System Design: Fun or Boring?

I wouldn't say I'm your typical analog engineer, by virtue of the fact that almost all analog engineers have more experience than I do. Really what it means is that I'm chasing the pack, trying to catch up. In some situations, that's really nice. Notable examples would be finding a mentor and getting help with the circuit during a typical workday. In other situations, it's quite a hindrance. Sitting amongst my age group and trying to talk microcontrollers, I'm often left out. Not because I don't understand them but because they're not my main focus.

And that is the focus of this introductory article, namely the fact that there are fewer analog problems to solve these days.

“Hunh?” says everyone reading this article on a site called Planet Analog, as well as the people who hired me to write on said brand new site.

You see, the problem isn't that analog has gone away as a physical phenomenon. Signals still exist in the environment, voltages go up and down, sensors need to be wrangled, signals need to be amplified, and in general we have more data about our environment that we want to capture, rather than less. So there's no lack of need or lack of desire to solve these problems. As a board-level designer working with analog components, I have many of these problems dropped on my desk on a regular basis. I'm guessing many of you sit in a similar chair.

What I'm really speaking of is how we attack these problems. The trend I continually see is chip vendors pulling functionality into their realm of expertise. Every day I see announcements about all-in-one solutions to solve some of the most vexing analog problems: Ultrasound! Weigh scales! Gas detection! The list goes on and on. If there's a specific and profitable niche application, a chip company is trying to roll a solution that will solve all associated problems.

Sometimes I'm sitting there cheering these advances as much as anyone else. I've got a job to complete. If there's a single-chip solution that solves that problem at a low cost and a low impact to the projects I'm working on (especially while keeping in mind long-term manufacturing issues, such as supply chain and obsolescence threats), then it's a beneficial decision for everyone. There is, however, one obvious problem with this:

It can get a little dull.

Is my day-to-day work boring? Not at all! There are still myriad problems to chase on a regular basis, often surrounding the little chip that is working its wonders, either amplifying or suppressing or digitizing signals from the real environment. And of course, one AFE (analog front end) cannot comprise an entire system, as much as large chip vendors would love that to be the case. (“We could charge hundreds per chip!” say the drooling business managers.)

Outside of the core chips there are signal integrity issues, power supplies to optimize, inputs to protect, boards to layout, systems to be designed, firmware to write (“ack!”,) and much more. I still can't help but feel a slight twinge of regret, though, as I thumb through old app notes. Not because I think solving problems was easier back then (quite the opposite!), but because the required solutions had an elegance that is now hidden beneath epoxy resin and layers of silicon and metal (which, by no small coincidence, is where many of the most interesting design problems reside these days).

Regardless of how many years you've been working on analog, you've likely seen another trend that also helps to feed this system-on-chip craze: It is almost required. Whereas departments were once filled with cadres of advanced analog experts, departments have been shrinking in general, and we've all been asked to do more with less.

So perhaps you're reading this article in a half empty office, not with the familiar grimace of an engineer longing for more op-amps, but instead greeting it with cheer. All you want to do is get your product out the door. This site will still be for you! The fact that you have a job to do quickly often means the resources provided here will be of even more value because of your time and resource constraints.

So what am I trying to say in all this? Namely that we face a changing landscape, one that has changed even since when I started in the industry less than 10 years ago. I've always enjoyed talking about and analyzing these changes as they come along. Hopefully, we'll continue to do so together here at Planet Analog.

7 comments on “System Design: Fun or Boring?

  1. Jim Connolly
    January 11, 2013

    Chris. Nice job of putting your role in perspective. From what you describe, the answer to your question has to be more along the lines of fun than boring.

  2. RichQ
    January 11, 2013

    Perhaps what has shifted is the taxonomy. When I was a young engineer there were clearly analog designers and digital designers. Then came MCUs and digital designer started merging with programming. Now programming of MCUs can be so complex that MCU programmers are re-emerging as a separate discipline.

    Perhaps what has happened to the category of “analog designer” is that it has split and blended with digital design categories. Now we have “signal integrity” and “power” and “high speed signalling” and “precision sensing” designers as categories.

    Most application areas need a blend of analog and digital techniques to resolve these days and all of us are “design engineers.” Perhaps we should stop bothering to parse the group further.

  3. bolaji.ojo
    January 11, 2013

    Old is good but fresh is even better and that's not the preserve of experienced hands alone. So, I am looking forward to your fresh (perhaps even inexperienced) angle on issues in the analog world.

  4. Chris Gammell
    January 11, 2013

    But then the site would need to be renamed!

  5. RichQ
    January 11, 2013

    Good point! Guess we should stick with the old taxonomy for awhile longer then. Or else create a Planet Power, Planet Signal Integrity, Planet Precision Sensing….

    We'd end up with a whole solar system. Best to leave well enough alone, then.

  6. TheMeasurementBlues
    January 11, 2013

    “The trend I continually see is chip vendors pulling functionality into their realm of expertise.” Like a smart phone in miniature. Everybody wants more features in less space. But in many ways, that's been going on for years. Yu're under pressure to get the product out and if you can buy a function rather than design it, that's what you do.

  7. Chris Gammell
    January 15, 2013

    No, I haven't had a boring day yet. It is simply a refrain I use with chip vendors when they tell me about new AFEs and other highly integrated analog components: “you're making it boring for me!”. I know there will still be MANY issues to deal with, they're just different issues now!

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.