Connect with us

UPS output waveform

Discussion in 'Electronic Components' started by kell, May 16, 2007.

Scroll to continue with content
  1. kell

    kell Guest

    I got a APC Smart-UPS 700 that somebody was throwing out. Works, only
    it had bad batteries. The question: what output does it have --
    "modified sine", or sine? A quick lood at the APC web site didn't
    reveal the answer, though it says not to run laser printers, so I
    suspect modified sine.
  2. I never understood that term 'modified sine' as a substitute for sine. It
    sounds like garbage. If you want a sine, and HAVE a sine to modify, the
    last thing you do is modify it.

    What they probably mean is a rectangle wave that has three states,
    positive, zero, or negative, and is switched between these with a timing
    that makes a waveshape which tries to approximate (very crudely) a sine.
    It's not a bad idea, as it halves the voltage step change that would happen
    if a square wave was used, is just as efficient, and is probably easier to
    filter. I've used one on various things that have their own switchmode
    PSU's without trouble, including a lab DPSS laser, but it's always a risk
    trying it with something untested.
  3. All of my Smart-UPS units output a sine wave. Back-ups are square wave.

    - Mike
  4. kell

    kell Guest

    Thanks for the info.
  5. James Sweet

    James Sweet Guest

    Plug in something with a transformer to it, like one of those big heavy wall
    warts or a stereo reciever. Modified sine wave will normally make the
    transformer buzz much louder than usual, while a pure sine wave will cause
    the normal soft hum you get plugging it into the wall.
  6. Arfa Daily

    Arfa Daily Guest

    A 'modified sine wave' is not a three state square wave, although you are on
    the right track. It is actually what you might describe as a "poorly
    digitised" sine wave - that is each of the half cycles of the wave is made
    from two or more voltage steps. It looks on a 'scope like the result you
    would get if you ran a pure sine wave through a 2 bit - (not the american
    phrase for something that's crap, but maybe that 's appropriate as well !! )
    & sign A-D converter. It is a common waveshape for the output of
    transformerless DC-DC power converters, and is called 'modified sine wave'
    because the voltage steps, whilst being clunky and crude compared to a true
    sine wave shape, never-the-less occur with a timing and voltage position,
    such that you could draw a true sine wave through the mean of them.

    My UPS 700 has a dirty great transformer in it, and as far as I am aware, it
    outputs a true sine wave, as did all of the UPS's in this series.

    Incidentally, I think that I have now gotten to the bottom of my one's
    intermittent failure when it either self-tests, or needs to actually provide
    backup power. As several people have suggested, it seems to be the spade
    connectors inside the thing - in my case, the ones to the batteries. When I
    obtained it, it had been standing for a long time, and the gel cells were
    knackered. The replacements that I bought had slightly smaller terminals,
    and it seems that the spade connectors don't make well to them, and when the
    converter kicks in, and tries to draw the many amps from the batteries that
    you forget that these things pull, the connections just can't cope. I
    actually watched the voltage at the back side of the connectors drop to
    about 22v, whilst the voltage at the actual battery terminals remained up at
    almost the nominal 27.4v that you normally read there. The unit then
    immediately detects a battery problem, and starts whistling furiously at
    you, and flashing its lights, including the 'battery fault' one.

  7. Thankyou. I agree. I was half asleep not wanting to wake up, and dreamed of
    stuff that I want solved, as I often do... I realised that two schmitt
    triggers (one each side of zero volts) with thresholds set to respond to a
    small sinewave signal might be an easier and cheaper way to get the
    waveshape required, than to model it with fast timers/dividers, so a true
    modified sine would be made. My error was in assuming that the sine was
    wanted, but in this case it isn't, because switching for efficiency is more
  8. Guest

    The question: what output does it have --"modified sine",
    It is most likely a modified sine wave unit.
    This doesn't have anything to do with the type of output coming from
    the UPS. A laser printer's fuser module will trigger sudden high
    current demand when it turns on. This surge will be hard on the
    inverter and may exceed the capacity of the UPS. Even if the UPS can
    handle the load, the operation of a laser printer's fuser will shorten
    battery runtime quite drastically.

  9. Guest

    The circuitry to generate a true sine wave is probably more expensive
    than circuits to produce an acceptable approximation of one.

    More expensive UPS units have true sine wave generation capability. I
    have an APC Smart UPS 1500 that is equipped with such circuitry.

    I've never had problems with a modified sine wave UPS, but some
    switchmode power supplies start "singing" when running from one. When
    attached to a UPS that outputs a true sine wave, these power supplies
    stay quiet, even when on battery.

  10. wrote in
    I thought so too, at first, hence my remarks about making one just to throw
    away its virtues, but that might not be the case. A Wein Bridge made using
    something other than a filament lamp might make a slightly inferior sine
    wave, but in this case that doesn't matter, and the frequency is fixed so
    you don't need expensive tuning pots, or have to worry about amplitude and
    shape changes while tuning changes. That means a very small cheap sine osc,
    which crossing zero and driving a Schmitt triggers on each supply pole, can
    be cheaper than trying to model the timing with faster oscillators and
    dividers and the logic required to set the timing of transitions. It
    certainly uses less parts, and has a higher accuracy despite needing only
    two voltage dividers to set thresholds.
  11. GregS

    GregS Guest

    Modified sine is somewhere between square and sine. A complete square wave
    has the disadvantage of not having enough voltage for lamps or electrical equipment.
    If you make it 120 VAC RMS and its a square wave, then the peak voltage is much less
    than a 120vAC RMS sine wave. You can make a stepped sine wave using changes on the transformers
    taps, which is efficient. You can also make a sine wave using switching technology. Making
    a sine wave with a linear amp makes a lot of heat.

  12. (GregS) wrote in
    It does... This interests me, how is a switching system arranged to get
    close to a sine shape without the linear amp? I can imagine that as you can
    derive switching points from a Schmitt trigger/comparator following a real
    sine signal to get two poles around zero to get the kind of wave that
    Statpower use in their inverters, using two main switching slements, you
    could have more than two switches to get finer resolution, but at what
    point does it become too awkward, and how many switches are optimal for
    close-to-sine output at reasonable cost and high efficiency? And is it
    humanly possible to make a more lengthy and cumbersome sentence than the
    previous one?
  13. James Sweet

    James Sweet Guest

    Generating the true sine wave is easy, that's not the problem. The
    difficulty is in efficiently amplifying it to drive the load.
  14. David

    David Guest

    Here is how a 120 volt true sine wave inverter using a high frequency
    switching supply works:

    You make a high frequency transformer isolated switching supply that can
    generate up to 150 volts or so and filter the supply so that attenuation at
    120 Hz is minimal. You modulate the regulator with a full wave rectified
    sine wave at the desired output frequency. The switching supply will output
    a voltage which looks like full wave rectified line voltage. You use MOSFETS
    in an 'H' bridge configuration to commute this into plus and minus voltages
    for the load on alternate half waves. You can also have two switching
    supplies with one supplying the positive half and the other the negative
    half. This simplifies the commutation circuitry.

  15. I think I get it, you mean it uses the high speed switching to allow small
    filter caps on the power output? I guess if so, this is also the basis for
    those amps for PA that do something similar to track an audio wave with
    arbitrary signals in it.
  16. David

    David Guest

    Yes, The switching frequency is 100s of KHz and small caps are used. You
    actually need a small load on the output to get the desired waveform.
    The so called class 'D' audio amplifiers use a form of pulse width
    modulation with positive and negative pulses followed by a low pass filter.
    Inverters can also be designed this way, but the ones I have seen use the
    method described earlier.

  17. Clint Sharp

    Clint Sharp Guest

    No idea on the waveform type I'm afraid but the laser printer thing is
    because it just isn't powerful enough, some fusers and motors in laser
    printers can draw *lots* of current. The APC I have has a label that
    states no laser printers on less than a 1500VA UPS.
  18. w_tom

    w_tom Guest

    Modifies sine wave' is an expression so that those who know only by
    feeling will then 'feel' they know reality. The answer to your
    question is in numerical specifications - especially its THD number.
    Without that number, anyone can subjectively claim a modified sine
    wave and a sine wave as same. Only other way to answer your question
    is using an oscilloscope. A neat little trick, using a 'big
    transformer' appliance, will identify the 'worst of the worst'. But
    still, some modified sine wave outputs may not create detectable noise
    in that test.

    Three different procedures to answer your question. Without
    underlying 'whys' or numbers, no reply has provided a reliable answer.
    Any answer without the 'whys' or numbers should be considered

    Meanwhile, what are you trying to accomplish? If UPS batteries are
    bad, then the entire discussion of modified sine wave is moot. What
    is the original objective? What are you trying to solve?
  19. kell

    kell Guest

    I see. Thank you.
  20. Arfa Daily

    Arfa Daily Guest

    Boy, I'm detecting a lot of negativity there ! I thought that most of the
    answers and discussions that have gone on in this thread so far, dealt with
    what we are talking about, pretty neatly and comprehensively. The original
    question was quite straightforward, and the basic answer is quite simple,
    not really requiring an in depth examination of THDs or any other numbers.
    The term "modified sine wave" is regularly used by UPS and other inverter
    manufacturers to describe what is readily understood by everyone to imply a
    digitally derived approximation to a sine wave, with rather less steps than
    would be desirable in an ideal world. How many steps, and what those steps
    represent in terms of distortion figures relative to a true sine wave, is
    neither here nor there in understanding the concept of the term.

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