Connect with us

Rogowski sensor questions...

Discussion in 'Electronic Design' started by Tony, Sep 11, 2004.

Scroll to continue with content
  1. Tony

    Tony Guest


    Rogowski sensors are effectively toroidal current transformers wound
    on an air core, commonly used in electrical power transmission and
    distribution systems. The output (after integration) is a good
    representation of the current flowing through the device, without the
    limitations of an iron-cored CT at the low end and the high end. But
    for accuracy, the winding must be ultra-accurate. Well that's my
    understanding of the literature, anyway.

    Has anyone seriously played with this technology? How critical is the
    winding geometry? What happens if the geometry isn't perfect? Can
    anyone in Australia wind them cheaply (I'm hoping cheaper than a cored
    CT, since there's no actual core)? Any gotchas to be aware of?

    Tony (remove the "_" to reply by email)
  2. Tony a écrit:
    At one of my customers, I saw this kind of current transformers made on
    a PCB.

  3. Tony wrote...
    Rocoil Limited has some nice Ward sensors,
    They make the point that having highly-uniform windings is necessary
    to reduce field pickup from other sources. Lack of precision is a
    commonly-mentioned problem, but Analog Devices includes an integrator
    in their precision AC energy-metering ICs, like the ADE7758, to use
    with Rogowski coils, so clearly precision sensors can be relied upon.

    Rogowski coils are not very insensitive, and are best suited for high
    currents. They can be mass-produced with printed-circuit boards, e.g.
    and US patent # 5414400, mentioned by Spehro Pefhany, etc.
  4. legg

    legg Guest

    The Rogowski coil sees use in lower sensitivity higher frequency
    situations where mechanically rigid toroidal sensors are impractical,
    or where the pulse amplitude of the sensed current is higher than
    conventional CTs can maintain.

    Typically, once wound, you can calibrate the integrator for accuracy
    at levels for which standard reference CTs are suited, with reasonable
    reasurance that the higher amplitudes will maintain this calibration.

    The only time that the coil's construction could be considered as
    critical, is if no calibration is permitted.

    Being air-cored, they are potentially more physically fragile unless
    wound on a former - therefore flexible types are often calibrated
    after being mounted into the application position.

  5. legg

    legg Guest

    Construction of simpler Rogowski structures is discussed in IAS 2000

    "High Performance Rogowski Current Transducers"
    W. F. Ray, C. R. Hewson
    Proceedings IAS 2000 pp 3083

    and air-cored or printed structures are discussed generally in EPE'99
    by various authors

  6. Marc H.Popek

    Marc H.Popek Guest


    they don't have to be strictly air cored. IN fact we make and use these
    puppies all the time to measure medium speed < 50 nS events in impulsed
    pulsed power applications. Just take a piece of coaxial cable, wind a
    "symmetric coil about the dielectric coated center wire and bring the return
    through the central wire core. wrap the spiraled coaxial cable about the
    area that has surface current running in it, and you have a vo = dB/dt.


  7. Tony

    Tony Guest

    Thanks Rob - once again - for the good references. I get the feeling
    that the effects of coil irregularities may have been exaggerated in
    some of the literature I have read - uniformity is certainly
    important, but no more so than one would expect.

    Tony Roe

    Tony (remove the "_" to reply by email)
  8. Marco Popek wrote...
  9. Uniformity is very important if the device is to have its sensitivity
    calculated to high precision. If it can be calibrated against another
    device, uniformity is not critical.
  10. I read in that Winfield Hill
    I think an important step was left out. 'First, strip off the
    shielding.'! Winding an inductor on a continuous layer of shorted turn
    is not a productive experiment.

    Long years ago, I visited a loudspeaker manufacturer and found on the
    production line that they were winding the crossover inductor on the
    metal-cased crossover capacitor. That's what happens when you leave
    Purchasing to decide which part to buy!

    Many years later, I recounted this story to the Chief Engineer of
    another loudspeaker manufacturer. Guess what we found half an hour later
    on the production line!
  11. John Woodgate wrote...
    Uniformity is _critical_ for good rejection of external fields.
    One needs to create equal contributions to the coil voltage,
    independant of the field location on the coil. This it the way
    uniform external fields can completely cancel.
  12. Tony

    Tony Guest

    I'd be surprised if a little non-uniformity in the windings varied the
    calibration constant appreciably - even if the winding is bunched up
    then spread out, each turn would still see the same flux, so I wasn't
    too concerned about that.

    And as I considered the ONLY source of flux around would be the
    conductor I'm measuring, interference rejection didn't seem to be an
    issue. But as things progress that would appear not necessarily to be
    the case - there would seem to be advantages to me in being able to
    reject the flux from similar nearby parallel conductors by 60dB if
    possible (although less would still be OK). I can't imagine how to
    quantify the rejection-vs-bunching factor, so I guess I'll need to
    make some and test them.

    Tony (remove the "_" to reply by email)
  13. Tony

    Tony Guest


    I've seen a pic something like you describe in a paper. But it
    intrigued me how one would wind a precisely uniform coil on a springy
    flexible former? Maybe if you had a machine that makes bass guitar

    Tony (remove the "_" to reply by email)
  14. legg

    legg Guest

    Major rejection of the structure is intended to be provided by the
    return path of the coil's conductor, shadowing the body of the coil.
    This is a feature often missing from pcb-printed coils that I've seen
    attempted. This can also form the body of a slotted shield. The
    Rogowski isn't simply an air-cored torus.

    The diameter of the winding is expected to be smaller than the sensed
    cross-sectional area being enclosed by the torus, in most situations.
    Under those conditions, winding irregularities, saddling and even
    overwinding tend to have little relative effect.

    Calibration-free sensors for accurate measurement are pretty rare.
    Consider the number of practical instances, however, where a 20%
    tolerance is considered usable - as in most overload limiters or loop
    control situations. Consider the alternative problems encountered with
    coaxial shunts in obtaining wide bandwidth accuracy.

    The Rogowski has it's application where weight, flexibility and pulse
    amplitudes make rigid cored devices or shunts unsuitable. Trying to
    make it do everything is probably a waste of time.

  15. legg

    legg Guest

    Consider also a circuit printed on flexible strip, looking somewhat
    like a watch strap.

  16. Marc H.Popek

    Marc H.Popek Guest

    Ill post a picture of the coaxial cable based rogowski coil.....
  17. Marc H.Popek

    Marc H.Popek Guest

    Yes John, the coil portion is wound on the Teflon insulation about the
    center conductor and then the return to the BNC is made to the center
  18. Marc H.Popek

    Marc H.Popek Guest


    The person I learned from wound as follows. the stripped coax is suspended
    from the ceiling, the wire to be wound about the center conductor and
    insulation is clipped with hemostats, the hemostat is sent flying about the
    center, and spirals about forming a pretty good symmetry, the coils are hand
    tweaked and tightened, some times for long rog coils > 2 feet active
    portion, the wrap is covered by shrink wrap to secure the wrap. Do you want
    one or two?

  19. Marc H.Popek

    Marc H.Popek Guest

    for 60 dB rejection of common mode the rog coil will fall short, you need a
    "B-dot" sensor and again for anything like 60 dB rejection of CM e fields
    you need the newest from a differential b-dot...

    for more detail what is the size of the structure you wish to meter? what
    is the range of the peak currents and the voltage that drives the flow?
    what is the frequency band (or effective band for impulses) you expect to
    see. With these metrics defined the sensor you need is designed by

  20. Well, I suppose it's the same technique. Put the former under tension,
    so that it's rigid enough to wind on. Rich Americans would put the
    former in chucks in the head- and tailstock of their lathe, and wind the
    tailstock back a bit. (;-)
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