Connect with us

peak

Discussion in 'Electronic Design' started by ducran lapoigne, Jul 29, 2003.

Scroll to continue with content
  1. hi all,
    i need schematic for remove the current
    peak when i switch on a 3.5 kva transformer
    anyone can help me ?
    thanks
    antoine
     
  2. Active8

    Active8 Guest

    what current peak? the primary resists fast current changes via the back

    EMF = -L(di/dt)

    you'll get a nice voltage kick on both sides when you cut the switch,
    but i can't calculate that or the current without more info.

    what's missing from this picture? max instantaeneaous voltage to be
    switched in... inductnce of xfmr coils, turns ratio, N ... coupling
    coeficient, k or mutual L, M ... the circuit itself?

    snubber? i.e. cap rated for AC service and a resistor. high current MOV?

    cut the primary voltage off and on either side you'll need to dissipate
    ..5LI^2 joules which is what? .5CV^2 so the cap would need to absorb that
    minus what the R dissipates, then it would send it back to the coil and
    so on until the R has dissipated enough energy to "remove the peak".

    mike
     

  3. Actually there can be a very large current during switch on not directly
    predicted by your nice and clean EMF = -L(di/dt) formula. When and if the
    core saturates the inductance drops like a rock and suddenly massive
    currents start flowing that are usually mostly limited by wiring resistance.

    Why does the core saturate you ask, even though it doesn't during normal
    steady state operation? To answer that I suggest you read an old thread
    about unexpected motor starting current peaks:

    http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&safe=off&threadm=IN8g8.63
    354%24vP.259578%40rwcrnsc51.ops.asp.att.net&rnum=2&prev=/groups%3Fq%3D2000%2
    Bhp%2Bmotor%2Bcurrent%2Bpeak%26hl%3Den%26lr%3D%26ie%3DUTF-8%26safe%3Doff%26s
    elm%3DIN8g8.63354%2524vP.259578%2540rwcrnsc51.ops.asp.att.net%26rnum%3D2

    Although that thread was relating to unexplained large starting currents in
    a motor the phenomena occurs just the same in a transformer. Pay special
    attention to Glenn Smollinger's reply to the thread since he was the first
    to answer the OP's question accurately.


    As far as a solution to this OP's problem... Since at this power level an
    NTC resistor probably isn't very appropriate, I suggest using a power
    resistor in series with the primary of the transformer. Place a nice big
    relay in parallel to the power resistor to switch it out of the circuit once
    the transformer has been energized appropriately. The actual relay coil
    delay control hardware will depend on exactly what voltages and signals you
    have to play with, but a simple RC circuit driving the gate of a MOSFET
    (which in turn drives the relay coil) should work just fine. The delay need
    not be very long, 200ms would probably be plenty (but it is only a guess) to
    prevent core saturation and therefore massive inrush currents.
     
  4. N. Thornton

    N. Thornton Guest

     
  5. N. Thornton

    N. Thornton Guest

     
  6. Active8

    Active8 Guest

    i can't get that link to work. prob the way my news reader mung'd it.
    but i understand saturation to a degree.

    i'd still look at that article, though. what was the subject line on
    that thread?

    mike
     

  7. The thread was titled "2000HP Motor Starting Issue" and was posted in
    alt.engineering.electrical back in the early part of 2002.

    Basically the jist of the situation is this:

    Suppose you have a normal sine wave AC voltage source. Now suppose you flip
    the switch at time t=0 (when the sine wave is at a zero crossing and is
    about to go upwards for a half cycle). The core flux density starts out
    intially at zero (we neglect residual magnetization since it is usually
    minimal), but increases continuously during that posititve part of that
    first half since wave because you are applying positive volt seconds. So at
    the next zero crossing (at time t=180 degrees), you have a substantial
    amount of flux density built up in the core (at least theoretically).
    During the time period between t=180 degrees to time t=360 degrees you apply
    exactly the same amount of volt seconds to the core as turing time t=0
    degrees to time t=180 degrees, except of the opposite polarity. Thus the
    net result is you effectively cause the flux density in the core to drop all
    the way back to zero, its starting value.

    In practice after several cycles the flux will recenter itself about the
    point where the flux density equals zero. That is, instead of the flux
    increasing from zero up to some point, then decreasing from that point back
    to zero, the flux will increase from some negative value, up past zero, up
    to some peak positive value, and then back down through zero and finaly back
    to some negative value again. The total flux density swing of the cylce
    will still remain the same, but now the peak flux density in either
    direction need only be half in peak magnitude from what it was when the flux
    swing wasn't centered about the core zero flux density point.

    In more practical terms in a real world situation, suppose you have a core
    material with a saturation flux density of 1.5Telsa (in either the negative
    or positive direction of flux). If the flux density swing is assumed to be
    centered about the point where the flux density is zero, then the peak to
    peak flux density swing that can be used is 2x1.5T=3Tesla, without causing
    the core to saturate. The only drawback is if you design your
    transformer/motor core to experience a flux density swing of 3Tesla, then it
    could easily saturate during those first few cycles until the flux density
    swing gets centered about flux density of the core equal to zero, provided
    that the user flips the switch at a zero crossing of the mains cycles
    instead of a peak of the mains cycles. If the user flips the switch at
    exactly one of the peaks of the mains cycle, then no core saturation will
    occur as the flux density swing will be centered about zero. In theory if
    the user flips the switch on at anywhere but the peak of the mains cycle the
    core will saturate, but the effect will by far be the worst at a mains zero
    crossing.

    The OP doesn't really need to use resistors to prevent the massive turn on
    surge current due to core saturation. Instead he could use some AC
    switching gear (most like a big TRIAC) and just make sure to only turn the
    power on at exactly (or very near) to any peak of the mains cycle.
    Unfortunately this isn't necessarily appropriate if your load also has a
    significant capacitive component, because the capacitor inrush current will
    be maximum if you do this. If the capacitor is too large, it could possibly
    destroy the TRAIC. Capacitor inrush current isn't likely to be as big of a
    deal as the core saturation effect overall however, since the capacitor
    stops taking so much current in less than one cycle (and usually amounts to
    only around a joule of total energy for normal sized capacitors), whereas a
    saturated core may take several cycles to fully settle down (meanwhile
    consuming very large current at full mains voltages thus dissipating perhaps
    hundreds of joules almost all directly in the resistance of the transformer
    primary and site wiring).
     
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

-