Connect with us

Closing Control Loops

Discussion in 'Electronic Design' started by Tim Wescott, Feb 24, 2006.

Scroll to continue with content
  1. Tim Wescott

    Tim Wescott Guest

    So, I've put the finishing touches on my book, and the publisher's
    marketing department is asking me questions. Some of them I don't have
    good answers to, because (a) I'm on the fringe of my target audience,
    and (b) everyone learns differently, and my way is through osmosis and
    thinking, which doesn't make it easy to cough up a bunch of specific
    information like conferences and magazines.

    The book's title is "Applied Control Theory for Embedded Systems", the
    subject should be self explanatory. So if you feel that you're in the
    target audience, and if you're in a generous mood, here are some
    questions I have for you:

    What do they call you at work (sorry -- what's your official job title?)

    What magazines do you read? Websites you visit regularly?

    Do you go to conferences? Which ones? (this is the one that motivated
    this posting, by the way -- I only go to the Embedded Systems
    Conference; I know there's a circuit-related one in Silicon Valley but I
    can't remember the name, nor do I know if there are any ones anywhere else.

    Have you had to close a control loop recently? Did you do a web search?
    What keywords did you use?

    Have you read any related books? Was it just a college text, or was it
    specifically directed at closed loop control for the practicing embedded
    systems designer? What did you think of it? What was it?

    Thanks in advance.


    Tim Wescott
    Wescott Design Services

    Posting from Google? See
  2. Rich Grise

    Rich Grise Guest

    Isn't that what academicians call "negative feedback"? ;-)

    Good Luck!
  3. Genome

    Genome Guest

    A self procreating **** up?
  4. Tim Wescott

    Tim Wescott Guest

  5. Hmm. That's the finest definition of life I've heard!

  6. Padu

    Padu Guest

    Hi Tim
    They call me Software Engineer (they call me other names too when they find
    bugs in the software, but I doubt you are interested in knowing), but my
    business card says Software
    Development Manager.

    Nuts and Volt, Servo, Motorcyclist, Computer, IEEE Potentials, Dr. Dobbs

    Google news, Orkut, CNN

    Not so often, I wish I had the money/time to attend to more conferences. The
    last one was IC-AI (International Conference on Artificial Intelligence)
    which I published and presented a paper.

    Yes. Yes. PID control; control theory; fuzzy control

    Yes. A couple of books actually.
    Got some hints from Schaums for control theory, and two other books that I
    borrowed from the university's library.
    The problem with all of them is that they go deep on the math of control
    theory but are not practical (no relationship between CT and embedded

    Do I get a free book? ;)

    Good luck with your publishing... A man has three missions on Earth: Have
    kids, plant a tree and write a book. I accomplish the first two... someday
    I'll write a book.


  7. I hopr you have lest out Nyquist diagrasm and Root Locus. Never seen anybody
    using them on a real design other than academics.

  8. On Friday, in article
    They need what they think they need for standard tick box list of checks.
    Sounds like a sensible idea fo a book..
    I used to have all sorts of job titles, as often every company has different
    meanings for job titles (let alone the made up ones that appear now).

    Having spent over 12 years working for myself in a one man company I
    have various titles (some legal, some other meanings) depending on what
    I am doing at the time and the questioner. Job titles only really have
    relevance to questions.
    Magazines IEE/IEEE periodicals (some), Circuit Cellar..

    Websites are not something I spent a lot of regular time on specific ones
    as it depends what I am working on or doing, except in bursts for datasheets
    or specific issues, it would be Dilbert, Google, Ebay.
    Never gone to a conference, occasionally seem some of the papers. From
    various sources it depends on the hosting country, speakers country,
    sometimes who is sponsoring their 'research' and what is deemed patentable
    under which jurisdiction as to whether you see enough detail anyway. I know
    of cases in many fields where published papers are very 'sanitised' for
    varying reasons to make you wonder if the paper is worthwhile.
    Mainly revisiting various loop controls/tuning for PID, servo loops and
    nuances. Considering some of the project time spans, recent could be two
    years (due to many other things).
    Most of them I found too abstract or too ideal mathematical models with lots
    of assumptions. some of the worst assumptions I have seen is we have a model
    for an actutator/motor that applies to all actuators/motors, where as in
    real life modeling of the subsystem is quite complex and cannot always be
    done theoretically but has to be done at first empirically as the full design
    is 'n' stages down the road. In one instance to get the model and even think
    of tuning it was complicated by differing response times of sensors to software

    Quadrature encoders
    position indicators
    optical position from line scan processing

    Further complicated by temperature, standing or moving before hand or worse
    still some of it was effected by different control loops controlling
    high voltage light sources and other parts of the system.

    That was even before you looked at mechanical interactions, tolerances,
    determining fault conditions (mechanical, electrical, optical).

    Too often the books forget about fault conditions, that can be directly
    determined or have to be determined from other sources.

    A lot of theory books are a bit like the old sciences joke

    A bookie wanted to know if there was a scientific way to
    determine the outcome of any horse race, so he asked the
    three scientists.

    The chemist said "Too many unknowns and variables"

    The biologist said "Too many factors to make it possible"

    The Physicist said "Yes"

    When the bookie asked how, the physicist said

    "For spherical horses in a vacuum"

    The worst demo of control software for computer (PC and larger) data
    acquistion systems I saw had classic limits and real world is somewhere
    else issues. Basically the demo consisted of a model of a simple oven that
    was heated and its temperature would rise, until the door was opened. The
    longer the door was closed or open the higher or lower the temperature would

    Now those here would hopefully know that if the door was opened the lowest
    temperature that could be reached was ambient in real life, in a model
    should be absolute zero. This model had no limits so the temperature could
    go below absolute zero or higher than the sun! So did not show the software
    in best light of showing how good the software could be.

    Hopefully your book will not be like that.
  9. Fred Bloggs

    Fred Bloggs Guest

    I knew it!
    The book is too small.
  10. Guest

    EETimes, Sensors, Medical Devices, Machine Design
    IEEE Spectrum, .....for some reason they won't send me ESP
    I got ESP back in 1995, but they won't renew my free subscription :(
    CircuitCellar, Nuts and Volts, sometimes Servo.
    No. Can't justify the cost. Way to expensive considering the
    lost time, conference fees and travel costs.
    Sure, but what does that have to do with the websearch?
    Quantum Programming. I thought it was great!
    A good summary of state machines and an alterative ways to use them.

    On Time and Under Pressure by Ed Sullivan. An old book. . A great
    overview of what makes a great organization tick, particularly the
    Version Control and Automated Build processes,

    Patent it Yourself, David Pressman. A great step by step process for
    submitting a patent.

    The Scientist and Engineer's Guide to Digital Signal Processing
    Free PDF book. I've been reading chapters. Great stuff.

    I don't generally read "controls" books anymore unless they are
    product manuals. For anything new, I'll read technical publications,
    or the web. For anything old I just refer to my bookshelf.
  11. Tim Wescott

    Tim Wescott Guest

    You should look over my shoulder some time, then.

    I use root locus when I'm brainstorming controllers, and I have my
    spreadsheet program display top & bottom Nyquist & Bode plots when I'm
    tuning from measured frequency responses -- I put a circle of diameter
    1/sqrt(2) on the Nyquist plot to indicate the 3dB sensitivity point, and
    tune to that.

    I _don't_ spend many pages showing how to construct root locus plots --
    while you do learn something from it there's a gazillion math
    applications that will do it for you.


    Tim Wescott
    Wescott Design Services

    Posting from Google? See
  12. Tim Wescott

    Tim Wescott Guest

    -- snip --
    Yes, that's it. Done, thanks be.
    That's three years worth of writing! Perhaps I've just managed to
    condense the wisdom down, eh?

    I'm going to take the first complimentary copy they give me and tape a
    red pen to it on a string, to make marginal comments. I expect that if
    it lives to a 2nd edition it will grow.


    Tim Wescott
    Wescott Design Services

    Posting from Google? See
  13. Tim Wescott

    Tim Wescott Guest

    -- snip --
    I tried to keep it grounded in reality, to live up to the "Applied" in
    the title. Most of the examples are simplified, so that I can talk to
    the point that I'm trying to make, but I try to keep the other issues

    I think there's a place for simplistic models -- I've done motion
    controllers where the emphasis was on getting a mechanism from point A
    to point B without toasting any electronics or breaking/jamming the
    mechanism, and accuracy was taken care of by other parts of the system.
    In those cases a very simple model that ignored such complications
    like backlash and friction worked very well -- as long as I remembered
    what the limitations of the model were. I tried to keep up a continuing
    theme that you need to remember what assumptions you made about your
    model, and what those assumptions mean.


    Tim Wescott
    Wescott Design Services

    Posting from Google? See
  14. I love
    "Readership: PRIMARY MARKET: ...
    SECONDARY MARKET: Engineering students at the *more practical*
    engineering and technical schools; ..." emphasis added ;/

    Lets out my Alma Mater.

    The advanced BS(Physics) students came to Phillips Hall to take our
    *REQUIRED* course in quantum mechanics. In the 60's I would have gotten
    more hands on "engineering" experience as a BS(Physics) student than I
    had a chance of as a 5-yr BSEE candidate. I fault IEEE and other
    professional societies for creating that problem with ill thought out
    course "standards". But the basic problem predates the various
    societies. My father got an ME degree, rather than persuing an EE
    degree, because (in the 1920's) the ME candidates received a much
    broader background in what would now be considered EE. At that time and
    this institution, EE was power plant, power distribution, and AC/DC
    machines. Whereas, he had the opportunity to study in an "applicable [
    *NOT APPLIED* ] math course" what we routinely use when working with

    I'll consider dismounting a major pet peeve hobby horse.

    PS. If anyone reading this is member of appropriate IEEE committee on
    educational standards, my reply-to is valid and I would be more than
    happy to contribute to improving the situation.

    Anyone get impression that this is a hot button issue for me ;/
  15. CBFalconer

    CBFalconer Guest

    I was about to mention that no reply-to showed on your article,
    when I decided to look at the full headers and discovered it
    there. Yet it didn't show up for normal headers on NS 4.75, while
    my own reply-to does! Then I realized it was identical to your
    'from' address, and concluded that NS had suppressed the non-useful
    additional header display.

    I still keep finding ways in which NS 4.7x is superior to
    Thunderbird 1.5.

    Some informative links:
  16. Genome

    Genome Guest

    It am easy try

    See and I have not even have to be member of committeeee to give you
    appropriate information to solve your problem plus you am not have pay me
    monee for find wrong answer........ Plus I went back and peer reviewed
    myselv so iz correct.

  17. Terry Given

    Terry Given Guest

    Hi Tim,

    dickhead. oops, senior design engineer
    IEEE trans:
    industry applications
    industrial electronics
    power electronics
    control systems (although almost entirely useless)
    yep. nope. did it all from first principles, using a pen and paper. then
    fiddled with resultant parts to optimise both BOM and response.
    a plethora of them. Computer Controller Systems, Astrom & Wittenmark are
    right at the top of the list. Slotine & Li, applied nonlinear control is
    damn good too, but not for the fainthearted. as is macejowskis
    multivariable control. I have a couple of dozen control books, some of
    the oldest ones are brilliant - far less use of brute-force techniques,
    much more emphasis on understanding. Automatic Control Systems
    Enigneering vol. 2, Langill, is fabulous.
  18. On Saturday, in article
    I realise that there has to be simplification, but a lot of theory does
    not USUALLY cover the applied like what to do when an external sensor
    or feedback in the loop says there is a fault and how to make sure the
    loop (and the system) return to the correct state (whatever that may be).

    An example would be printing shop guillotines that usually require two
    spaced out buttons to be pressed at the same time for the whole of the cut
    process. On the basis that if both hands are on the buttons then the
    operator's hands are not in the machinery! Now dpending on how the system
    operates there are basically three safety stop states - stop, move back
    a little or return to start position.

    I have seen some control systems really screwed by an 'abort' function
    that need complete restarts or worse after that, basically because the
    control loop hardware/software could not effectively reset correctly.
    Quite often I have seen customers who do not understand the limitations
    and assumptions in their systems let alone any model they think they have

    Some of the major control loop problems I have seen is applying open loop
    blocks to an overall system that is closed loop. Mainly because they
    did not understand the limits of the blocks they were using, most
    notably delays, determinicity and sampling restraints.
  19. Fred Bloggs

    Fred Bloggs Guest

    TW must be talking about WPI , pronounced Wuss-duh
    in those parts.

  20. Tim Wescott

    Tim Wescott Guest

    Fred Bloggs wrote:

    If it was me it was supposed to be off the record.

    My Thesis advisor once told me that the first time he visited WPI he
    spent about an hour on the interstate -- he was looking for Wu'sta. He
    passed by Wor-ches-ter several times before it sunk in...


    Tim Wescott
    Wescott Design Services

    Posting from Google? See
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day