• Additional Title: Business Owner, System Architect, Product Manager

How did I get into electronics/engineering?

I grew up in an agricultural community where the kids were always building tree forts, underground hideaways, and running wild. I was always taking things apart to see how they worked: fixing appliances, building electronic kits, rebuilding engines, and making go-carts.

Grandpa Jones would come over for Sunday dinner every week. I remember getting all excited one Sunday because I was rebuilding an engine for a homebuilt go cart and wanted to ask questions about lapping valves. Grandpa Marshall would come over to fix something and I was his sidekick.

Eventually, I became the “fix it” man around the house. I was pretty good at it too, but I had a devious side. When I was 13, I was making a snack and things went bad. I was making cheese on toast in the toaster oven. I put it in the oven, turned it on, and went to my bedroom and forgot about it. I heard this “ding” and panic struck because I knew the toaster was on bake, not toast, so it should not pop open. I ran down the hall and the toaster had flames high enough to be hit the cupboards!

Smoke was billowing out while my mother was taking a nap on the couch. I put out the fire, took the whole toaster oven apart, steel wooled the whole thing inside and out, and put it back together before my mother awoke. After her nap she asked why it was smoky in the house, and I told her I burnt some toast! Ten years later when I told her the story, she just laughed.

That curious and sometimes dangerous little boy has moved on to power electronics and other fun things.

What are your favorite hardware tools?

I don’t really have any favorite tools, but I do have preferred user interfaces. For example, I prefer Tektronix scopes over Agilent scopes. But this is more a matter of familiarity. I live with a scope in my hand because most power electronics designs operate in the time domain. Rigol is my workhorse (its cheap), but I love Tek DSOs with deep memory and serial digital protocol decoding. As for cool, my Oscium scope that I plug into my iPad takes the cake. I was lucky to be able to play with early versions and give them some feedback. I can’t wait for their next cool tool! I know what it is, but I can’t tell you.

What are your favorite software tools?

I work a lot in .Net and dsPIC C. I would prefer to use Haskel, ML or Eiffel, but those languages are not used much in the commercial world. F# is a nice language, but finding knowledgeable developers would be difficult. How many people do you know that have taken a course in Lambda Calculus? Therefore, C# has become my universal solvent for general programming and users interfaces.

Its funny, I prefer to use a Mac rather than Windows, I do most of my design on a Mac, but I have to run a Windows virtual machine to program in C#. If you watched me work you would see my spreadsheets, schematics, layout, modeling on Mac, and my PIC code and C# code on Windows 7. This combination to works quite well. I don’t browse the Internet on Windows or add/remove many programs, so it performs well, and when it gets messed up, I can revert to a snapshot. The Mac just does not have many problems. I tell people that I run Windows in a Mac VM so that when Windows crashes, my mail still works. But the truth is, Windows 7 does not crash that much, so I guess I am a snob.

If there is one tool I just can’t live without it is Subversion. I have a team of developers all around the world and Subversion scales well. There are also good Subversion client tools for Mac and Windows. I also use Guiffy a lot for tree differencing workspaces after merges.

What is the hardest/trickiest bug you have ever fixed?

I can think of two bugs that drove me crazy, one hardware, one software.

I was working on integrating our MOSFET tester with a handler on a test floor in Malaysia. The tester has a parallel cable that receives a start signal, and sends back test results to the handler. For those that don’t know, a handler is a mechanical robot that mechanically connects the MOSFETs to the tester’s electrical probes. One would think what could be easier than a simple parallel interface? Well, add 10 feet of cable, noise, Chinese speaking engineers, some Filipino engineers, an American with jetlag, and time pressure… The handler and tester would run for hours and then lockup in the middle of the night. We did not have a scope capable of catching the signal when the problem occurred, so we performed experiments with time delays and grounding and we just got lucky once and caught it in a scope trace. It turned out the handler had a bug in its latching logic and we had to change our signaling to work around the problem. A simple delay circuit in the cable fixed it. The lesson was, never assume the other equipment is bug free.

The software problem was integrating a wafer prober with tester using National Instruments IEEE 488. I flew to Japan to work with the prober manufacturer. The first step was to get a simple *IDN? Command to give back a valid string. Nothing worked. The command failed every time. After an hour we finally found an IEEE 488 spy tool and looked at the transactions. Everything looked right. We then found the original prober designer. He took one look at it and then explains in Japanese, through drawings and translators, and we find out they did not implement it! In a million years I could not imagine not implementing the most basic command in the industry standard spec! Eventually it is all working, and then lockup!

Our customer and one of the VPs of the prober company were standing right there, and lockup! I had two hours remaining for debug and then I was supposed to go have a nice dinner with my hosts, celebrate, and fly off to Malaysia. The two managers just stood there 3 feet away and watched me debug. No pressure!

Eventually I discovered there was some strange threading issues interacting with callbacks from the NI Driver and devised a workaround. Oh, and did I mention? No Internet available to search for solutions.

In general, integration problems are the hardest because you are not working in a nice lab with all your favorite tools and coworkers. It is more like a pressure cooker with a stuck relief valve.

What’s on your bookshelf?

• The Innovators Solution
• Platform Strategy for High-Technology Companies
• New Product Management
• Switch-Mode Power Supplies
• Transformer and Inductor Design Handbook
• Critical Realism

Do you have any tricks up your sleeve?

“It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.” -Twain

The ability to hold all knowledge as fallible and limited is the greatest trick I know. I highly recommend reading Critical Realism by Andrew Collier if you can find a copy. Critical Realism is a Philosophy of Science.

In this book is a description of open vs. closed systems. Engineers build closed systems. All physical laws are understood. All variables controlled. The digital world and software worlds come close to this. Now let me introduce you to the analog world. This is the world of open systems and just when you think it is closed and all is in control, it bites you in the… well, you know where. Oh wait, the digital world is really analog with states.

What has been your favorite project?

Working on the FTI 1000 MOSFET tester has been a great experience. I am an architectural thinker by nature and this system uses FPGAs, USB, .Net and is full of digital and analog instruments. Imagine an 800V voltage source along side a 100A 10V current source all being controlled by Windows via .Net and USB. Test times of this system are less than 100mS, so all this instrumentation has to be fast and accurate.

I was involved in the initial Verilog/USB design, the communication protocol, and the software architecture. Later I worked on an instrument for measuring gate resistance. It is great fun to work in multiple disciplines at the same time. This is why I am currently a prototyping ultra low noise isolated switch mode power converter. I get to work with magnetics, switching devices, control loops, heat transfer, and code.

Do you have any noteworthy engineering experiences?

Some of my best memories were tricks played on me by my coworkers. One of the best happened at National Semiconductor during my first two weeks. I was a “fresher” as they say in India. I worked in the hybrids group as a test engineer. The test floor was full of probers, handlers, and testers.

The test floor was run by a bunch of Filipino women. That’s how it was in Silicon Valley at the time. Out on the test floor, one of the leads pulled me aside and asked me if I could do them a favor and give some documentation to Mr. Handler. Being new and wanting to make an impression, I took the paperwork and marched off to the other side of the test floor looking for Mr. Handler. I get to the back corner and he is nowhere to be found. I look up at a piece of equipment and there is a placard that says: Handler #1. Next thing I know all the ladies are all giggling.

I have had a great time working in Asia. I have been to a Brahman wedding in India, and a Muslim wedding reception in Malaysia. I have been thrown out of a Hindu temple for taking pictures (I swear, I did not know it was wrong). Working with people that are different from you is both challenging and fun. Street life is a very rich experience for an American.

What are you currently working on?

I am working on an SMPS: prototyping really. The goal is to develop an isolated topology that can scale from 200W to 1KW with noise less than 100uV and control it by PM Bus. When I was consulting to Linear Tech I worked on the LTC3880, which is a buck with PM Bus. This got me thinking that one could build custom converters with PM Bus using a dsPIC or Piccolo. Because I am designing converters for instruments in low volume, I need more flexibility than an off the shelf converter can provide. I am also using digital control so that I can dynamically tune the feedback.

Digital control loops are quite popular with motor control, but they are a challenge with SMPS design. At 20-40 MIPS it is difficult to get the entire math calculated in an interrupt fast enough for the next PWM clock cycle. You also have to deal with limited ADC and DAC resolution, noise, etc. Flexibility does not come for free. Development is more work and the cost is higher.

How did Focused Test come about?

The founders of Focused Test have a long history of success with semiconductor test. A couple of the founders worked at TMT, which was acquired by Credence. All five founders have backgrounds in test; many of us have worked in traditional analog semiconductor companies like National Semiconductor, and in test equipment companies like Credence.

One of the founders got the bug to bring another analog tester to market. It is sort of an addiction I think. Once is not enough. The original two guys lived in Boulder, which is not Silicon Valley where all the test guys are. I happened to live in Colorado Springs, with a shorter commute. We picked up a couple more guys in Silicon Valley and got started. Most of our sales are in Asia, so we opened an office there. Now it does not really matter where we live because we are global.

What direction do you see your business moving in?

I can’t really tell you too much. I don’t want to tip off the competitors. However there are some general trends that every company has to deal with. Back end test and assembly starting moving to Singapore and Malaysia over 10 years ago. Each time a new low cost opportunity emerged, companies moved in. Now there is a lot of activity in China.

Meanwhile, the industry starting consolidating. Credence and LTX, etc. The industry is moving towards the “Rule of Three.” There will be three big competitors and a bunch of bottom feeders addressing the small niches that the big guys can’t address efficiently. First Taiwan jumped into these smaller niches and now China. So we have to compete globally even as a small company, using our cost structure in the US. One advantage we have is bench depth. With founders that have walked the walk, we have a deep understanding of market needs.

What challenges do you see in your industry?

I saw a quote from Databeans the other day. It said that 67% of semiconductor sales are discrete components, such as MOSFETs, and they account for 7% of the revenue. That means very high volume, and low margins.

It is challenging to build equipment that is both low cost and high quality. The semiconductor industry prides itself on quality, and tremendous pressure to reduce production cost. This pressure is felt the greatest in contract manufacturing. Because the volume of tester sales is low, and the capital required to design it is high, this compounds the problem. The larger ATE companies that have consolidated can share technology between internal platforms to achieve a lower cost. But the young companies have to invent their way to success. Lean and mean; smart and fast are what it takes.

What are some of your hobbies outside of work?

I am involved in Product Camp. A friend and I started Product Camp in Colorado because we felt Product Managers needed a way to network and help each other. We had our first camp last November and had well over 100 people. We expect to have over 300 this October. If you have not been to a Product Camp, do a web search and find a local one and go have some fun.

I have some hobbies for stress relief. I love Celtic music and play mandolin. I also play penny whistle, but my wife hates the piercing sound, so it sits in my drawer. I love philosophy and theatre. I am lucky to have 3 daughters that were in all the high school plays. Every once in a while I have to put down my scope and be human.

I also like hiking and backpacking.