorrectly performed. He attempted, unsuccessfully, to block the State of New York from acquiring one of his AC generators for their new "electric chair," which would make its debut with the capital sentencing of a street merchant named William Kemmler. After a series of failed legal appeals, Kemmler was strapped to the chair on August 6, 1890 for not one, but two gruesome jolts of electricity. Westinghouse is reported to have said, "They would have done better using an axe."


In his wildest imagination, author John Grisham could not have conjured up the next chapter in the Current War. Just three weeks after the botched electrocution of William Kemmler the following headline appeared on the front page of The New York Sun: For Shame, Brown! - Disgraceful Facts About the Electric Killing Scheme; ***** Work for a State's Expert; Paid by One Electric Company to Injure Another. The article below detailed the scheme between Brown and Edison to surreptitiously acquire Westinghouse generators for New York's electric chair. It also told how Edison's company paid for the printing and distribution of anti-AC pamphlets to members of the Missouri legislature, then debating which system to install in their state. Turns out that somebody had broken into Harold Brown's office and stolen 45 letters—correspondence between Brown and Edison. The Sun was happy to use the letters as a basis for the article, which severely compromised both Brown and Edison's public stature and cast doubts of the veracity of their anti-AC bias. But both men, especially Brown, would quickly be rehabilitated following the October death of a man named John Feeks.


Feeks was a Western Union lineman working high above a downtown Manhattan street when he touched a shorted high-voltage AC line. Randall Stross, in his 2007 book The Wizard of Menlo Park, describes how "...the jolt entered through his bare right hand and exited his left steel studded climbing boot. Feeks was killed almost instantly, his body falling into the tangle of wire, sparking, burning, and smoldering for the better part of an hour while a horrified crowd of thousands gathered below." Feeks' death was not only gruesome, but it was very public. His death came after at least six others— five linemen and one innocent bystander— were electrocuted by high-voltage AC lines.


That the electrocution of the unfortunate lineman Feeks occurred near the center of New York City’s government is significant. Apparently nothing galvanizes public officials into action faster than seeing a man roasted in the air above them. They quickly passed legislation requiring that all utility lines be buried deep in the ground. Naturally, the electric companies fought back, but the law was upheld and the overhead wires were cut down. (There was an almost comical side note to this part of the story, as corrupt and do-nothing city workers did, well, nothing for months, leaving large sections of New York dark during the subsequent winter.)


Ironically, after so many fireworks (sorry about that) the Great War of the Currents ended quietly. Westinghouse had been carefully purchasing companies with technology that solved some of AC's biggest problems. On the business side he acquired a meter that could accurately measure power consumption. Westinghouse also bought patents that improved both the efficiency and safety of AC systems, which quelled much of the resistance to AC. One of those patents was from a disgruntled former Edison employee who felt financially slighted by the Wizard. His name was Nikola Tesla, and he's worthy of his own blog. I'll put that on my list.


Distracted by his new interest in ore processing, Edison gradually released control of his power company to the company's board, which in 1889 approved a merger of Edison Electric with General Electric. At the same time, the Wizard lost control of his own company and most of his interest in the fight for DC. The new Edison General Electric Company spent the next few years buying up AC technology, which they were now free to pursue as the future of power distribution. This included buying their chief rival, Thomson-Houston, lock, stock, and transformer. In a sign of how disengaged Thomas Edison had become from power distribution, the story is that he didn't even hear about the purchase or that the new name, General Electric, would not include his own until after the sale went through. GE and Westinghouse would continue to snipe and sue each other, but would also be thrown together on projects such as the Niagara Falls Power generation and distribution system.


Direct current systems continued to be used in some cities. Helsinki distributed DC until the 1940s, Boston supplied it along stretches of Beacon Street into the 1960s, and remnants of Stockholm’s DC system survived until the 1970s. In New York City, where the War of the Currents began, it wasn't until 2007 that the last customers requiring DC (used mostly for elevators) were finally migrated to AC. As for "Old Sparky," after a decades-long national moratorium on the electric chair ended in 1977, only six states still use it as their primary form of execution, if the inmate chooses it over lethal injection.


It is my opinion, based on the historical record, that Thomas Edison absolutely believed in the superiority of DC over AC, until the safety and transmission issues of AC had been resolved. Ironically it is Edison— the DC proponent who watched dozens of animals (although not an elephant) get electrocuted to prove a point— and not Tesla— the man whose genius solved many of AC’s safety and transmission issues— that remains associated with the modern AC power grid.


A blog for another day. Time to turn out the lights on this one...

In my last several blogs I've been talking a lot about radiation testing.  At the beginning of the series of blogs I've been posting we looked at the various types of radiation testing that are performed for products going into space applications.  Recall that we looked at TID (Total Ionizing Dose) and SEE (Single Event Effects) testing where TID is a type of cumulative effect testing and SEE is more of a real time test.  We spent a lot of time specifically looking at the various types of SEEs - Single Event Latchup (SEL), Single Event Upset (SEU), Single Event Transient (SET), and Single Event Functional Interrupt (SEFI).  Upon going over the basics I applied that to the case of high speed ADCs and looked specifically at these single event effects are observed as well as how we look at TID with such devices.  This month I am taking a step back and looking at the why instead of the what.  Hop on over to Planet Analog and take a look at my latest blog: Why All This Stringent Testing?.


In my latest blog we look at the different sources of radiation as pictured above.  The three sources are solar radiation, cosmic rays, and trapped radiation around the earth.  Devices going into space have quite a harsh environment with all these sources of radiation as well as drastic temperature changes due to passing in and out of the path of the heat of the sun (satellite applications specifically).  I hope you'll take a look and learn a little more about space.  Feel free to leave a comment here or over on Planet Analog.  I'd love to hear your feedback. 


Also, if you are reading this today on 7/27/18 don't forget to check out the live streams of the lunar eclipse.  It is not viewable in North America, but you can catch it from various feeds from Europe:


2:30PM EST:


2:00PM EST


You can also find it on the Weather Channel app at 4:00PM EST.


Happy Viewing!!

It was April 21, 1916 and the people of Chicago were enjoying a beautiful Spring day, their city's grueling winter behind them and the unrelenting heat of summer a few months ahead. The Chicago Cubs baseball team had opened their first season at the two year-old Weegman Park the day before and, despite the Cubs' fifth-place finish the previous season, hopes ran high for a return to the World Series. Overall, 1916 was a grand year for a city intent on burnishing its broad-shouldered reputation. This would be the year a young Carl Sandburg published a poem in which he described his city as the "hog butcher for the world," something any visitor who ventured anywhere near the slaughterhouses during warm weather could have guessed from the odor that pervaded the country's second-largest city. That wasn't stopping the Republican Party from coming in less than two months to hold their quadrennial convention, when they would nominate the Supreme Court's Chief Justice, Charles Evans Hughes, as their candidate for the Presidency.


April 21st was also the day that Cook County Judge Richard Stanley Tuthill would rule on a case that had been argued before him over the previous two weeks. The ruling, although a much less well-known event than the baseball game at what would soon be called Wrigley Field, would grab headlines, not just in the Windy City but across the globe. Judge Tuthill had ruled that William Shakespeare was a fraud, and that philosopher, scientist and inventor Francis Bacon was the true author of all those plays, poems and sonnets.


Now let's be clear that although the ruling was covered in newspapers around the world, the publisher of "The Complete Works of Shakespeare" was not racing to change the title to "The Complete Works of Bacon." Nor was the Royal Shakespeare Company removing the Bard's name from their marquee. That's because what every member of the press corps knew - and wasn't shy to say in print - was that trial was, in fact, nothing but a publicity stunt conceived by two men; movie producer William Selig and millionaire George Fabyan.


Chicago Newspaper story on Trail

Chicago Tribune, April 22, 1916


Fabyan was a former Bostonian who had inherited a successful cotton goods manufacturing company. With the inheritance he bought large estate in Geneva, Illinois which he called Riverbank Labs. There he reveled in his financial ability, as he once said, for "...spending money to discover valuable things that universities can't afford." Those valuable things included research into such mainstream endeavors as acoustics and genetics. (The acoustics lab is still regarded today as one of the world's best.)  Other investments, such as the search for the elusive perpetual motion machine, were less regarded by mainstream scientists. Such scoffing seemed only to embolden Fabyan, who also funded the work of a fellow Bostonian named Elizabeth Wells Gallup. Mrs. Gallup had written extensively about her belief that a secret code had been placed in the First Folio, an early printing of the Bard's work. This code, when properly deciphered would prove Bacon's authorship of the writing attributed to Shakespeare.


However far-fetched the theory may have seemed to most people it was based on one universally accepted fact: that Francis Bacon had invented a way to hide secret messages inside any text. He called it a Biliteral Cipher, and it should look familiar to anyone who has ever worked with computers, specifically in binary code.


Biliteral Cipher

Bacon's Biliteral Cipher


Every letter of the alphabet, known as the plaintext, is represented by a group of five a or b (sorry, U and V, but you have to share) which make up a binary code.  The plaintext "A" is represented as aaaaa, the plaintext "B" would be aaaab, and so on.  Using this code Bacon demonstrated how easy it is to hide a secret message that is the exact opposite of the one the public would read.  Let's take the message STAY WHERE YOU ARE as an example.  Using Bacon's cipher one would write out the message so it looked like this:




Notice that some of the letters are bold and italicized.  Next we organize the letters into groups of five:




The last step is to use the table to convert the groups of plaintext into the deciphered message. So STAYW is hiding the code aabab (two letters in normal font, one in bold/italic, followed by a letter in normal font, ending with a bold/italic letter) which we convert to the letter "F." HEREY is ababa which converts to "L," and OUARE is babba which is "Y."  The decoded message? FLY, the exact opposite of the original plaintext message. Computer-savvy readers will also note the similarity between the Biliteral Cipher and ASCII code, although Bacon was two bits shy. (Nope. Not going to apologize for that.)


Now the notion that Bacon wrote Shakespeare had been around for decades. There was even a magazine for believers called Baconiana (which today can be read on the Francis Bacon Society's website.) Mrs. Gallup (who had attended the Sorbonne and was no intellectual slouch) fell sway to this theory, as did many prominent people of the day including Mark Twain, Walt Whitman, Sigmund Freud, Henry James, Henry Miller, and Helen Keller.  At Riverbank (and on Fabyan's dime) Gallup would pursue that truth - scientifically - by having a copy of Shakespeare's First Folio photographed and analyzed for passages whose plaintext hid Bacon's Biliteral cipher.


Mrs. Gallup was not a photographer, but as luck would have it Fabyan has just brought to Riverbank a bright young geneticist named William Friedman from Cornell.  Friedman had been hired to work on developing a strain of wheat that would grow in an arid environment, something Fabyan thought would come in handy in places such as Palestine. When Gallup learned that Friedman's hobby was photography, she drafted him onto her team. Friedman likely would have volunteered, anyway, as he had become smitten with a lovely young librarian who was working for Gallup. Her name was Elizebeth Smith (whose parents deliberately spelled her name with an "e" to set it apart.) What happened next is the kind of stuff you can't make up, because no one would believe the story. But the following is all true.


Snippet from The Tempest, First Folio

A snippet of The Tempest, from the First Folio


Gallup used enlargements of Friedman's photographs to measure and quantify the slant of each letter to determine which were "a" and which were "b" in Bacon's cipher. When mathematics failed to provide a definitive answer she let intuition rule, and from that process came forth decipherings such as "Bacon had writ these works" and other revelatory passages that proved Baconians right. While Friedman and Smith mooned over each other, Gallup collected the evidence and by early 1916 was ready to publish her great work exposing the fraud that was Shakespeare. That's when Fabyan was hit with the lawsuit filed by Selig, the producer, whose film company was about to release films based on four of Shakespeare's plays in celebration of the tercentenary of the Bard's death on April 23rd. The suit claimed that the reputation of Shakespeare "would be shattered" by Gallup's "alleged decipherings" and the public would not go to see his movies. Selig asked the court to rule that Shakespeare was the real author.


In early April, as the Cubs decamped from spring training in Florida and headed for their North Side ballpark's National League debut, Judge Tuthill gaveled the trial of William Shakespeare into session. Sadly the official transcript of the trial has disappeared from Cook County records. My father, who wrote extensively about the history of cryptology, made several efforts to find the transcript, but to no avail. In the early 1990s PBS sponsored a "Mock Trial Determine the Authorship of Shakespeare's Works," but no mention was made of the Biliteral Cipher which was the foundation of Mrs. Gallup's claim of authorship. In J. Ajlouny's play, The Trial of William Shakespeare (recently published by the Fresh Ink Group) the author has the audience play the role of the jury after hearing evidence presented in his courtroom re-creation. But little time is spent on the claim of a hidden cipher. In my similarly titled play, published by Eldridge Plays & Musicals, I included a trial scene with the Biliteral cipher presented as a key defense. (Somehow, Tony Award committee missed its one production.)  Ultimately, all any of us can do is surmise what was said in Judge Tuthill's courtroom. That's a real shame, because both Fabyan and Selig (two men both known for having dynamic personalities) choose to act as lead counsels for their side, meaning some great theatrics have likely been lost to history.


What has not been lost is the remarkable series of events that followed Judge Tuthill's ruling which, as we can see from the press clipping above, resulted in precisely what both Fabyan and Selig had set out from the start: gobs of free publicity for Selig's films. And many seemed to have had a good laugh over the idea that Chicago had legally declared Shakespeare a fraud. One wag, tongue firmly in cheek, asking if the ruling meant the city was not obligated to change the name of Shakespeare Avenue to Bacon Avenue. Other Chicagoans were not so sanguine, most notably the other Cook Country judges. On May 2, 1916, just ten days after Tuthill's verdict, they reversed his ruling, thus restoring Shakespeare to his literary perch.


So how do we get to Pearl Harbor from here? To tell the next part of the story we need to back up just a little. Friedman's doubts about Gallup's deciphering methods had been growing since he got pulled into her Bacon project. By the time of Tuithill's ruling, he was almost regretting accepting Fabyan's original offer to come to Riverbank. I say almost, because soon after the trial he married Gallup's lovely young assistant, Elizebeth Smith. Around that same time Fabyan made him the "Director of Riverbank's Department of Codes and Ciphers." Such was the genius of George Fabyan, who recognized that the Bacon skeptic possessed an extraordinarily sharp analytical mind that could be leveraged in the field of cryptology. Friedman would reward Fabyan by writing a renowned and, in some cases, groundbreaking series of papers that put Riverbank on the map.


The timing of the publication of Friedman's papers could not have been more perfect. In early 1917 a telegram (known as the Zimmerman Telegram after the German official who sent it) was decoded by British Intelligence.  It was an offer, by Germany, to give Mexico large portions of the southwestern United States if the U.S. were to declare war against Germany, which it did in April. Fabyan immediately offered the services of his Department of Codes and Ciphers to the War Department, which they gladly (and with no small amount of desperation) accepted. Almost immediately Friedman was challenged with another coded German message, this time to Indian activists agitating for independence from England. With that success William enlisted and went to France, where he worked directly for General Pershing. His decoding of intercepted German wireless messages is said to have resulted in the saving of thousands of Allied lives. After the war he continued writing, quite literally inventing the new field of cryptanalysis.


Before we continue with William's story, we first need to note that Elizebeth was no mathematical slouch. She lent her estimable decoding skills to the Navy and then to the Bureau of Investigation (as the FBI was known during Prohibition) to help smash rum-running rings operating off the coast of Florida and in Houston. Later, she would be instrumental in busting a narcotics ring operating out of San Francisco. Elizebeth was often called upon to testify at the resulting trials, a sometimes dangerous act given the systemic violence of smugglers. All this during an era in which women were not always welcome in the fields of math and the sciences. Or law enforcement, for that matter.


Now let's get back to William. Along with his groundbreaking writing on cryptographic theory, he spent a great deal of time studying the flaws in a new kind of code machine, an electro-mechanical device that used rotors to generate the cipher. In much the same way that companies seeking higher levels of cyber security will hire hackers, Friedman exposed the flaws in the machine's original design.  Then, in collaboration with Frank Rowlett, a mathematician he had hired in 1930, he went and created a version that was so good it became the United States' highest security cipher machine in World War II.


Rotor-based Code machine

Late 1930s version of a rotor-based code machine


Friedman would become the stuff of legend when, before the outbreak of hostilities between the United States and Japan, he and his team broke the Japanese cipher known as PURPLE by building a replica of the machine without ever having seen one. I'm going to say that again, because it is an extraordinary achievement to build a copy of something, in this case an extremely complex cipher device, using only intercepted messages to surmise how the device might have been built. How those decoded messages got used was another matter.  As the Wiki page on Friedman reports, "One such intercept was the message to the Japanese Embassy in Washington, D.C., ordering an end (on December 7, 1941) to negotiations with the US. The message gave a clear indication of impending war, and was to have been delivered to the US State Department only hours prior to the attack on Pearl Harbor. The controversy over whether the US had foreknowledge of the Pearl Harbor attack has roiled well into the 21st century."


Friedman's accomplishment was not without its costs.  As explained on his Wiki page, " 1941, Friedman was hospitalized with a "nervous breakdown," widely attributed to the mental strain of his work on PURPLE." Friedman recovered, and continued to contribute to the war effort. To further the Allies advancements in code-breaking his team would trade one of their replica PURPLE machines for details on how the British broke the German ENIGMA code.  (That effort, led by Alan Turing, would later be dramatized in the Oscar-winning film "The Imitation Game," starring Benedict Cumberbatch.  I include his picture here because Cumberbatch is my wife's favorite actor, so I figure maybe now she'll read one of my blogs.)


Benedict Cumberbatch as Alan Turing in The Imitation GameCourtesy of STUDIOCANAL UK

Benedict Cumberbatch as Alan Turing in The Imitation Game

(Courtesy of STUDIOCANAL UK)


After the war, Friedman became the chief cryptologist for a new government agency built for the Cold War: the National Security Agency, or NSA. While leading the agency's cryptologic efforts, he also continued to write textbooks, some of which are still used today to train young cryptologists. By 1956, when Friedman retired, the rotors and gears of the Enigma and PURPLE machines had been supplanted by vacuum tubes in the first mainframe computers. Tubes would later be retired by discrete transistors, and those by semiconductors.


Retirement?  William and Elizebeth didn't retire the way most people do. It had always bothered them how the Biliteral Cipher, which was the very first binary code (and the foundation of all modern computing) had been misused by Mrs. Gallup. With Elizebeth once again by his side, William Friedman returned to the First Folio and together they wrote a book called "The Shakespearean Ciphers Examined" which exposed the flaws in Gallup's methods. There is something very satisfying about this final chapter in their lives, one that reflects not only their great devotion not only to each other, but to the facts.


The Genius Who Failed - A lot

Posted by DavidHTO Employee Jun 26, 2018

When he arrived in Boston in the summer of 1868 he did not look like a man who would someday change the world.  Clutching all his worldly possessions in a canvas bag, he was the very definition of a "hayseed," a country boy who came to make his mark in the big city. So he passed, undistinguished, through the downtown crowds as he made his way to the local Western Union office, holding a letter of introduction from a friend back in Ohio. At the telegraph office he displayed the one skill that set him apart from the crowd, a seemingly natural gift to be able to receive and send Morse code very quickly. The telegraph system was the backbone of communications after the Civil War (author Tom Standage called it The Victorian Internet) and skilled operators were highly valued so, despite the office manager's misgivings about the appearance of this "country bumpkin" the young man was hired, although it was for the overnight "graveyard" shift. Still, it must have been an exciting time for him as he copied and sent messages from across the country.


Western Union office NYC post-Civil War

This Western Union office in New York was similar to the one our young Ohioan worked in Boston


The overnight shift actually suited the young man. He didn't sleep much, anyway. Too many ideas floating around in his head for devices that he wanted to build and sell to what he assumed would be a line of investors. No, our young telegraph operator didn't want to spend the rest of his life just "pounding brass," he wanted to be rich. So after his shift at the Western Union office he walked over to this building, at 109 Court Street, in a section of the city known then as Scollay Square (where Boston's City Hall and the JFK Federal Building are located today.)


109 Court Street

109 Court Street, Boston, MA


Charles Williams manufactured telegraph equipment here, but he also rented out space in the attic to about a half a dozen people. In every way this place was just like today's high-tech "incubators," where like-minded experimenters went to pursue their dreams of building the next great thing. Williams not only provided space, but also access to wire, motors, transformers, and other material needed for electronic work.


Our young telegrapher knew exactly what the world needed. Just about anybody involved with telegraphy did. In 1868 it was only possible to send one message down a single wire at one time which meant if we wanted to pass more messages, we had to string up more wires. Which we did, with abandon...


NYC 65th Street

65th Street in New York City in the 1880s


That's why the Holy Grail of inventions was the elusive Multiple (or Multiplex) Telegraph, a device that would permit more than one telegraph signal to pass through a single wire. Our young inventor thought he had a way to do that.  Working diligently in his off hours from the telegraph office he designed such an apparatus. But after repeated attempts he failed to get it to work. Frustrated, he put the idea aside and went to work on a completely different type of machine - an automatic vote counter. The device, which worked precisely as designed, earned him his very first patent in June of 1869. Now he was a real inventor! The next step was to find a buyer for the device, and through an investor he was able to secure a demonstration in Washington before several Congressmen. He was sure they would jump at the opportunity to use a device that would speed the legislative process.


Edison's Vote recorder

Automatic Vote Recorder


The actual response from the committee chairman?  "If there is any invention on earth that we don't want down here, that is it."


What our young inventor thought would be a welcome way to reduce the gridlock in Washington was seen more as a disruption of time-honored Congressional traditions such as vote-changing and filibustering. He returned to Boston not as much frustrated as motivated to build something else, only this time to make sure that it was something he could sell. Today we call that market research, back then it was the pure fire of his determination. Perhaps, also, to prove the teacher wrong who said to his mother "you might as well take the boy out school, he's addled."


By now you've probably guessed that the name of our young telegrapher/inventor was Thomas Edison. Now in Boston almost two years, he had failed twice to achieve his goal of being a successful inventor, the first time because the device (the multiple telegraph) didn't work, the second time (the automatic vote counter) because he couldn't sell it. For his next invention he would then display the knack that would lead to over 2,300 patents and a multi-million dollar fortune. He found a problem, in this case the need to synchronize multiple stock tickers in financial offices and banks all over the country (the currently deployed device, invented by Edward Calahan a few years earlier, required exchange companies to deploy workers to manually reset and calibrate the tickers).  For Edison, that meant thousands of ready-to-buy customers. All he needed was a working prototype.


He spent his last few months in Boston attempting to build one, but none of them worked. He had managed to find a few investors, but their patience ran out in late 1869 and young Edison, who had such faith in this new idea that he had quit his job at the telegraph office, was now as broke as when he arrived two years earlier. He moved to New York where his luck changed after meeting a telegraph executive and inventor named Franklin L. Pope, who happened to work for the company that owned the patent to the previous generation of stock ticker. Pope gave Edison some space at the company’s Wall Street headquarters to pursue his concept. What happened next is legendary. 


As explained in the Menlo Park Museum website, "...within a few days, the master ticker tape machine had a major breakdown throwing the entire office and many New York businesses into turmoil. Since Edison was always around the office, he offered to fix the machine. When he did so within two hours, he was offered a job the following day as Pope’s assistant."  Using his knowledge of the inner workings of telegraph devices (garnered from years as a telegraph operator) he not only designed a way to synchronize multiple tickers, but in a later model devised a way to print characters to represent shortened names of the company along with numbers for the current price. This was a huge deal, because up until this time the machines printed in Morse code, requiring employees with that skill, and the time to transcribe the data.


Stock Ticker

Thomas Edison's Stock Ticker with Alphanumeric print


As the Menlo Park website explains, "...after working for Dr. Laws (the company president) Edison set up his own engineering business and was soon hired by Western Union to be in charge of all of their equipment. Within a short span of time, his boss at Western Union offered to buy out all his new inventions and improvements to the equipment for a lump sum price of $40,000."


An invention that worked and money to show for his effort. What could be better? Edison used the funds to start his own incubator in an empty factory building in Newark, New Jersey. A few years later he would purchase property in the New Jersey town of Menlo Park where, thanks to monumental inventions such as the light bulb and phonograph, he would earn the public's adoration and the nickname "The Wizard of Menlo Park." This, despite his alleged ruthless, painful  electrocution of an elephant to prove a point. But that's a subject for my next blog.


Oh, by the way, Edison was not the only inventor to pursue his dreams in that attic laboratory at 109 Court Street.  Look again at the picture at the top of this blog. That's a recreation of the laboratory of a Boston University Professor of the Deaf, who had his own ideas on how to solve the problem that eluded Edison, that of multiple telegraphy. This professor (along with his assistant, named Watson) got sidetracked when he realized that he could send more than just dots and dashes. Because it was here in 1875, in the exact same place where Edison had built his first patented invention, that Alexander Graham Bell first heard the sound of the human voice through an electric device.


As Paul Harvey would have said, "and now you know the rest of the story..."

Hello once again!  I hope that you have been enjoying my recent series on single event effects with high speed ADCs.  So far we have looked at TID (Total Ionizing Dose), SEL (Single Event Latchup), and SET (Single Event Transient).  This month the focus is on SEU (Single Event Upset).  This is quite closely related to the SETs we looked at over the last two months.  It is possible that a SEU could result in an SET depending on the ion strike and the resultant behavior of the circuit.  For the purposes here though we will look at them separately to help provide a bit of clarity into the underlying cause.  For a high speed ADC we will consider an SEU as a configuration register upset.  The AD9246S is the device under test.  As it turned out there were no configuration register upsets observed for the AD9246S but in my Planet Analog blog I discuss what an SEU might look like.  A brief idea is that an ion strike results in a configuration bit(s) upsetting.  It is expected that the bit(s) would not permanently flip but would revert back to the expected value.

SEU - Single Bit

Configuration Register SEU - Single Bit

SEU - Multiple Bits

Configuration Register SEU - Multiple Bit


Recall that it is important to perform this testing so that the device behavior in space can be predicted.  The Weibull fit curve is generated from the data and then input into a model factoring in the expected orbit to generate a probability of device upset.

Example Weibull Fit Curve

Example Weibull Fit Curve - AD9246S


It is imperative to perform the testing to generate this data and predict the device performance in space since, as we all are well aware, it is not very easy, dare I say impossible in many cases, to repair/replace devices once they are put into an application in space.  Obviously there are things like the ISS (International Space Station) that are routinely visited and repaired, but something like Juno which is orbiting Jupiter that would not be able to be repaired nor worth the cost required to attempt to do so. If you'd like to learn more I encourage you to check out my blog over on Planet Analog.  I hope you have learned a lot so far and found SEEs interesting.

Have you ever listened to AM radio at night and wondered how you could hear stations from great distances away, but not hear them during the day? So have scientists and researchers, who have spent decades trying to understand what's going on above us to cause radio signals to "skip" like that at night. That led to more questions, such as why a signal sent from the same location at the same time of the day - but at a different frequency - skips farther or shorter distances? Or why signals from the same location at the same time of day and at the same frequency - but sent during a different season of the year - can behave differently? Or how in some years, even with all those conditions (location, time of day, frequency, and season) the same the signal travels great distances, but a few years later will barely make it to the border? With all of these variables at play, are we able to predict the behavior of the ionosphere and its effect on communication?


Sure we can.


But no, we really can't.


Okay, so first a quick bit of history to put these seemingly conflicting answers in context. Humans have long observed - and early on exploited for navigation - the fact that a magnetized object which is allowed move freely will point north. In 1839 a scientist named Frederick Gauss postulated that there was region of space above the earth that was electrically charged. For the next several decades a lot more postulating, theorizing and experimentation went on that lead to the development of an early form of wireless communication called spark gap. The name is very descriptive. You know how sometimes you will touch someone and you both receive a small electric shock? That's basically spark gap. When early experimenters (such as Guglielmo Marconi) figured out how to detect the spark from a distance, radio communication (using a variation of Samuel Morse's code) was born.


Among its many disadvantages was that spark gap required massive amounts of power to work over any appreciable distance, and was limited to a small range of frequencies, near today's AM band. What happened next reminds me of lining up dominoes and watching them fall against each other, because the timing could not have been better for the advancement of radio and our understanding of the ionosphere. The first domino is the passing of the Radio Act of 1912. Since wireless "state of the art" in 1912 was spark gap, which produced signals below 1200 KHz (basically the middle of today's AM radio band) the Act graciously handed experimenters and ham radio operators - what Congress blithely considered to be a useless part of the radio spectrum, above 1500 KHz.


Spark Gap Transmitter at Awanui, New Zealand

Spark Gap transmitter at the Awanui, New Zealand station.  Note the blowers in the bottom right of this photo.  These high-power stations got hot.  Very hot.


Hear that sound?  That's another domino falling.


Just before the First World War the pioneering work with vacuum tubes by Lee DeForest, Walter Schottky, and others enabled the design of transmitters that produced a signal that was more than just a hash of static, it was a tone called Continuous Wave, or CW. Amateur experimenters who could not afford the extremely powerful transmitters and massive antennas of commercial and military spark gap stations found they could make CW contacts with much less power and smaller antennas. These experimenters also began to notice patterns in their ability to reach certain distances based on some of the variables mentioned earlier such as time of day and frequency. Leveraging that experience, hams were soon making record-setting connections between the Eastern seaboard of the United States and Europe, and the West Coast and Hawaii by radio. Now get ready for another domino to fall. In the beginning most hams were crowding close to that 1500 KHz Congressionally-mandated wall. It got very crowded down there very quickly so, leveraging the ever-improving frequency range of the new tubes, higher frequency transmitters and receivers were built.


Now a seriously important domino was about to fall because in the early 1920s, as experimenters moved up in frequency they found longer distance communication was easier at certain times of the day. As more and more data was collected (hams are notoriously good record-keepers, making this like a global lab experiment) the physicists put the pieces together and, in 1923, concluded that the ionosphere - postulated 87 years earlier by Gauss - was responsible.  Now a whole line of dominoes began to fall as several things happened. First of all radio, considered to be a toy in its early days, was quickly becoming commercially viable. Hundreds of local stations were being built around the country, their owners happy to sell airtime to local hardware and feed stores, evangelists, and hawkers of patent-medicines, all of whom were happy to pay to have the ears of thousands of potential converts and customers. Listener-ship skyrocketed with many manufacturers getting into the radio receiver market, including the purveyors of cheap crystal sets, which allowed folks whose homes did not have electricity to hear those broadcasts.


Crystal Set

No electricity?  No problem with a crystal set!


The government quickly realized that this new industry required oversight by an organization dedicated to taming this new "Wild West," so in 1926 it formed the Federal Radio Commission. A year later the Radio Act of 1927 was passed which, among other actions, rescinded the Radio Act of 1912. Now fall more dominoes, as those frequencies above 1500 KHz that the government had tossed away as "useless" back in 1912 were now being allocated to commercial, amateur, military and other government services. This put thousands more technicians and experimenters up and down the frequency spectrum, working with more efficient transmitters and better performing tubes, collecting tons of data on radio reception. In 1926, building on the work of theorists and physicists such as Oliver Heaviside, Arthur Kennelly, and others physicist Robert Watson-Watt first used the term "ionosphere" to describe that place where radio signals propagate or "skip" (as it was more popularly known by excited radio listeners) in the atmosphere.


What they found was that a radio station operating at the right frequency and at the right time of day could be heard thousands of miles away - even across oceans. This was not lost on some governments in Europe, where tensions had been rising since the end of the First World War. Germany and England, two nations at the forefront of wireless research, built stations that could be heard here in America, allowing them to present their views unfiltered to the world - not unlike the way Twitter and other social media are used today.  Being market-driven, it wasn't long before manufacturers of consumer radios included one or more shortwave bands in their radios. This meant that many Americans who were listening to Franklin Roosevelt's "Fireside Chats" broadcast here in the States were also able to hear Hitler, Churchill, and other foreign leaders via shortwave, thanks to the skipping of signals in the ionosphere.


 Zenith Model 6-S-222 "Cube" Radio from 1936

 Zenith Model 6-S-222 "Cube" Radio from 1936, with three radio bands


Ionospheric skip took on even greater geopolitical importance after the Second World War, as both sides in the Cold War built mammoth shortwave stations, pumping out up to a million watts of power to extol the virtues of their political system and the disadvantages of the other. Yet our understanding of the ionosphere, especially our ability to predict the behavior of signals up there, was not advancing fast enough for those tasked with the development of rockets, whether they were being designed to deliver nuclear payloads to the enemy or humans into orbit. But it wasn't just the ionosphere that was of concern. Based on observations made as far back as the early 20th century, scientists were speculating that there was a region of charged energized particles beyond the ionosphere. In January 1958 experiments aboard Explorer 1 (the first United States satellite to reach orbit) proved the existence of what became known as the Van Allen radiation belt.


Very early observations of the Van Allen belt confirmed what had been speculated by scientists, that from 93 million miles away the sun was producing a "solar wind" of electrically charged particles that pushed against the belt at various times and at varying strengths. While there was some predictability to it (such as an 11 year cycle of sunspots) the sudden appearance of solar flares were very unpredictable. So, for those of you keeping score, between the discovery of the radiation belt and the sun's impact there were dozens more variables added to an already long list (frequency, time of day, time of year, etc...) that need to plugged into any equation for ionosphere skip prediction. And what an equation it is! (Don't worry, I won't give you any more math than I can handle... which is why I am giving you none.)


At the very beginning of this piece I asked if we could predict the effect of the ionosphere on communication. My answer back then was: "Sure we can." And then I said "But no, we really can't." I hope you can see that I was not being deliberately evasive. The fact is that even today, after over 180 years of theorizing and testing and collecting data (from both terrestrial and space-based observations) physicists and mathematicians kind of throw up their hands when asked for a definitive predictive model for what's going on up there by giving us this word:


stochastic: /stəˈkastik/ adjective

Derived from the Greek, the word means "randomly determined; having a random probability distribution or pattern that may be analyzed statistically but may not be predicted precisely." What that means for us is that instead of using lined up and orderly dominoes for my ionospheric analogy I should have employed the game of Craps, because Einstein was wrong when he said God does not play dice with the universe. Not only does He play dice, but He's constantly shooting for a hard eight.


I know this from personal experience as a ham radio operator, and now that you have all the background, I want to tell you about the time I cashed in on the long odds. To start, hams are always looking to improve their chances of making contacts with other hams in far-off places (in ham parlance, we call that DX) and there are a number of online tools which provide predictive reports on ionosphere skip. Here's an example of one such report, posted on the American Radio Relay League's website, which shows what happens when all those variables we've been talking about get plugged into a tool that charts predicted activity between any two parts of the world, in this example the East Coast of the United States and Eastern Europe: 


East Coast to Europe Propagation Forecast

Source: American Radio Relay League


The vertical axis is frequency in MHz (for reference, an AM station at 1030 would be down at 1.3) and the horizontal axis is any 24-hour period at UTC (what used to be called Greenwich Mean Time) during the month for this chart. Now, here is why we have to swap our dominoes for dice, because the three different colored lines represent only the probability that at a given time and frequency two hams could make a connection between those two places on the planet. Red represents a 10% chance, green is a 50% chance and blue is your safe bet (kind of like using house money) with a better than 50% probability that those two hams could hear each other at that time and frequency. 


The chart says that If I tuned my radio to 14 MHz at 0800, I'd have a 50/50 chance to connect with a ham in Prague. If, at the same time, I transmit at 10 MHz I'd have a better than even chance. That drops to only a one in ten chance were I to be on 18 MHz. So 18 MHz at 0800 is not a good place to try to make that contact. But we also see (by following the green line) that the chances for a contact at 18 MHz improves as it gets later in the day. Come back in six months, when the northern hemisphere is tilted differently towards the sun and the chart will be different. Come back in six and half years (during the other half of the sunspot cycle) and the chart will look very different, as well.


So thank you Gauss, Heaviside, Van Allen and others for all your pioneering work. Thank you to the folks who collect data from satellites, radio beacons, solar observatories and other sources to create this and other reports. But you know what?  While they are great guides for those of us who surf the ionosphere, I have personal experience with the predictably unpredictable nature of skip and how, sometimes, even when all the charts and tables and reports say otherwise, you gotta go for the hard eight.


In 2010 the sun was at a low point in the sunspot cycle, meaning that the ionosphere was not conducive to skipping signals, especially up on the 28-MHz band. The data said that at 0200 UTC (9:00 pm EDT here in Massachusetts) there was a very low probability for making a contact with anyone more than a few miles away. After being on my radio I was ready to concede that as fact but, before I shut the radio down, decided to tune across the band just in case. Suddenly, I caught the sound of a voice, weak at first, but quickly getting stronger. As weak as his signal was, I could tell he was calling CQ - that's ham shorthand for "can anyone hear me?"  Within 20 seconds it was as if he were in the room next to me.


"This is E51JD in the North Cook Islands calling CQ and listening for a call."


Wait. What? No, that's impossible. The North Cook Islands are (...type a quick lookup on Google Maps...) over 7,000 miles away.  And the charts and tables say that... okay, just in case... "E51JD this is WB2HTO in Massachusetts, do you copy?"


North Cook Islands


"I do and good afternoon, WB2HTO this is E51JD on North Cook Island in the South Pacific. Name here is Jim..."


So I began a conversation with a location that the charts and tables said I should not be able to have. It was a quick exchange of basic information (signal reports, names, and a few other details) because just as his signal had been growing in strength it was now starting to fade.  So we quickly signed off while we could still hear each other. As soon as we were done I could hear other stations trying to reach Jim, but whatever was happening up there in the ionosphere was changing so rapidly that I'm not sure if anyone else made a contact. I had been listening at just the right time and on just the right frequency to take advantage of what the physicists call a sporadic E opening, the "E" referring to a layer of the ionosphere that sometimes acts... stochastic.


Since, like Forrest Gump's box of chocolates, you never know what you're going to get (or how long it will last) when the E layer starts acting up like this, I moved up in frequency (to give Jim a clear space to make other contacts) flipped on the microphone and called out a "CQ" looking to see if any stations would reply. No sooner did I release the mic button that I heard "WB2HTO this is VK2HOT in Port Macquarie Australia, how copy?"


This... is... impossible. Port Macquarie Australia (another quick Google search) is 2,000 miles farther away than the North Cook Islands.  And I'm talking to someone there. At night. On a band that is supposed to be dead.


Port MacQuarrie

The distance from Port Macquarie, Australia to Boston is over 9,000 miles


A quick exchange of basic information followed and the exchange ended while we could still hear each other.  Meanwhile the band continued to shift. I only had two data points - the North Cook Islands and Eastern Australia - but it looked like the skip from Boston was moving westward. You can bet I tried calling for another contact, but after several tries it was clear that whatever magic had happened up there in the ionosphere was over. Or perhaps there just wasn't a ham in the right place at the right moment to hear me.


Several weeks later I got something in the mail from my new friend. It was a QSL, or confirmation, card. They are mostly sent electronically today, but I'm glad to have this hanging in my radio shack to remind me that the ionosphere is predictably unpredictable. Or, perhaps, unpredictably predictable.



Confirmation, or QSL card, from the contact that the numbers said should not have occurred

If you've been reading my blog for the last several months you know I have been discussing a lot of radiation effects and have given an example with these effects on a high speed ADC.  Last month I started talking about single event transients (SETs).  I felt like the subject did not quite get fully covered so this month read along as I spend a bit more time on the topic before we move on to single event upsets (SETs).  I realize this topic is most important to those working with the space industry and is not often thought about with commercial products.  However, this could be shifting as process technologies shrink and devices become more susceptible to low level radiation that makes it through the Earth's atmosphere.  In addition to that there are companies like SpaceX who are working to put more COTS products into you commercial guys may not be off the hook! 


I encourage you to take a look at my latest blog continuing the discussion on SETs over at Planet Analog here: Planet Analog - Jonathan Harris - Single Event Effects (SEEs) with High Speed ADCs: Single Event Transient (SET), Part 2. Here is an example of what an SET for a high speed ADC might look like:


AD9246 SET Run 64


If you have not read any of this series of blogs yet hop on over to Planet Analog and check them out.  You can find all my blogs here: Planet Analog - Jonathan Harris - Latest Content. The radiation effects series started in my December 2017 blog. There are also lots of other topics that I've discussed on my blog at Planet Analog.  Feel free to add comments are questions there or here on EngineerZone.

Working as I do here for the global leader of high-performance signal processing solutions, I often speak or write about Analog Devices and how our semiconductor products represent the "state of the art."  The challenge for ADI (in fact, for any company seeking a technology advantage) is to advance that state of the art and make it commercially viable.  As I've learned in my hobby of antique radio restoration, the same was true in the early days of mass-market consumer radio.  The technology might be 80 to 90 years old but, as I'll show you now, we can see great examples of how circumstance and ingenuity led some companies to find - or miss - that sweet spot.


From its founding in 1892 as the Helios Electric Company (manufacturing carbon-arc lamps) then, in 1906, as the Philadelphia Storage Battery Company (making batteries for electric vehicles) and finally, in 1919, adopting the name by which we know it today, the Philco Company (when it turned its attention to providing consumers with storage batteries for the burgeoning radio industry), the company was always looking for technology that could meet the needs of the market.  For example, because those early batteries were expensive and messy (they required the monitoring and refilling of liquids) in 1925 Philco would introduce the first "Battery Eliminators."  Today we call them power supplies.  Back then it was a boon for those home owners who had electricity because now they could run their radios directly from a wall socket.  (Just a quick side note: many in rural areas did not have power in their homes and still required batteries to run their radios.  Philco continued to market their batteries to serve that market and maintain their brand so that when New Deal projects such as the TVA brought power to farms across the South, Philco was the preferred brand when home owners upgraded.)


In 1926 Philco decided to get on the broadcasting bandwagon, but it took almost three years before they released their first radio.  That's because they saw how radio manufacturers, such as Atwater-Kent, were individually (and expensively) assembling their products and, as a result, having to charge high prices, limiting the market and, by extension, the radio audience.  The planners at Philco saw how Henry Ford was dramatically reducing the cost of manufacturing with assembly lines, which they co-opted for the mass production of their radios.  According to Wikipedia, only a couple of years after introducing their first mass-produced radio in 1928, Philco was already the leading maker in the country, grossing $34 million with the sale of over 600,000 radios.  For the mass market these included the well-made, beautiful, yet inexpensive "Cathedral" model, and for customers who liked cutting edge tech they developed and sold an innovative, one-tube wireless remote, which they marketed as the "Mystery Controller."


Philco Mystery Front

The Philco Mystery Controller was the first mass-market remote


Then, came the war...


Following U.S. entry into World War Two in late 1941 the economy, still moribund due to the Great Depression, was jolted into a frenzy of war-related production.  Philco was among the top 100 providers to the government for the next four years, as they pushed aside the design and manufacture of just about any product that was not war-related.  In 1945 the war ended, and as restraints on commercial production were lifted manufacturers, Philco among them, looked for products they could quickly produce and sell to a country anxious to enjoy the benefits of peacetime.  Both of the following radios are Philco model 46-250, meaning that they were produced that first full year after the end of hostilities.  They look the same, don't they?


Philco 46-250 Bakelite

From 1946, two Philco model 46-250 radios (author's collection)


Both radios are housed in Bakelite cases with only two knobs; one was a combined On/Off Switch & volume control, the other for tuning up and down the AM radio band here in the U.S.  Inside, both have an almost identical "All American Five" design found in most tube radios built from the 1930s onward.  AA5 radios eliminated the big, bulky power transformers prevalent in the large consoles that were popular before the war, which greatly reduced the cost of manufacturing and owning a radio, something especially important during the Depression.  The "trick" to an AA5 was, first, to reduce the number of tubes needed to the bare minimum.  I'll cover this in a future blog, but for now it's important to know that in the 1920s and 1930s our improved knowledge of tubes (and what caused them to distort a signal) allowed designers to build tubes that combined two or more stages of the receiver into one tube, which helped to bring the tube count down to five.  That typically included an RF converter, IF amplifier, audio detector/first amplifier, audio output, and rectifier which enable the most cost-efficient, best sounding radio.* 


Was there a drawback to the AA5?  Well, there was that nagging problem of electric shock.  That's because one side of the power line was connected to the metal chassis, so you didn't want to touch it when the radio was plugged in and turned on.  And who wouldn't want an appliance like that for the wife and kids?  (Full disclosure: I received a couple of lessons on AA5 power supplies the hard way.  Hurt like hell, too.  You'd think I'd have learned after the first time.  Or the second. Or the... never mind.  That was one of the selling points of Bakelite, by the way - it's a great insulator.)


Okay, so now let's talk about these specific radios and how they play into the story of changing the "state of the art."  As stated before both are the same model number (46-250) but, as we see from the stickers affixed to the base of each radio, there is a difference in the code numbers.  These numbers are also called chassis numbers, which delineate versions of the basic circuit employed in each radio:


Philco Inside side-by-side

Product stickers for the two radios


The primary difference between the two chassis is that the 122 on the left uses a mix of "Octal" and "Loctal" tubes, while the 125 on the right uses Loctals and what is called a miniature tube.  Octals, as you can see above, have a black base that was made with Bakelite.  They had thick pins that fit snugly into metallic sleeves arranged circularly in a Bakelite socket that was mounted on the chassis.  Octals were the workhorse of radios for decades.  Loctals were a relatively new type of tube developed by Sylvania in the very late 1930s for use in automobiles.  They had an aluminum alloy base that, as the name implies, locked into place in the socket.  As detailed in a Wikipedia page on tube sockets, Loctals had the advantage of being "pin-for-pin" compatible with the older Octal tubes.  Interesting to note that pin-for-pin compatibility is a selling point still used today in the field of semiconductor manufacturing.


Loctals are in both versions of the 46-250 we're talking about today, taking the roles of RF converter, IF amplifier, and audio detector/first amplifier stage.  But chassis 122 on our left still has two Octal tubes: a 35Z5 for the rectifier and a 50L6 for the audio amp.  This tells us that the 122 is an older model, as it is well-known that manufacturers such as Philco did not want to throw away their stock of older components.  But, as the three loctal tubes indicate, Philco's stock of old Bakelite tubes for the first three stages must have reached a point where it was financially viable to use the newer loctal technology.


As it turned out, the move to Loctal tubes is a great example of incorporating "state of the art" that sometimes fails to live up to the promise.  The tubes, according to Wikipedia, were "...prone to intermittent connections caused by the build-up of electrolytic corrosion."  And if you tried to take advantage of the pin-for-pin compatibility, you found the smaller pins of the Loctal tubes in the bigger Octal sockets had a tendency to "wobble."  So much for an upgrade.  Those problems would not surface for a few years, and Philco would use loctals in many of the home tabletop radios they built after the war.


So loctals have kind of a sketchy history, and demonstrate one challenge of developing new technology that may, at first, be considered "state of the art" but later turns out to have many flaws that diminish its marketability.  However, we will now see how the story of the audio output tube will, no pun intended, light the way forward not just for radios but all electronics.  We turn again to Wikipedia, which explains that "in 1938 a technique was developed to use an all-glass construction with the pins fused in the glass base of the envelope. This was used in the design of a much smaller tube outline, known as the miniature tube..."  The advantages of the mini tube are echoed today in the semiconductor industry where companies (such as Analog Devices) build and market products that use less power and therefore dissipate less heat (in the case of the mini tube, because the filaments were that much smaller.)


Despite the reduction in size the mini tube outperformed its predecessors and so for the next version of model 46-250, Philco designers tweaked the audio output section of the radio to accommodate a next-generation mini tube, the 50B5.  Clearly the 50B5 presented enough of a cost benefit because, as you can see in this close-up of the chassis where the Octal 50L6 previously sat, Philco installed a mini tube socket in its place. One can see that an Octal-sized hole had already been punched into the 125 chassis, requiring the riveting of the smaller mini socket inside:


Mini Socket adapter

Mini Socket adapter the Audio Output tube


Lower power dissipation and smaller tube size meant that Philco, along every other radio manufacturer who wanted to stay competitive, would abandon Bakelite cases and start housing their radios in cheaper, lighter, and less expensive plastics developed during the war.  They were not as resistant to heat, but didn't have to be.  State of the art and commerce often walk hand-in-hand, which they did in the case of Bakelite vs. Plastic and octals vs. loctals vs mini tubes.


Me?  I have a few colored plastic pieces in my collection, but I like the old school Bakelites with the bigger, hotter octal tubes.  As one of my ADI colleagues is fond of saying, "real radios glow."  You can hear the 46-250 125 chassis in action in this YouTube video.  





* The AA5 design worked (shock potential aside) because the filaments of the radio's five tubes were connected in series and the voltage drops across the tubes came close to line voltage (a number that changed over the years - there's another blog for another day, for now let's just settle on 120V as the voltage drop goal).  Since the leading number in a tube designation is the voltage drop across that tube, we can see the Philco 46-250 adds up to 106V.  Philco added a 2 watt 80 ohm resistor to the series of tube filaments to drop the additional 9V.  And yes, that meant that inside the cabinet not only did we have heat dissipating from the tubes, but also off that resistor.  That's why radio backs had large openings, to provide air for cooling.  


Philco PS with 80 ohm resistor

Close-up of the Philco 46-250 power supply showing the 80 ohm resistor

We had our most successful year as a group of ADI teams at the World Championship events in Houston and Detroit. It was record-setting, both for individual teams and for the program! We had a total of eight teams attend Championship and ADI hosted a booth again in the Robot Service Center this year at both events, where we got to talk to teams from all over the globe about what we do and what we have in store for teams next year.


ADI had four teams in Houston this year spread across three subdivisions. We were lucky to have the Robot Service Center right behind the Turing Division field, where Team 2655 Flying Platypi and Team 2471 Mean Machine were competing. One field down from us was 254 in the Hopper Division, and another two down from them Team 1577 Steampunk was competing in Galileo. I have to admit I was particularly interested in watching the Turing division teams, not just because my team was there but because... yeah, no, it was because my team was there. But what made Platypi's situation so unique was that we had zero control over our own destiny because of the robot's design. We couldn't climb, and we couldn't touch the scale. Winning or losing a match was up to our alliance partners. The only ranking point we could control was the Auto Quest. But we knew our robot could fill a niche that very few others could, and we were praying for other teams to see that value we could bring to an alliance.


After finishing 55th in the Turing Division, 2655 was selected to be part of the 8th seed alliance by our friends 1533 Triple Strange, another Greensboro team whom we've worked with nearly enough times to earn a spot in the "most successful team-ups" list on The Blue Alliance. It was going to be a long road clawing through this division as the number 8 seed, particularly for the first set of matches. To get out of the quarterfinals, we had to win against the top-seeded alliance, who was strongly favored to win. And for me, this was so difficult because this meant playing against 2471 who is also supported by ADI, and I've come to know this team as great friends in this community.


To be honest, I actually completely missed the first match. I came running up the stairs to the stands as the crowd erupted at the match score. And to my shock, the 8th seed had upset the 1st seed. As I emerged at the top of the stairs and looked to where 2655 was sitting with 1533, they were going crazy. And I looked closer to see that members from every North Carolina team were sitting with them, all cheering. I ran over to sit with the team while we waited for the next match. One field over, I could see 254 competing on the Hopper field, partnered with 148. This was an all-star alliance, the favored champions. I watched as 148 distracted teams so that 254 could load the scale with no interruptions. With the reality of 2655 advancing to Einstein looking more and more likely, the prospect of playing against this alliance became more and more real. On the one hand, I knew that realistically we probably weren't going to make it to the Einstein finals. But what if we did!?!


So match 2 between the 8th seed and the 1st seed comes up, and it was SO painful to watch. Alliance partner 1296 died in the middle of the match, 2655 threw a chain and couldn't drive. The 1st seed won the game with ease. I stood up and ran down the stairs to the pits with one of our other head coaches to see what was wrong. One of our drive team members ran up to the rail and explained that one of the drive chains had busted and fallen completely off. After a nail-biting and excruciating seven minutes, the robot was fixed just in time to head onto the field for the rematch.


And would you believe it, the 8th seed won it. I watched from the stands as the 8th seed alliance went on this Cinderella-esque journey to the Turing Division finals. I hardly remember each individual match and who did what, I just remember shouting, cheering, losing my voice, and then entirely losing all resolve when at the end of the trial we made it. We made it to Einstein. I shouted with what little voice I still had, I cried right along with the students (so, so much ugly crying, it was kind of pathetic). Hugs were going around everywhere. It was such a beautiful moment to see all of the North Carolina teams celebrate the success of these two teams as one, to watch these students I've worked with for so many long months achieve something they never thought they could do. At that point, I didn't care about any matches past that. We made it to Einstein and won our first Championship award for the team's fantastic business plan. I couldn't have asked for more.


At this point, I had no voice left. Zip, zero, zilch. I watched as our battered robots struggled on the Einstein stage, conversing via text with the drive and pit crew down on the field. When it became clear that we were locked out of Einstein finals, I looked up at the screen to look at who our final match was against. And surprise, it was against the "Black and Blue" alliance from Hopper. I only had one thing to say to the drive team before their match:


"Give them the best defense they've ever seen, and go have fun. No, actually, just go have fun out there. I don't even care about the score. It's your last regular season match, just have fun with it."


The match score was actually pretty sad. FMS sent our robot the wrong information for autonomous, so we scored on the wrong side of the switch and never fixed the switch position. The match score was left at 15 for the majority of the match for our alliance. At first, I was frustrated. Then the most beautiful thing happened, and I just about fell out of my seat.




It wasn't even planned. They just did it. Teams 148 and 1296 just decided to spin out of control because they had literally nothing else to do for the rest of the match. It was by far the funniest thing I've ever seen on an FRC field. I don't think I'll ever be that excited and tickled to lose a match ever again. The only way the season could have ended better was for us to have operational robots and defeat that alliance. But I almost prefer the way it ended because we didn't play to win the match, we played to have fun. In my opinion, that's more important than winning.


The Black and Blue alliance went on to the Einstein finals in Minute Maid Park, and they swept the finals to win the event. The Cheesy Poofs became the first team in FRC history to go an entire regular season undefeated.


Tune in next week for a wrap-up on the Detroit Championship and a look inside the Robot Service Center Booth!



This blog is part of a series covering the 2018 season of the FIRST Robotics Competition, FIRST POWER UP. Stay tuned for more updates, including coverage of the Championship Events in Houston and Detroit at the end of April! Get to know the ADI teams, learn more about our donation boards, and meet the employee mentors that make it all happen!

I’m back again to talk about my hobby of ham radio. Last time I wrote about the experimental antenna I recently built on my property for operating on the low (1.8 – 2.0 MHz) Amateur Radio bands. Today I’m going to go back to my early days in the hobby when I first got to know OSCAR. No, not the Grouch or Felix’s roommate, this was the acronym for the Orbiting Satellite Carrying Amateur Radio. 


Hams have always been at the forefront of communication technology, starting with the first wireless stations which sent and received Morse code. Hams would later experiment with some of the first AM, FM, and even television stations, so it’s no surprise that they were early in communication utilizing space. In fact, the Space Race was only four years old when OSCAR 1 was launched on December 12, 1961. OSCAR 1 was a very humble beginning to ham radio in space, as it was literally used as ballast to balance the payload of the rocket, although it holds the distinction of being the first satellite to be ejected as a secondary payload and enter a separate orbit. According to the OSCAR Wikipedia page, “…the satellite carried no on-board propulsion and the orbit decayed quickly. Despite being in orbit for only 22 days, OSCAR 1 was an immediate success with over 570 amateur radio operators in 28 countries forwarding observations to Project OSCAR.”


In 1974 after a series of successful launches of increasingly sophisticated satellites, Oscar 7 was launched, and that was followed fours later by the launch of Oscar 8. Both satellites were in low earth orbit (about 115 – 120 miles up) which meant they would only be in range at certain times and for a short period of time, but could both send and receive Morse code signals, permitting actual conversations. Both satellites acted as “repeaters,” receiving signals (the uplink) on the 2-meter band (144 – 148 MHz) and transmitting on the ten-meter band (28.000 – 29.000.)  As a youngster who grew up thrilled by the exploits of Shepard, Grissom, and Glenn it was exciting to think that in some small way I could experience the Space Program first-hand. Using a polar azimuth projection of the earth and a plastic overlay of Oscar 7 orbits (which had been included in an issue of QST, a ham radio magazine) it was easy to determine when the satellite would pass within range.


Azimuthal projection

From an edition of QST magazine.  Plastic overlays (with either Oscar 7 or 8 orbital paths) sat over this map


I set my transceiver (the rig was a Tempo One) to the down-link frequency on 10 meters and, when the timing was right, I could actually hear a number of CQs and QSOs coming from space. That was pretty cool. But I knew that the real fun would begin when I transmitted through the satellite, and so in the spring of 1977 I bought a used Hallicrafters VHF1 “Seneca” 2 meter transmitter, then constructed a 2-meter ground plane out of coat hangers, which I hung outside the shack window. Nothing fancy, as I had recently purchased a used TH3 beam for mounting on my parent's roof (these were GREAT parents) and knew eventually I would stack some sort of antenna for 144 - 148 MHz up there. Once I had the 2-meter transmitter working, I checked my QST Magazine for the satellite's latest schedule and, with my high-tech plastic overlay, saw that Oscar 7 would pass right over New York on the afternoon of June 6th, 1977.



A rendering of OSCAR 7 


I started hearing CW signals from other stations a few minutes after 1:00 pm (EDT) on 10 meters and began sending CQ (the internationally recognized abbreviation for “calling any station”) on 2 meters (remember that OSCAR was a 2 meter receive/10 meter transmit repeater). That’s when I heard… nothing. Huh. Well, since I didn't have a directional antenna I kind of figured the satellite, which was orbiting about 115 miles above, would have to be a lot closer to pick me up. So I tried again and - BINGO! - there it was, coming back to me on 10 meters, my own signal! Now here's the really cool part: as I'm transmitting and the satellite is racing towards me at 17,000 miles an hour, I could hear the tone of the signal getting higher in pitch - the Doppler shift of my signal coming back from space. It got stronger too, as Oscar 7 passed almost directly overhead, then almost immediately the tones began to shift lower as the bird raced away and the opposite Doppler shift took effect. Well, I just thought that was the coolest thing to be able to actually hear that happening. As of this blog, the experience is as vivid as any ham radio memory I have. 


Repeated attempts at connecting with another station were unsuccessful, not surprising given the transmit antenna was non-directional and, geez, made of coat hangers. Two weeks later I installed the TH3 beam (for 10, 15, and 20 meters) and a Cushcraft circularly polarized 2-meter antenna, stacked on top of the HF beam. Here’s how it looked shortly after installation. (My mom says she still gets heart palpitations thinking about me climbing on the roof.)


Kruh residence in New York, topped with two beamsThe roof of my parent's home in Merrick.

That's me, my TH3 HF beam and Cushcraft 2 meter antenna.


With both transmit and receive directional antennas in place, I was ready to go for a contact during OSCAR’s next pass.  On July 24, 1977 (Oscar 7's 12,294th orbit) I made my first space-based QSO, with a ham in Florida (WA4JID). I subsequently would connect with dozens of hams up and down the Eastern Seaboard and western Europe through Oscar 7 and then, in the summer of 1978, dozens more through Oscar 8. Yet I don't recall any of those QSOs as vividly as I do when I first heard the dits and dahs of my Doppler-shifted CW signal coming back at me from space.

In my last few blog post we've been looking at radiation effects on high speed ADCs.  We started the journey discussing TID effects and moved over to the current topic of SEEs (single event effects).  This month the topic in particular is a class of SEEs known as SETs (single even transients).  These are events that are transient in nature as its title suggests.  These types of events occur for short periods of time and do not require a device reset to return to normal operation.  With these events some sort of transient device upset is observed and after a short period the device returns to normal.  In the case of a high speed ADC this can be illustrated by a short duration where the output code is beyond a specified threshold and returns back to normal levels without requiring a reset of the ADC.


I hope you are enjoying the journey so far looking at all these radiation effects.  You can find more details on my blog at Planet Analog: Planet Analog - Jonathan Harris - Single Event Effects (SEEs) with High Speed ADCs: Single Event Transient (SET) .  I trust you will stay tuned as we continue looking at SEEs in my next blog where we will take a closer look at SEUs (single event upsets). 

It surprises some people to learn that, though cell phones offer inexpensive, instantaneous communication worldwide, there are still millions of active ham radio operators around the world. We like to say that there is something for everyone in our hobby. Some of my fellow hams are actively engaged in supporting disaster relief efforts (they were an especially critical part of the still-ongoing recovery in Puerto Rico). We have astronaut hams orbiting in the International Space Station who have set up a station to talk to us earth-bound hobbyists. Other hams design and build high-gain antennas for bouncing signals off the moon. We do these things-and more-while using methods that range from state-of-the-art digital modes (that run on our computers) to good old-fashioned Morse code.


The facet of the hobby I enjoy is operating HF, on eight different sets of frequencies between 1.8 MHz to 28 MHz, called bands, that have been allocated to ham radio. Lucky placement of trees on opposite ends of the long side of my property provided me with a place to hang a 100’ center fed dipole. Dipoles are great. They are inexpensive - they’re just wires, really, with a feed line and some rope to hold up both ends in available trees - plus they are easy to design, install, trim, and fix - and for a bonus they even have a little gain (although it’s directionally fixed because they don’t rotate.)


Here’s an aerial of my property, showing the location of the dipole and, in the inset, a picture of what is called “twin lead” (the feed line from the transmitter to the antenna) and the antenna tuner, intended to “fool” the transmitter into seeing a 50Ω load (even though the actual antenna is not resonant on the frequency.)


Aerial of my property

Using this dipole on HF frequencies I made contacts with all 50 states and over 200 countries as far away as southwest Australia – practically on the other side of the planet. However, when I operated on the lowest ham band, 160 meters (1.8 to 2.0 MHz,) performance was sub-par, since the dipole was only 25 feet above the ground, and that’s just a fraction of a wavelength on 160 meters.  That meant most of my signal was going up, not out. My longest-distance contact on the band was only about 1100 miles away.


During a lunch-time conversation with Analog Fellow Woody Beckford (WW1WW) I explained the problem. He suggested that I convert my 100-foot dipole into a vertical by twisting the leads together and feeding them into one side of a balun. A vertical would provide a lower “take-off” angle for my signal, increasing the distance that my signal would be heard. This is dramatically shown by EZNEC, an antenna analyzer program popular with hams: on the left is the model for the pattern of a dipole at 25’ and on the right side was the model for a 25’ vertical at the same frequency.  Note the lobes in the right-side model showing the lower take-off angle of RF:


Antenna model for dipole and vertical

 Because the dipole – and therefore the vertical - would be limited to 25’ height (a small fraction of a 160 meter wavelength,) we knew getting it to tune up on frequencies as low as 2.0 MHz was problematic. Using EZNEC, Woody calculated parameters for a base-loaded coil that would not only electrically extend the antenna but also eliminate the need for a balun. Here’s the coil I designed from Woody’s specs, with 38 turns of #10 wire around a 2" piece of PVC.  The wire out of the top connects to the twisted pair leads of the twin-lead, an SO-239 provides easy connection to coax from the shack, and the lugs allow for easy addition and removal of the radials:


160 meter Coil


Ah, yes, radials. You see, at this point, the antenna was quite literally only half complete. That’s because verticals need something that dipoles don’t. Vertical antennas can be said to be only "half there," the other half being a "reflection" in the ground, which means they rely on the return currents through the ground, via wires radiating from the bottom of the vertical (hence the word “radials"). So basically that means if the ground system sucks the antenna will, too. It may load up nicely but as Doug Grant - another former ADI employee and ham operator (K1DG) reminded me - so does a dummy load.


To quantify the effectiveness of the ground radial system we can use radiation resistance (R) which, in one term, expresses loss from the entire system: the antenna, the feed lines, and the radial system. We need R – and a special meter (I used an MFJ-259 antenna analyzer) to help answer a number of questions: How many radials do I need to install? How long should they be? Should they be elevated or buried? What effect will my house, my neighbors’ houses, soil conductivity (which can change from season to season) and buried gas, water, sewer, telephone, and cable lines have on R or the vertical’s broadcasting pattern? There were simply too many variables to calculate. Only testing (by laying out radials) would reveal the answers. And this is where the fun began.


With a 1000’ spool of insulated wire in hand, I began a few months of laying out the radials and measuring SWR, R, and X (which I haven’t mentioned yet, is reactance, which is the opposition to alternating current due to the combination of capacitance and inductance inherent in any antenna system), all of which we also want as low as possible. For the purpose of this blog we will focus on R, since X tends to follow R up or down and although a naturally low SWR would be preferable, I could always use the tuner to present a 50Ω load to the transmitter. The first test was with four radials at ground level, laid out at the edge of my property and next to the house, as shown in this annotated aerial:


Annotated Aerial of property with radials


The results were encouraging although at its lowest R, was above my soft target of 50Ω and gradually increased with frequency until, at the top of the band, it was over 60Ω. More radials were clearly going to be needed, but where could I put them? The answer lay in a series of articles authored by Rudy, N6LF, titled QEX articles on verticals and radials in which he details how elevated radials can be as effective - and sometimes more effective - than ground-based radials.  After reading this I ran a pair of radials approximately 3 1/2" off the ground, along the upper support of a wooden fence that ran around my backyard. The results were brilliant, with a drop of at least 10Ω across the 160 meter band.


With nothing but time and plenty of wire left from the 1000’ spool, I experimented with the addition of several radials in various layouts, including one test in which I figured “hey, if one set of elevated radials reduces R by 10Ω, why not add a second set along the middle of the fence?” Imagine my surprise when I measured R actually going up across the entire band! (Who wants to tackle the math to explain that?) Ultimately, after trying a few more layouts and lengths, I found that the addition of two more 70’ radials laid out on opposite sides of the vertical performed best, reducing R across the entire 160 meter band by about another 5. Here’s a plot of performance of the original four radials (in red), with elevated radials on top of the fence (in blue), and with the final two, 70’ ground-level radials in place (in green).


showing measured results of 4, 6, and 8 radials
After sharing these results with Woody and Doug (and asking "what should I try next?") I was given my favorite piece of advice ever: Stop, already, and just get on the air! By then it was October and, with winter coming (a time when 160 meters has optimum operating conditions) I was happy to stop digging up my backyard. Over that winter - and since - the antenna has performed way above my humble expectations, with contacts made in all the lower 48 states and over 35 countries, some with hams over 4,000 miles away in Eastern Europe, Russia, and South America - a HUGE improvement over the dipole. Well worth the effort and strange looks from the neighbors as I was digging and burying wires in my yard.


For more information, data, and pictures of this vertical project, please visit my ham station website.

As the early competition season wraps up, the season for many teams is now over. But for lots of our ADI teams, the excitement is just beginning. New England Championships is this coming weekend, and other teams across the globe are now learning if they have qualified for the World Championship in Houston and Detroit.


North Carolina

Team 2655 has had one of the wackiest and most successful seasons in the team's 10 year history. Platypi went to their first competition with a tall robot with a custom elevator system one of the students developed himself, and it was a struggle. Every match we were tweaking and re-tensioning chain, with a handful of matches ending with the robot flopped over. After the first competition, the team convened and collectively decided to re-work the entire robot from the drive base up. The elevator was stripped off, and a pneumatic arm put in its place. Sure, the robot could no longer reach the scale, but we were now blazingly fast. Platypi held their own at their second event and went on to win their event in Forsyth County near home, earning enough points to squeak into the state championship. Team 900 Zebracorns also came home with a hard fought win for the Chairman's Award at the same event, earning them admission to State as well.


Platypi rose through the ranks at the state championship, earning enough district points to advance to the World Championship in Houston. This year has been a record-breaking year for the team. This is the team's first year without full access to a well-stocked machine shop and tools, yet the Platypi brought home their first ever Robot award in the team's history for the complete redesign and rebuild. The team also took home the Entrepreneurship award twice including at the state championship.


2655 is the only ADI team of three from North Carolina advancing to the World Championship in Houston.


Pacific Northwest

Many of you will remember my interview with Quality Engineer in Camas, WA Bruce Whitefield. His team, 2471 Mean Machine, has been making quite the ruckus in the FRC community, and for good reason! This team had one of the most talked-about robot reveal videos, featuring some impressive autonomous work and a solid "buddy climb" support system which has allowed them to climb to the top ranked team in all of the Pacific Northwest District. With two event wins and two well-deserved Robot Design awards under their belts, Team Mean Machine earned their rightful place at their district championship. Team 2471 had a strong showing and went on to win the event, earning them a spot in the World Championship in Houston. Personally, I'm expecting to see this robot on the Einstein fields! Check out what they were able to do with our ADIS16448 IMU board!




Team 1577 Steampunk has long been an Israel powerhouse, and this team even made the FIRST Updates Now Network FRC Top 25 for Week 5, which takes opinions from the greater FRC community to decide who the best robots are each week. This year they finished all three events, including their district championship, ranked 1 or 2 at the event. Steampunk earned their place as the top ranked team in Israel this year and a rightful place at the World Championship in Houston.


Come Find ADI!

These teams aren't the only ones going to Houston - ADI will once again be present in the FRC Robot Service Center with gyro/IMU support and a show off some of the ways that ADI sensors you see on your robots help revolutionize the way we interact with the world. Come stop by our booth at both Houston and Detroit!


New England


Curious about our New England teams? Their district championship is this weekend! Here are all the teams you should watch for at the event! And don't forget to check back on the FIRST District Rankings website to see who qualifies for the World Championship in Detroit!


Going to the New England Championship Event...

  • 1153 RoboRebels
  • 5422 Storm Gears
  • 4909 Bionics
  • 5962 perSEVERE
  • 5735 Control Freaks
  • 4905 Andromeda One
  • 1058 PVC Pirates
  • 2342 Team Phoenix




This blog is part of a series covering the 2018 season of the FIRST Robotics Competition, FIRST POWER UP. Stay tuned for more updates, including coverage of the Championship Events in Houston and Detroit at the end of April! Get to know the ADI teams, learn more about our donation boards, and meet the employee mentors that make it all happen!


I may work at the world's leading designer and manufacturer of analog and mixed signal semiconductor solutions, but my first love was an old tube radio in my parent's basement. Growing up during the 1960s I would listen to Top-40 AM radio stations from all over the country on this old Philco. (I later had a very brief career as an AM radio disc-jockey, but the story of my misspent youth is for another blog.) Though I later got my Masters in Engineering, my colleagues will attest to my continued love for old wooden and Bakelite radios, as I have more than a few in my cubicle here on the Wilmington campus.


Like many of my fellow antique radio collectors, I always seem to have a few “someday I’ll get to it” radios; the ones we collectors buy, put on a shelf, only to gather dust as other radios get priority. So it was for me with this GE220, a post-war Superhet Bakelite tabletop that I bought for $20 at a flea market many years ago. The fact that the radio was missing its back didn’t trouble me - I liked that it had a shortwave band and was a lot heavier than later, lighter, All American Fives that would soon flood the market. It felt sturdy, almost like manufacturers wanted to assure the American public that our wartime and Depression years of sacrifice were over.



This past summer, after a nice layer of dust had settled on the rig - and with no other radios to work on I took it off the shelf and began with the basics: cleaning the piece (inside and out), a recap and new power cord. Once I had the confidence that the radio powered up safely, I then tackled the challenge of an antenna.

Now the following comment comes from a long-time marketer: it might have been a trend of the times, but GE really went overboard with the brand names. The radio itself was called a “Musaphonic,” (probably an early positioning against upstart FM which would threaten - and eventually conquer - AM radio’s dominance for broadcasting music.) But the marketers at GE went a step further, even giving the loop antenna its own name: “Beam-O-Scope.” Other models, such as the floor-model combination tuner and phonograph H-77, H-78, and H-79, were equipped with the equally impressively named ‘Super Beam-O-Scope.’” (The italics are theirs, from the Rider Manual, Volume 11.) Other models, such as the pre-war tabletop GE L-740, included another grandly named antenna called the De Luxe Beam-O-Scope.


So what is the “Beam-O-Scope?” The GE220 schematic didn’t elaborate, but in the documentation for GE’s H-7x series it explains that: “The ‘Super Beam-O-Scope’ is essentially a tuned coil antenna wound on a frame and shielded by a Faraday screen against electrostatic disturbances” (Again, looks like the battle against no-static FM has already been joined!) But, marketing aside, this left me without an antenna - one that, as seen in the schematic below, was more than just a few loops of wire; it also included a “pick-up” loop for an external antenna that required a 470 Ω resistor and .002 µF capacitor in series. Furthermore L1, the built-in broadcast antenna (the heart of the Beam-O-Scope) also had a 1.5 – 15pF variable cap in parallel.


GE220 Schematic

The loop antenna is critical, since it is part of the first tuned circuit in the radio (feeding the grid of the 12SK7 RF Amp.) If I were to create a replacement I would have to get within the right range of inductance required for the tuned circuit. Further complicating the task was that there were only four wires coming from the radio, three from under the chassis and one off a variable cap that was tied to the ganged capacitor of the RF detector stage.  While researching for other GE220 owners I found this old thread on the forum, one that had been started by another collector with a GE220 and the same issue – no Beam-O-Scope antenna and four wires coming from the radio. The thread included photos of working rigs and one with a list, wire-by-wire, of the five connections from the radio required to the antenna to complete the RF detector circuit.


Unable to find someone with an existing Beam-O-Scope to sell, I resigned myself to having to build a replica from scratch. But then I remembered that a few months earlier, at the New England Antique Radio Club spring flea market, I had spent $5 on this Philco E-808:


Philco E808

Philco E808-5


“Worth five bucks,” I told myself. “It’s not that nice-looking but I can use it for parts.”  When I went to the shelf I was happy to see that the Philco's loop antenna was there.  A quick check with an ohmmeter showed the loop was unbroken (that would have been a bummer) so I began the process of converting it into a Beam-O-Scope.

From a picture posted on the thread I counted 25 loops of wire on the Beam-O-Scope. Now, we all know there’s a whole lot of math that goes into the design of a loop so it will collect RF within the broadcast band and at just the right level to feed the grid of RF Amp (in this case a 12SK7.) That math includes many variables, including the number of loops, the width and permeability of the wire and the size of the space in the middle, just to name a few. But, with all due respect to the designer of the “Beam-O-Scope” and associated circuitry- this isn’t rocket science, and I banked that the tolerances were pretty wide and that the 12SK7 would accept signal in a range that the Philco loop, although smaller than the GE’s Beam-O-Scope, would provide.  The picture below shows the Philco antenna soon after the conversion was started, showing the .005uF fixed and 5-15pf variable caps and 470 ohm resistor and connecting lugs that I added.  I then laid a single loop of wire around the outside of the main antenna for the pickup loop.


Philco E-808 loop antenna at the start of conversion



That left just one more connection to be made: that missing fifth wire from the chassis, which the schematic showed going to the side of the main loop with the junction of the 5 – 15pf variable and C1-A (one of the three sets of ganged capacitors in the tuning section.)


I found it interesting that the person who wrote the original post on that forum had a radio that was also missing the fifth connecting wire. A design or manufacturing flaw, perhaps, that induced the wire to come loose?  Whatever the reason, it was a simple matter to trace the AVC (Automatic Volume Control) line in the radio, finding what looked like the connecting point, and securing a wire there and to the main loop of the hybrid Beam-O-Scope I had created.


GE220 schematic with AVC highlighted


What a treat to have it work the first time, as you can hear and see in this YouTube clip:  The re-assembled radio now sits in a more public place, on a shelf upstairs in the house. I cannot walk past it without feeling a bit of pride, having channeled Dr. Frankenstein to produce a working radio from the parts of two. It’s ALIVE!