How to approach filter design

Discussion in 'The Projects Forum' started by urb-nurd, Mar 6, 2015.

  1. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    After many cycles of errors and smoking components, i am finally at the stage in my project where i am ready to design a filter to convert my chopped AC waveform into a smooth variable amplitude sinusoid.

    I was wondering if anybody had knowledge of any good places to educate ones self on this application.
    ( i apologize for my apparently lazy question, I have looked but have come across nothing too helpful)

    I am dimming a bulb using PWM at 25 Khz and intend to filter this to eliminate some EMI problems stemming from switching 10+ amps at 25Khz.
    Additional information:
    Line frequency = 50Hz
    Switching frequency = 25 Khz
    Bulb voltage = 25V rms

    Any guide is appreciated as always.
    Thanks AAC!
     
  2. crutschow

    Expert

    Mar 14, 2008
    13,039
    3,243
    If the EMI is being conducted through the mains then you could add filters at the mains power connection. Inductors in series with each main's wire and/or a common-mode choke in series with both main's wires should help. Or an inductor in series with the lamp should also help. The required inductor values are likely best determined experimentally.
    A commercially made EMI line filter should also help.
     
  3. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    i also just realised i created a thread about this a few weeks back that i totally forgot about.

    I downloaded and played around with the filter design tool mentioned in my other thread, though i am a bit overwhelmed by the volume of input parameters.
    [​IMG]
    The image above is the first result i got from the filter design tool.
    The result is attached and seems to filter out the PWM effectively, though i genuinly have not much of a clue when it comes to this field.

    When you suggest an EMI line filter, is that to reduce EMI going back to the lines or is that for filtering the PWMed waveform?
     
    Last edited: Mar 7, 2015
  4. Papabravo

    Expert

    Feb 24, 2006
    10,152
    1,793
    What level of mathematical sophistication do you bring to the table? Our ability to help depends on your answer.
     
  5. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    likely the minimum that is necessary for an electronic and electrical engineering student.
     
  6. Papabravo

    Expert

    Feb 24, 2006
    10,152
    1,793
    Filter design starts with an accurate specification. For example: How much attenuation in dB is required in the stopband? How much ripple is allowed in the passband? What is the bandwidth of the transition band?
    From these and other specifications we can select a type of filter, it's complexity, and it's component values.
    A copy of Van Valkenburg, Analog Filter Design, will be most helpful to you.
     
  7. MrChips

    Moderator

    Oct 2, 2009
    12,449
    3,364
    So, are you comfortable with calculus and Laplace transform?
     
  8. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    To a degree. Meaning i have done Laplace transforms in the past 2 years - but couldn't tell you about them off the top of my head.

    This is where i fall short. i am not sure about the attenuation, nor the ripple.

    My intention is to try and recover a sinusoid from a chopped waveform.

    I am using PWM to chop a 50hz sine-wave, and i would like to filter this chopped waveform to provide a solid sinusoid to my halogen bulb.

    I read a paper on the development of a dimmer for AC leds which gave me the idea to employ filtering in my application also
    [​IMG]

    Unfortunately the paper didn't cover the design of the filter, but focused on the testing of the dimmer circuit.
     
  9. MrChips

    Moderator

    Oct 2, 2009
    12,449
    3,364
    Does the halogen lamp care whether the signal is sinusoidal or chopped?
     
  10. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    I believe the mechanical integrity of the filament to be bettered with less sharp rises in current.
     
  11. Papabravo

    Expert

    Feb 24, 2006
    10,152
    1,793
    I guess I don't see the point of chopping the waveform and then trying to recover the waveform. I think of filters as methods of information recovery or elimination. Power conversion is something else. Would a variable transformer accomplish your purpose?
     
  12. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    I think that switching high currents at 25Khz could be quite problematic with regards to EMI.
    I am using a configuration very similat to that pictured above - but my load is of much higher power than the LED's
     
  13. crutschow

    Expert

    Mar 14, 2008
    13,039
    3,243
    Without know the level of EMI suppression you need, then it's all just a guess.
    I suggest you try a common-mode choke that is rated for the maximum current you have with as high an inductance as feasible. Try 0.1μF to 0.47μF for the filter capacitors.
    For further suppression you can add an inductor in series with the lamp.
     
    urb-nurd likes this.
  14. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    Thanks guys.

    Ill look into this some more today and possible be back if i get stuck .

    I will also be asking lecturers at my university so i can hopefully return with a good concise explanation of what it is am really need.
     
  15. urb-nurd

    Thread Starter Member

    Jul 9, 2014
    269
    3
    I have been playing around and managed to get a filter that appears to work, however its feasibility in the real world is where i am unsure.
    Filtercircuit.PNG

    The circuit shows i am using a capacitor and inductor with my load (bulb).
    Filteredwaveform.PNG

    The waveform shows that the load waveform is clearly filtered and a nice enough sinusoid.

    However, the supply and capacitor currents appear to be alarmingly high pulses of up to 260A.

    What should i make of this?

    EDIT: this arrangement seems to not work, altering the duty of my PWM does not vary the amplitude of the sinusoid as i thought it would.
     
    Last edited: Mar 10, 2015
Loading...