filters

Discussion in 'General Electronics Chat' started by arun2009, Aug 30, 2010.

  1. arun2009

    Thread Starter New Member

    Aug 30, 2010
    3
    0
    hi frnds im new to electronics,kindly answer my questions
    I was doing a lpf and hpf using both rc and opamp. It worked for the desired freq but it is also accepting a wide range of freq outside the desired range, for eg:if i designed a 1khz lpf, output is obtaing for upto 25khz??there is no problem wit my ckt design??can u help me in figuring out d problm???

    Regards
     
  2. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    It's either misdesigned, misassembled or the Q of the circuit wasn't that high to begin with. No filter is perfect, they usually attentuate by a certain amount of dB per octave - per pole of the filter.

    A single pole LPF is usuaully 6 dB/octave.
     
  3. gootee

    Senior Member

    Apr 24, 2007
    447
    50
    What are the input and output amplitudes at 25 kHz?
     
  4. arun2009

    Thread Starter New Member

    Aug 30, 2010
    3
    0
    thanks for your reply..
    I will explain wat i hav done, i used a .1uf cap and 1k resistor..with which i'll get a RC lpf having cut off freq of 1.592k hz. As i was new i tought dat lpf will not pass freq greater than dis, but later i understood that it will reduce the signal amp after 1.5 k hz,dis is wat expected rt?..but for me the signal was getting distorted after 1k hz itself..kindly give some suggestions and corrections if im wrong.
    Thanks and regards
    Arun
     
  5. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    Can't say as I've ever seen such a high value resistor in a crossover circuit - except perhaps in some of the most complicated and expensive ones in the HPF section.

    You're dealing with 8 ohm (or somewhere thereabouts) loads driven by an amplifier with a low output impedance.

    Can you make some sort of a sketch of what you're trying to hook all together and how? Knowing the model # of the speakers involved would be of great help too.

    Given that, I'd be glad to design you a simple, inexpensive yet effective crossover solution.
     
  6. gootee

    Senior Member

    Apr 24, 2007
    447
    50
    Yes. That IS exactly what is expected.

    A low-pass "1st-order" RC filter will only give three (3) dB of attenuation (i.e. 1/2 of the original amplitude) at its "cutoff" frequency, and six (6) more dB (i.e. 1/4) attenuation of the amplitude per decade of frequency (i.e. every time the frequency increases by 10X). That is not usually considered to be a very steep/fast roll-off of the amplitude.

    Regarding the distortion, there could be other reasons too, but yes, a complex signal with multiple frequency components could have different frequency components attenuated by different amounts, and could then appear to be distorted.

    You can make higher-order filters by adding more RC stages in series, probably with unity-gain opamp buffers between the RC stages (OK if you're doing small-signal stuff). But note that there are also much better filter topologies available, with various desirable response-characteristics available for different applications.

    There is an in-depth discussion of filters in two chapters of this rather-excellent free on line book:

    http://www.analog.com/library/analogDialogue/archives/39-05/op_amp_applications_handbook.html

    And you can download the (free) FilterPro software, from http://www.ti.com, to help (very much) in designing higher-order opamp-based small signal filters.

    With the use of higher-order filters, you can get a _much_ sharper after-cutoff slope, i.e. a much-steeper attenuation slope above the corner frequency of your filter, theoretically enabling you to get almost arbitrarily-small amplitudes very soon after your cutoff frequency.

    I believe that national.com also has an on-line web-based filter design application, available for free.

    Cheers,

    Tom
     
  7. arun2009

    Thread Starter New Member

    Aug 30, 2010
    3
    0
    tanks for your reply..as directed by you i have gone thro' all topics and got a gud idea on filters

    Thanks and regards
    Arun
     
  8. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    I might also mention a very interesting feature that JBL incorporated into their earlier automotive high-end audio crossovers and some of the amps they had in the same model line.

    The crosssover frequencies weren't adjustable by the pots you usually see, it was done with a setup that involved 16-pin DIP sockets and the blank 16-pin headers you can buy to plug into them. They offered standard frequency modules or gave you the formula for building your own if you wanted to use specific frequencies.

    In effect, all the resistors involved in setting the active op amp crossover frequency were soldered into those headers so changing the frequency of any particular channel was a simple matter of unpluggging the existing and plugging in the new.

    They have since taken most of the JBL schematics off the web. I downloaded all of them years ago but onto a mirrored drive that lost one side of the mirror and I was never able to read the data out of the other side. I'll bet some of them are still out there though, I'd just have to search through some of the models I knew that had them. They were the famous "White series" amps they made about 20 or so years ago.

    I'll try to find one of them this weekend and post it as a reference to the way they designed the filters but it was the idea of being able to change frequencies on the fly that intrigued me. No stupid pots to guess with, that introduced noise or that changed over time due to extended bumpy road driving.
     
Loading...