No announcement yet.

Test Crossover/ Calculator.

  • Filter
  • Time
  • Show
Clear All
new posts

  • Test Crossover/ Calculator.

    Thought I would try the calculation way before WinPSD. Basically a Bessel first order at 275 Hz. Then a 2200 Hz Bessel 3rd on the tweeter. Well I was wondering. Does the 3rd order have to be in the mid? Can it be removed to a first and keep the 3rd order on the tweeter? The online calc said to add it this way. This was my design for the HDS as a woofer. 4 ohm SDS as the mids. I took the db down to 85 db for the mid/tweet because of woofer baffel loss(space loss) I know a baffel step would also be needed. Supposedly.

  • #2
    Not only baffle step, but you also have no consideration for phase, impedance, or cone breakup. This is why no one here will recommend or endorse a cross over based on an online calculator. Just too many things that are not accounted for.


    • #3
      Ok, thx for the advice. I'll skip onto WinPCD soon.


      • #4
        OK, so I plugged your numbers into a modeling program. Here's how you did.
        The woof DOES roll off (at -3dB/octave), but it actually rises +4dB between 2.5 and 4.5kHz (due to cone breakup). It actually DOES cross to the mid (pair) near 300 Hz.

        The mid (pair) is padded about -6dB, but only ends up running near 86dB, not TOO bad of alignment w/the woofer (at 87dB), but yields basically NO baffle-step.

        For some reason (too much series resistance in its L-pad?) the tweeter ends up attenuated -22dB, which: causes it to cross from the mids near 4kHz, but it's much, much too quiet.

        Overall, your response is rather flat from 200 Hz (invalid data below that point, but THAT will depend on your box alignment) up through about 1.5kHz, then drops at -9dB per octave, so it's down around -20dB @ 10kHz (still dropping).

        I can't say that your notch or Zobel on the tweeter do ANYthing. Generally that kind of notch only needs to be used on a NON-ferrofluid tweeter (the Silkie HAS FF), and a Zobel isn't typically used on a tweeter (but your mids may benefit from one). Also, there's no reason to start the tweeter's high-pass filter PAST the mid's 72uF cap. Its own 4.5uF series cap makes the 72uF cap (in the TWEETER circuit) inconsequential. This cuts the number of elements in the tweeter filter down from 11 to only 5.

        Spoiler alert !

        So, if you toss the notch and Zobel on the tweeter, and lower its L-pad series resistor down from 33 to 6 ohms, it ends up running around 83dB, w/a -3dB down pt. near 3kHz. BETTER YET, just go 2nd order w/an "L": 5uF series cap, 0.80mH shunt coil, for L SR=6n (ohms) and PR=10n.

        On the woofer I'd just up it to 2nd order w/a 6mH (low DCR - iron core probably) series coil, and a 50uF (npe) shunt cap (to gnd).

        For the mids, I'd suggest a 2nd order bandpass + a Zobel and L-pad: 33uF series cap and 0.90mH series coil (#20 OK), then THREE shunts: a 4uF cap, a 3.0mH coil (high DCR ok here), and a 10ohm+10uF Zobel, lastly the "L": SR=7n and PR=6n. You also need a notch ("tank") cap across the smaller (0.90mH) coil. It's a very small value cap, a 1.0uF.

        Runs around 82dB w/full BSC. Attenuations: mids -10dB, Silkie -7dB. Fcs around 450 and 2.2kHz. Hardly ever drops under 8ohms.

        Off you go.


        • #5
          Thanks for taking the time. I can't run the software for a few days! So the mids do run 10 db higher. I used the calc. It said no gain. 5 or 6 wouldn't be so bad. The woofer is only 1 db less, (in anechoic) well actually almost 2 by the lower response frequency graph. Ok, will see..
          Ya I realized shortly after the Impedance correction circuit was not needed.


          • #6
            I know what u saying!
            I'm figuring on a higher return from these woofers due to room placement/effects.. I do have a boost also. I have an eq in my amp.
            So judging a 3db loss for room. Gain from my eq amp. There's no reason I shouldn't hit a sensitivity of 85-88 in room bass response.

            I have to check baffel step. Crossed over low. I'm wondering if the baffel step compensation goes on the mids?