1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Back emf

  1. Dec 25, 2011 #1
    Hi,

    I have a question about back emf

    So, my understanding of back emf ([itex]\epsilon_b=-L\frac{dI}{dt}[/itex])is that its the counter voltage that's induced by change in current, whether it be from a battery or magnetic field ([itex]\epsilon_0[/itex]). We can model this as an LR circuit where [itex]\epsilon_0 +\epsilon_b=IR[/itex]

    But when we add the back emf, don't we change the value of dI/dt? i.e. if we differentiated the equation we'd be left with

    [itex]\frac{d\epsilon_0}{dt}-L\frac{d^2I}{dt^2}=R \frac{dI}{dt}[/itex]

    And then, the time derivative of current would be offset by an additional 2nd derivative of current (multiplied by negative L). I'm assuming we calculated change in current as [itex]\frac{dI}{dt}=\frac{1}{R}\frac{d\epsilon_0}{dt}[/itex], hence the offset.

    Does this model then only hold for cases where current changes linearly/not at all? And if this is correct, how would we model it for such cases where current changes with respect to 2nd or higher derivatives of time?
     
    Last edited: Dec 25, 2011
  2. jcsd
  3. Dec 25, 2011 #2
    If the resistance in the circuit = 0 then the induced emf,e, must be equal to the applied emf, E,
    This means that e = E = -L.dI/dt
    This means that the current will increase at a constant rate of dI/dt = E/L
    If there is resistance in the circuit then the emf appearing across the L = E - Ir
    This means that L.dI/dt = E - Ir
     
  4. Dec 25, 2011 #3
    Thanks for the clarification. So the back emf and IR must add up to the applied emf.

    My confusion/question still remains though.
    That is, if we apply a back emf don't we change the rate at which current changes over time (if it has a nonzero second derivative)? If so, wouldn't this induce another back emf because the "first" back emf has induced a change in current?

    Would we then model this as a summaiton of some sort? i.e.
    [itex]\epsilon_b=-L(\frac{dI}{dt}-\frac{1}{R}\frac{d^2I}{dt^3}+\frac{1}{R^2}\frac{d^2I}{dt^3}-...)[/itex]
     
    Last edited: Dec 25, 2011
  5. Dec 26, 2011 #4
    I think your confusionis inthe phrase '.......if we apply a back emf.....'
    A back emf is not something that is applied.
    When a battery (emf) is connected to a circuit and the switch is closed the current begins to increase. AN Increasing current will result in an induced (back) emf as a result of any inductance in the circuit.
    The induced emf = -LdI/dt (it is not an applied emf,it is caused by the increasing current)
    If there is resistance in the circuit you are correct to state that Ir + induced emf = applied emf, E.
    E = IR + LdI/dt
    the solution to this equation is
    I = Io(1-exp-(Rt/L))
    After a long time the final (steady)current will be Io and the initial rate of increase of current will be = E/L
     
  6. Dec 27, 2011 #5
    Oh, okay. Yes, that makes sense.

    I feel sort of silly now. The backemf is generated differentially so it's just a simple diffeq to solve for it..

    Thanks!
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Back emf
  1. Back emf (Replies: 1)

  2. Back emf (Replies: 3)

  3. Back Emf Paradox (Replies: 3)

Loading...