Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Low cost EFT generator design

  1. Apr 13, 2017 #1
    I am trying to make a EFT generator in low cost which can generate bursts specify by IEC61000-4-4.
    The circuit which I am working on is given as below. This is for making a high voltage of atleast 400V. I got this design from a paper. But it is not mentioned the values of R1 and R5. I am not getting the expected out put too. Can anybody tell what mistake I am doing? high_v.PNG
    The Pspice files are given in attachment

    Attached Files:

  2. jcsd
  3. Apr 13, 2017 #2

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    What is the rate of current rise through your inductor-resistor combination R1-L1 ?
    What is duration of your on-time ?

    what is "U" and what are those acronyms TD TR TF et al ?
  4. Apr 13, 2017 #3
    pulse source.PNG
  5. Apr 13, 2017 #4

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    Thanks !

    Still you haven't said what is "U" in the TD, TR, and TF terms. If it's short for "micro" why didn't they use μ like they did for PW and PER ?

    I dont know a doggone thing about spice.
    If i just assume they're sloppy typers and "U" also means "micro" ,
    then every 200 microseconds we apply a 50 microsecond pulse of 12 volts through resistance of 2 ohms plus whatever is R of the mosfet ?
    Were the mosfet perfect it'd be zero ohms, so let's assume that for simplicity

    Current in a RL circuit is

    so at end of 50 microseconds I = 12/2 X (1-e-2X50E-6 / .002 ) = 6 X (1 - 0.9512) = 293 milliamps (check my arithmetic please)

    which seems mighty little for a 5 amp transistor.

    From where did your on-time come?
    I think but do not know you probably need more current through your inductor so that when switched off it gives a healthy inductive kick.
    The energy you add to the output capacitor's electric field must first get built up in the coil's magnetic field, ½LI2 is all that's available to get dumped into the capacitor where some of it (hopefully a lot of it) becomes ½CV2 .
    That's how "Flyback" converters work.
    We run into that in automobile ignitions. When the points are set too wide, they do not stay closed long enough between cylinder firings at high RPM for the coil current to rise high enough to make a healthy spark for the plugs. That limits how fast the car will go. You'll hear gearheads talking about "Dwell" which is a measure of what fraction of the time points stay open .

    Since you're simulating, try 10X wider pulse width 1/10th as often and see what happens. Also shorten up fall time to a few nanoseconds.

    Wait a second what's this ?


    V1 is shown as 12 volts adjacent the battery but 5 volts in the tabulation
    and V2 is drawn adjacent a source that looks like a pulse but assigned 0 volts in the tabulation

    You and i know what you meant but computers are sooooooo nitpicky...... better double check what it's up to.

    What does your simulation program do when you lengthen the pulse width and shorten the fall time ?
    Last edited: Apr 13, 2017
  6. Apr 13, 2017 #5
    Yah, U is for micro
  7. Apr 13, 2017 #6

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

  8. Apr 14, 2017 #7


    User Avatar
    Science Advisor

    The user can entered the multiplier in either case. With the old FORTRAN SPICE it was always upper case, but now anything is possible. U and u are both 1e-6. Some spices can replace u with µ if the option is enabled. M and m are both milli = 1e-3, while the three letter MEG, written in any case is 1e+6. A 1uF capacitor will be 1e-6 farads. 1nF is also sensible. But F or f is not always farad. If it comes first, it is femto. So a beginners 1F super capacitor will have almost no capacitance.
  9. Apr 14, 2017 #8

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    Another little experiment, real easy since it's just a computer program

    big mosfets take a fair amount of gate charge
    from its datasheet at

    to squeeze that 63 nanocoulombs through your 470 ohms in 50 microseconds takes how many volts? I get 0.59 , meaning your mosfet gets on;y 4.4 volts of gate drive.


    try raising pulse amplitude to 10 volts?

    old jim
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted