Not my father’s workbench, but similar. Photo courtesy of

by George Taniwaki

My father, who died in 2017, was a self-taught expert in electronics troubleshooting and repair. As a teenager in Japan before World War 2, Michio Taniwaki started his own radio repair business. He returned to the US after the war and began his career just as the market for color television and home VCRs took off. He worked for well-known television distribution and sales companies in Denver such as Ward Terry and Co. and Valas TV and Appliance. He practiced his craft at a time when most consumer electronic devices were quite fickle. They contained vacuum tubes that burned out and required replacement and analog components that required frequent repair and adjustment.

As consumer electronics began to incorporate solid-state transistors, the repair market declined. Michio took his skills to the nascent computer industry. His first stop was NBI, Inc. a leading maker of standalone word processing machines, second only to Wang Laboratories. However, after the introduction of personal computers, especially the IBM PC, the market for word processing machines collapsed. NBI could not make the transition and closed.

A group of former NBI employees, including Michio, went on to create Outbound Systems, Inc.  This company made laptop computers compatible with the Apple Macintosh. Apple licensed its Mac OS, but did not license the ROM BIOS chips needed to run the Mac OS. Thus, Outbound had to buy older Macs and cannibalize the ROM chips it needed from them to manufacture its devices. Eventually, Apple released its own line of laptops and notebooks and discontinued licensing of its operating system, causing Outbound Systems to close shop.

While cleaning out the garage and crawlspace at my mom’s house, I gathered up all the electronic test equipment my dad owned with the intent of selling it all on Craigslist. The pictures below show what I found along with some comments.

Analog test equipment

My father had a very nice collection of electronic test equipment. Most were made by RCA, formerly Radio Corporation of America. RCA was the leading producer of consumer electronics from the 1930s until the 1970s. They also made most of the test equipment used to maintain their products.

His most important tool was the oscilloscope which is used to display analog signals on a video monitor. His original scope failed a long time ago and is gone. I think it was made by EICO, not RCA. His latest scope was a Tektronix model (third row right) that he purchased from NBI after they closed.

Each RCA test device has a different function. But they were all designed to fit in the same steel case with leather carrying handle. The face of each device was an aluminum sheet with black engraving, a red RCA logo, and gray-blue knobs. The audio generator (top left) has a easy-to-read aluminum dial engraved with the frequency. The marker generator (third row left) has a clever design that includes a paper scroll with various frequency ranges printed on it. The knob on the left advances the scroll up-and-down to select the range while knob on the right moves a pointer left-and-right  to select a specific frequency within the range.

Sweep generators and marker generators were used to produce VHF and UHF signals to test television components and circuits. A sweep generator produces a range of frequencies within a band. A marker generator is used with a sweep generator and injects a high amplitude peak at a desired frequency inside the band. These signals are fed into the television circuit and compared to the output signal to confirm the circuit is working correctly. The input and output waveforms can be displayed using the oscilloscope.


RCA WA-44C audio generator (left), RCA WR-51A stereo FM signal simulator


RCA-WR69A television/FM sweep generator w/WG-295C video multimarker (left), RCA WR70A RF/IF/VF marker adder


RCA WR-99A crystal-calibrated marker generator (left), Tektronix 2213 oscilloscope


Simpson 262 multimeter with leads and RCA battery tester (left); Assorted test leads

Power supplies

My dad had several low and high voltage power supply units. He also had a makeshift 120V power supply made from a lamp cord with two alligator clips on the other end (lower left).


EICO 1050 battery eliminator and charger (left); RCA WP-700A power supply


Adjust-A-Volt (Standard Electrical Products) power supply (left); Unknown vendor high voltage power supply


Makeshift 120V power supply (aka, The most dangerous item in the house)

Cables and parts

Finally, every good craftsman has a junk cabinet. My father’s contained lots of cables, soldering guns, and spare parts. He also had an entire dresser filled with unopened boxes of RCA and Sylvania vacuum tubes of various sizes and types. There may have been over fifty total. Several years ago he (or more likely, my mother) tossed out the dresser and all the tubes. Too bad. Vacuum tube stereo amplifiers are now quite popular among audio aficionados and replacement vacuum tubes are now selling for $10 to $100 each on eBay.


Assorted audiovisual cables (left); Assorted soldering guns, desoldering pumps, and solder wicks


Assorted potentiometers, transformers, coils, and capacitors (left); Assorted resistors, varistors, transistors, switches, plugs, and light bulbs

All photos by George Taniwaki

Update: Added hyperlinks to references.

by George Taniwaki

Last week I had two small electronic devices fail. I tried to fix them both and failed. Here’s a short description of my attempt to fix the first one, a USB KVM switch. Along the way I ponder the role of industrial design and intellectual property rights in China.

I have two computers, a desktop PC and an iMac. Rather than have separate keyboards and mice, I share them and toggle between the computers using a KVM switch (the abbreviation stands for keyboard, video, and mouse). My setup is shown in the photo below.


Figure 1. My home computer setup, the KVM switch is on the desk, above and to the right of the keyboard

Last week the KVM switch quit working (after only a year!). I opened it up but could not see any loose connections. My guess is that the clock chip failed. Clocks are cheap but so is a new KVM switch. So rather than spend time and money trying to debug and fix this device I bought a new one for $9.50 on Amazon.

When it arrived, I noticed it looked just like my old one right down to the fact that the button on the switch was slightly off-center (see photo below). The only apparent difference was the printing on the case and the UPC codes.


Figures 2a and 2b. The front of the old KVM switch on left, new one on right

A UPC code consists of two sets of 6-digit numbers. The first 6 digits is a unique identifier for a given manufacturer. Each manufacturer is then able to assign the next 5 digits to products as they wish and then calculates a checksum to assign as the last digit.

The two switch boxes shown above had different manufacturer IDs and different product IDs, indicating they were made by different manufacturers.

This intrigued me. Both KVM switches are made in China. This made me wonder about why the UPC codes don’t match. Some possibilities I came up with:

  1. One company reverse engineered (i.e., brazenly stole) the other’s design for the KVM switch (even given the poor reputation Chinese companies have for respecting intellectual property, this seems unlikely)
  2. One company made the original design and licensed it to the other (seems unlikely, once the first company did all the work to line up suppliers why would it want to help a competitor enter the market?)
  3. Both manufacturers licensed the design from a third company that only does design and lets others manage the supply chain and manufacturing details (given the low-cost of the item, this seems an unlikely product category for a successful design firm)
  4. Between the time I purchased the first KVM switch and the replacement, the company changed its name, was acquired, or changed its manufacturer ID and product ID for the switch for some other reason (this seems most likely)

I was a bit concerned about the reliability of my new switch. If it has the identical circuit design of my old switch, it might fail prematurely. So I opened up both KVM switches to see the circuit board.


Figure 3a and 3b. The circuit board of the old KVM switch on left, new one on right

Surprise! They are not identical. The new one has three chips compared to four for the old one. And some of the chips are smaller. These smaller chips in the new switch have more pins (indicating higher complexity).  Fewer chips should make it more reliable (hopefully). Both have a 12MHz clock chip (which is the full bandwidth speed for USB 1.0 specification). Both have an identification mark of MT-201UK-CH which indicates they were both designed by the same firm.

Someday, if this KVM switch fails, I might buy one that supports HDMI video connectors. Then I will have the left monitor dedicated to Windows, the right monitor dedicated to Mac and the center monitor toggle between them along with the keyboard and mouse.


Go back to the photo in Figure 1. Notice that the image on the monitors on the PC are much bluer than the image on the Mac. I’m a color printing expert and I’ve futzed with the white point and color balance on both machines and can’t get them to match. Why is it so hard to profile and calibrate monitors?


Part 2 of this story is posted Aug 2013.

All photos by George Taniwaki

Today’s Wall St. J. (subscription required) reports that several medical companies are adopting Apple’s iPad for use by their sales force when calling on doctors. This is a major coup for Apple since pharmaceutical and medical device manufacturers each have thousands of sales people.

Presumably these sales people already have laptop computers that they can use when calling on doctors. So why add an iPad to the arsenal of sales gadgets? The following  anecdote told by Mike Hedges, chief information officer of Medtronic, a manufacturer of implantable heart devices, provides the key.

“Mr. Hedges said he had attended a dinner with several cardiologists when one asked about Medtronic’s drug-eluting stent. Normally, the conversation would have shifted to another topic, but one of his salesmen quickly pulled out his iPad with the product information, which kicked off several hours of discussion.”

It appears that there may be several factors that can make the iPad (or other slate device) an effective tool for sales. First is the slim form factor. It is easy to pull an iPad from a briefcase without stopping a conversation. It probably can be done with only one hand and avoid the disrupting need to pick up the briefcase, dig through it, and pull out a bulky laptop with a bunch of cords attached to it.

Second is the quick start-up time. Since the iPad is normally in standby mode rather than off, launching takes less than one second. So once it is visible to both the sales person and customer, they don’t just sit waiting for the device to do something useful before letting the conversation continue.

Third is ease of navigation. The sales person can navigate using fat fingers only. There is no need to use fine motor control and manipulate a pointing device, keyboard, or touch pad. Again, the conversation can naturally continue while this is happening.

Finally there is the visual appeal of watching someone use an iPad. The device itself is cool looking. But also appealing is the ability of the sales person to share the view of the desktop with the customer while navigating. This again helps keep the conversation flowing and engages both the sales person and customer when showing a video or other presentation. Compare that to the same scenario using a regular laptop. The screen faces the sales person until the presentation starts running. Then the sales person will rotate the laptop to show it to the customer. Again, the laptop faces one person at a time, and is not conducive to sharing. It supports show and tell, not a natural conversation.


Apple iPad, a sales weapon. Image from Apple

Today’s Tech. Rev. describes an epaper technology developed by Nemoptic that the company hopes will be cheap and efficient enough to be used as disposable unit price tags, replacing the ubiquitous paper unit price tags currently used on store shelves. Very cool.


A paper unit price tag. Image from

About 20 years ago, I came up with an idea for wireless shelf tags (just the concept, with no idea how to build a working prototype) and showed it to my mother, who was a manager in a supermarket at the time. Given prices and technology at the time, my thought was the tags would be permanent, not disposable and would require an external power source. I explained to my mother the labor savings from having a tag that could be updated without requiring a clerk to print and attach it. I also explained how the tags would always display the same price as the scanner at the cash register because both would get price information from the same database. (This doesn’t necessarily mean the price is correct though, just consistent.)

Her reaction was interesting.

First, she wanted to know how the tags would be attached to the shelf. I said probably with a couple of screws. She explained to me that the tags had to be easily moved as shelf space needs changed. This is easy to do with paper tags with the current plastic tag holder. She was afraid that securing electronic tags to the shelf would make it too hard to move them.

Then she said the biggest problem with paper tags is not that they need to be updated. It was that they constantly need to be replaced because they are stolen by customers. Really!

Even if my electronic tags were screwed to the shelf and were useless when disconnected from the shelf, they would be an irresistible target for malicious thieves. My mother said that sometimes the only difference between a profitable store and a money losing one (in the same chain) was the amount of shrinkage. So preventing theft and damage is a key consideration in selecting materials used in a retail environment.

In response, I proposed replacing the tag with a bar the full width of the shelf. That way it would never have to be moved. The clerks would “move” a price tag by sliding it along the bar. She then asked how durable the display bar would be. Even if customers couldn’t steal them, she was sure some would attempt to pull them off the shelves.

She also wanted to know how a clerk would move a tag from one shelf to another and how much training the clerks would require to do this. And she wanted to know, how would clerks prevent malicious customers from watching them and then copying those actions themselves throughout the store. The user interface and security issues stopped me from pursuing the idea further.

The advantage of the paper shelf tags was that they were easy to move and cheap to replace. Factors like easy to update and absolute accuracy were not important to her. So even if disposable epaper unit price tags are less expensive than paper ones, they may cost more to implement, if customers take them and you need more of them.

by George Taniwaki

Every year Mattel introduces several new editions of its ”I can be…” Barbie doll. They include the clothing and accessories necessary for this famous doll to be successful in a particular career. (Me thinks this may have more to do with selling toys than broadening the career sights of little girls, but I digress.) This year, the company decided to let girls vote for their favorite career. The top vote getter would be featured as this fall’s new edition. Mattel heavily promoted the contest on social media sites like Facebook and Twitter. (Not sure how I missed it.)

On Feb 12, at the New York Toy Fair, Mattel announced that the winner of the popular vote for the 125th special edition Barbie was computer engineer. It’s hard to tell from the drawing below, but Barbie is wearing a Bluetooth headset, has a t-shirt with computer code written on it, and has a smartphone strapped on. As my friend Jim Reichle would say, “The heat, the heat.” (It’s an inside joke from Caltech.) But to compensate for that bit of nerdiness, she’s color coordinated with hot pink laptop, glasses, and wristwatch (maybe it’s a revival of a Spot Watch).


Winner of the popular vote, computer engineer. Image from Mattel

You can preorder a computer engineer Barbie from the Mattel web site. I’m sure it will soon become a popular collectible among a certain crowd here in Redmond.


I thought it was rather odd that computer engineer would beat out the other careers that Mattel offered girls to choose from: architect, environmentalist, news anchor, and surgeon. I didn’t think many girls consider computer science as an attractive career. After all, it isn’t very popular among women entering college. And it turns out that computer engineer wasn’t the first choice of the girls, news anchor was.

An article in Wall St. J. Apr 9. reveals that a viral campaign started by computer engineers hijacked the voting for Barbie’s new career. Computer engineer Barbie became a cause célèbre among the digerati. For instance, a writer for SQLblog encouraged his followers to vote. The influential GeekGirlCamp ran an appeal asking readers to “Please help us in getting Barbie to get her Geek On!”

In the end, Mattel realized the power of social media cuts both ways and decided to have two winners. News anchor was declared the winner of the girls’ vote while computer engineer was the winner of the popular vote. Mattel will release an anchorwoman Barbie in time for this year’s holiday season.


Winner of the girl’s vote, news anchor. Image from Mattel


Why did computer engineer Barbie attract so much attention? Well, I think part of it may be the odd sense of duty (or sense of humor) that geeks have toward promoting their culture. (Had I known about this contest, I certainly would have voted for computer engineer.) But part of it may have been to actually raise awareness of computer science as an attractive career for women.

Many formerly male dominated professions such as law, accounting, mathematics, medicine, and biological science are now much more gender balanced, or in some cases becoming female dominated. However, engineering, physical science, and computer science are not.

In fact, the proportion of men in many of these professions never fell below 70% and are actually on the rise again. A Wall St. J. blog post states that the number of women in computer science has been falling while the total number of workers has been growing, causing a steep rise in the male-to-female ratio. (And I don’t think it was caused by girls hearing Barbie say, “Math class is tough.” A study published in Science Jul 2008 shows that the gender gap in math achievement as measured by standardized tests has disappeared. So it is likely something else is causing it. The article points to a problem with standardized tests themselves and the pernicious effect of the No Child Left Behind legislation. But I digress again.)

Most of my own college education and work experience has been in heavily male dominated fields. My freshman year was spent at California Institute of Technology, where in 1977 fewer than 10% of the undergraduates were female. I transferred to the Colorado School of Mines where the proportion of females was about double that. At both schools there were almost no female graduate students or professors. Even after almost 30 years, a Amer. Assoc. Univ. Profes. 2006 report cites Caltech as the doctorate-level school with the lowest proportion of female full professors (14%) in the U.S. The next lowest school? Mines at 16%.

In its first 100 years (from 1874 to 1973), Mines graduated a total of 14 women. The percentage grew quickly thereafter and was still rising while I was attending. But then it stopped. A recent story in Mines Magazine shows that the proportion of women at the school has remain steady for the last twenty years at about 25%. However, the type of student may be changing as women now hold about half of the student leadership positions.


A professor at Mines, a woman named Tracy Camp, authored a paper that appeared in Comm. ACM Oct 1997 that highlighted the falling enrollment of women in computer science programs and warned of its consequences to the U.S. economy and global competitiveness. She urged action to identify and counteract the forces that were, and still are, leading fewer women to seek degrees in computer science and careers in the IT industry.

I wondered if Dr. Camp was one of the adults who voted for computer engineer Barbie. When asked, she said, “Yes, I voted for the computer engineer Barbie. I also sent an announcement out on my networks, which helped add a lot more votes.” She doesn’t feel bad at all about adults hijacking the vote, “Research has shown that we need to change the image of computing to get more girls interested. Barbie may help.” (One of the great things about writing a blog is that I can send impertinent emails to busy people and they respond, but I digress.)

So there you have it. Computer engineer Barbie is a child’s toy, a collectible, a role model for career-minded girls, an Internet meme that provides a lesson in how social media is changing marketing, a symbol of U.S. economic competitiveness, and a partial solution to the gender gap in engineering. Who knew?

My friend Carol Borthwick sent me a link to 20 photographs of vintage computers featured in Time. The photographs are by Mark Richards, who has a book out entitled Core Memory: A Visual Survey of Vintage Computers.

CoreMemory    Image from

Most of those pictures are for devices that are before my time. So maybe you want more recent images of the 10 worst keyboards of all time as judged by PCWorld in 2007. It’s amazing to see how much variation there was in the past for something that is so standardized today.

WorstKeyboards  Image from PCWorld

But of course why dwell on the quaintness or horrors from the past. At the other end of the scale is a brand new computer keyboard that does everything. The Optimus Maximus keyboard has an OLED screen on the top of each of its 113 keys. Each key-top screen is programmable so that it can display a variety of images, or even video at 10 frames per second. And it’s yours for only $2,400.


Video from ThinkGeek

There are actually some practical uses for a keyboard that has programmable key labels. First, if you like to use keyboard shortcuts or are unable to use a mouse because of a disability, then having the feature list appear when you press a modifier key like Shift, Alt (Option on a Mac), AltGr, Ctrl, Windows (Command on a Mac), or Function would really aid in recall of the features available.

Second, if you customize key bindings, which someone who uses keyboard shortcuts a lot does, then being able to modify the key label would be helpful to remind you which key is associated with which command. It would also serve as a notification to a guest user that the keys have been mapped in a nonstandard way.

Third, if you are bilingual and often type in another language you may discover that you need two keyboards, one for each language’s keyboard layout, and switch between them. Alternatively, you can stick with the keyboard for your primary language and use a character map utility to find and insert characters needed for your secondary language but not available on the keyboard of your primary language. Using either method is a big hassle. Having a keyboard with programmable key labels would allow you to use a single keyboard and switch mappings to enter text for both languages easily and conveniently.

Finally, in the very rare case that you are a professional software tester specializing in internationalization and localization or a prodigious multilingual savant, then you can switch between all the available keyboards in the world and type every Unicode character without switching between physical keyboards.

Of course, keyboards will be obsolete soon, if you believe the leaders of the natural user interface (NUI) movement. We’ll soon be using gestures and voice to interact with computers. So this whole blog entry is moot.

Recently, Apple released a beta of the latest version of its iPhone operating system, called iPhone OS 4.0. As part of the release, it is distributing a new version of the software development kit (SDK) so that developers can write software that will take advantage of the new features of iPhone OS 4.0. The biggest advance is the ability to run multiple applications at once, called multitasking.

Currently, if you are using a third-party app and you answer the phone, the app quits. Further, third-party apps cannot run in the background and update (like you would want for weather, navigation, IM, etc.) The iPhone, like any modern personal computer runs many processes simultaneously. However, the ability of external developers to control this was limited in earlier releases of the iPhone OS. (Note that this limitation only applied external developers. Apple’s own apps multitask easily. For instance, you can take pictures or get maps while using talking on the phone.)


iPhone 4, now with multitasking. Video still from Apple

Some developers have reported that Apple has made some subtle but significant changes to the SDK license agreement for iPhone 4.0. These changes increase Apple’s ability to generate revenue from users to the detriment of developers.

First, Apple added a clause to prohibit the use of third-party compilers to create code for the iPhone. Most developers already use the Apple tools and are not affected by new restriction. The exceptions are developers who use, or want to use, Adobe Flash. Just a few days before Apple released the new iPhone SDK, Adobe Systems released the latest version of its tools to create web sites and rich Internet applications (RIA) called Creative Suite 5. One of the big feature of CS5 that Adobe had been touting was the ability to create apps using Flash and compiling them to run on iPhones.

After several days of tense negotiations and public sniping, Adobe has finally thrown in the towel. In his blog, Mike Chambers, the Flash product manager, says:

“We will still be shipping the ability to target the iPhone and iPad in Flash CS5. However, we are not currently planning any additional investments in that feature.

…if you want to develop for the iPhone you have to be prepared for Apple to reject or restrict your development at anytime, and for seemingly any reason.”

By refusing to support Flash, either in its Safari browser or in applications, Apple ensures that developers for the iPhone must use its tools to create the applications. In turn, this means the developers must use the iTunes store to distribute them. These restrictions ensure that Apple will get a cut of the revenues and prevent distribution of apps that are not desirable. Adobe’s failure to anticipate Apple’s moves or create a viable mobile strategy is described in today’s Business Insider.

A lot of the content on the iPhone is free, so it may seem that Apple has generated a lot of negative publicity for itself for little benefit. However, having complete control over the single source of content for the iPhone will become even more important as Apple negotiates distribution of apps and content for its new iPad. If every book publisher, magazine publisher, television network, and movie distributor has to get approval from Apple for its content to appear on the iPad, then Apple is in a position to control pricing, promotion, and exclusive distribution rights. This can be very powerful and profitable indeed. [Disclosure: I do not own any stock in Apple or Adobe Systems, except as part of mutual funds that I have no control over.]


Last week, another developer revealed that Apple has also added a clause to its iPhone SDK license to prohibit “the use of third-party software to collect and send device data to a third-party for processing or analysis.”

In passing, this would be seem to be a good thing. It appears to prevent application developers from invading the privacy of Apple’s customers, or using location services to display annoying ads targeted by geolocation data. However, it is important to view this ban as part of the wider battle between Apple and Adobe (and others) over mobile strategy. That’s because Apple doesn’t want to prevent developers from tracking users and showing them ads. It just wants to prevent them from using another company’s tools when doing so.

Last fall, Adobe Systems acquired Omniture, a web analytics firm, for $1.8 billion. At the time, I didn’t understand the logic. Adobe sells web and print design tools to designers. Web analytics software is purchased by marketing and advertising professionals. They are not the same people. Why would anyone care that they could now buy both products from the same company? But now as Adobe releases the latest version of its Creative Suite 5 web and print design software, it is pretty obvious what the advantage is. CS5 makes it easy for designers to integrate analytical markers into web content that can then be measured using Adobe’s web analytics tools. And as much of the web moves to mobile devices, distributing location-based advertising could be a critical source of revenue for Adobe.

Most software developers for mobile devices are not advertising-savvy. To help them, Apple recently purchased Quattro Wireless for $275 million and is using the technology from that acquisition to launch as its new iAd mobile advertising platform. Apple will sell the ads and give 60% of the revenue to developers.


iAd mobile ad service, I insist you like it. Video still from Apple

However, by altering the SDK license to ban application developers from using third-party software, Apple is locking in developers to its service and shutting out Google, Adobe, and others. Through its actions, Apple has strengthened its ties to its most loyal developers, pushed away the ones that were using outside tools, stopped its competitors, and done it all in a way that isn’t highly visible to end users. These strong armed moves may not withstand antitrust scrutiny. But if they work, they will produce great revenue streams for Apple.