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

Question on filter

  1. Apr 24, 2009 #1
    I have built a LC parallel tank circuit, and I am having a problem with it. I have a ground that is placed between two polarized capacitors in series. With the positive leads grounded. And an output on the opposite of the input. The problem I am having is when I connect the ground the input signal shorts out to the ground, and I get no signal on the output. And when I disconnect the ground connection the circuit doesn't filter out the unwanted frequencies.
    Can someone please let me know what I'm doing wrong.
  2. jcsd
  3. Apr 24, 2009 #2


    User Avatar

    Staff: Mentor

    Could you please post a sketch or schematic? You are trying to make a bandpass filter with the parallel LC circuit? It doesn't sound like you are using the polar capacitors correctly, for one thing. What voltages (DC and AC) comprise your input signal?
  4. Apr 24, 2009 #3
    The signal is AC. Also I don't know how to post a sketch. I have one. Can someone tell me how to post it.
  5. Apr 24, 2009 #4


    User Avatar

    Staff: Mentor

    If you can put it into PDF or JPG format, you can upload it as part of your post. In the Advanced Reply window, click on the paper clip to go to the Upload Attachments window.

    If you can scan it to a JPG, or if you have a soft copy and can write it to a PDF (like with the free writer PrimoPDF), then you can upload it. You can also upload it to an image hosting website like ImageShack, and post a link here to it.
  6. Apr 24, 2009 #5
    circuit diagram 315png.jpg
  7. Apr 24, 2009 #6


    User Avatar

    Staff: Mentor

    The pic helps, thanks.

    That's not any filter topology that I'm familiar with. Where did you find it?

    More typically, you will use either a "T" or a "PI" filter configuration. For a T configuration, you would put two inductors in series, with a capacitor to ground between them, and the left end of the top of the T is the input, and the right end is the output. You can also skip the right inductor, and just have an LC lowpass filter.

    If you have capacitors across the top of the T, with an inductor to ground in the middle, that has a different transfer function compared to the first T I described. BTW, the output impedance of the "input" signal, and the input impedance of the circuit connected to the "output" end, both influence the transfer characteristic of the T or PI filters.

    If you mean to have a parallel LC resonant circuit as part of a simple filter, you could ut the L and C in parallel (drawn vertically) as the vertical part of a T filter, and use resistors for the input and output coupling to it (as the left and right sides of the top of the T).

    Here are some images of PI filters: http://images.google.com/images?hl=en&rlz=1T4GGLL_enUS301US302&um=1&sa=1&q="pi+filter"&aq=f&oq=

    and of T filters: http://images.google.com/images?sou...&q="t filter"&um=1&ie=UTF-8&sa=N&hl=en&tab=wi

    Oh, and showing the caps as polarized is also an error, unless your signal has a negative DC bias voltage greater than its peak amplitude. You should be using unpolarized capacitors, unless a DC bias is an explicit part of the circuit.
  8. Apr 24, 2009 #7
    the diagram is a resonant tank circuit. basically the same as an old fashion radio tuner. Meant to discard unwanted frequencies to the ground and allow the resonant frequency to pass. So should I be using two non-polar capacitors in series with a ground?
  9. Apr 24, 2009 #8
    Idea04 filter is set up for a negative voltage DC filter for DC with ripple. Use either a resistor or inductance for input to this filter.
  10. Apr 25, 2009 #9


    User Avatar
    Science Advisor

    The effect you are getting is quite normal.

    That is a Pi network and needs to be near resonance for it to do anything useful.

    If you are using polarized capacitors, they are probably too big for RF applications.
    Polarized capacitors are usually more than 1uF which is far too big for RF.
    They also don't work well at RF. The input capacitance in your network is probably almost a short circuit at RF so that is why your signal vanishes.

    Pi networks are used for impedance matching and they have to have exact values of inductance and capacitance to work properly.

    To find the inductance of the coil you have, try putting a known capacitor (say 1000 pF) in series with it and measure the RF voltage at the junction of the coil and capacitor. At some frequency there will be a large peak in RF voltage.
    Then use the following formula to find the inductance of the coil.

    Resonance formula:
    .........................................L = 25330 / (F^2 * C)

    .................................where L is in uH, C is in pF, F is in MHz

    If you just want to drive your circuit from a signal generator you can use two broadcast radio tuning capacitors as the capacitors. Fixed capacitors in the range 100 pF to 1000 pF will also give results. These will normally be ceramic capacitors.

    As a test, and assuming you have an RF signal generator covering at least 3 to 30 MHz , you could put 1000 pF as the input capacitor, 20 turns of wire on a piece of plastic water pipe as the inductor, and a variable capacitor from a broadcast radio as the output capacitor.
    You should be able to peak up the output at a range of frequencies by varying the output capacitor.
    Last edited: Apr 25, 2009
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook