hi ,
I am making a active 3 way system with 4th order linkwitz riley filters , which has the mid range playing from 120hz to 2500hz and then the tweeter takes over , I have modeled the mid range in a sealed enclosure with a Otc of 1.2 will this high of Otc create issues ??
https://diyaudiocart.com/Raw-Driver...d-High-fidelity-8-Ohm-6.5-inch-Woofer-S16NI-8
I am making a active 3 way system with 4th order linkwitz riley filters , which has the mid range playing from 120hz to 2500hz and then the tweeter takes over , I have modeled the mid range in a sealed enclosure with a Otc of 1.2 will this high of Otc create issues ??
https://diyaudiocart.com/Raw-Driver...d-High-fidelity-8-Ohm-6.5-inch-Woofer-S16NI-8
It will ring around 120, I'd increase the enclosure size or add a lot of stuffing to get it down somewhat lower. The ringing may not be a big deal, a little extra "kick" or warmth subjectively, but I prefer a more damped performance in general. Of course, that's all disregarding the XO, and you're not providing much detail here. What's the Fc? What driver? Active or passive XO?
OK, so with an active XO at 120hz, you're going to have a combined 6th order eventual slope pretty near the passband. To keep the filter performing as expected, you definitely want to push that Fc downward along with the Qtc as much as possible, and potentially move the XO point up a little bit if it's variable. With a 4th order active your best results are generally with a decent amount of flat(ish) clean response on either side of the filter. That's actually true of all XOs, and 4th is pretty steep, but you don't want an extra 12dB filter right next to your XO point.
Putting a driver in a box resulting in a high-ish Qts is not such a major concern. This is because the driver-in-box has a second order high pass response (around resonance). By adding another second order high pass filter you can create an overall 4th order response. By choosing the Q of the additional filter correctly you can get a smooth rolloff with no ringing. At 120Hz you would probably want to use an active crossover, but a passive crossover can be used in the same way.
Take a 4th order Butterworth alignment. This consists of two second order sections that are cascaded (connected in series) both having the same "corner frequency". One section has a Qts 1.31 of and the other has a Qts of 0.54. It does not matter which is the "driver" and which is the "electrical filter".
In your case you only have Qts=1.2, so you will need to use a filter with Q around 0.6. It's not super critical, and you could even use the Q to tune the knee and crossover to the sub/woofer to your desires.
Even in cases when the driver Q is somewhat high, keep in mind that it's the combination of driver and electrical input that govern the overall response produced by the driver. This is why you can use an additional filter to eliminate any ringing in this situation.
Take a 4th order Butterworth alignment. This consists of two second order sections that are cascaded (connected in series) both having the same "corner frequency". One section has a Qts 1.31 of and the other has a Qts of 0.54. It does not matter which is the "driver" and which is the "electrical filter".
In your case you only have Qts=1.2, so you will need to use a filter with Q around 0.6. It's not super critical, and you could even use the Q to tune the knee and crossover to the sub/woofer to your desires.
Even in cases when the driver Q is somewhat high, keep in mind that it's the combination of driver and electrical input that govern the overall response produced by the driver. This is why you can use an additional filter to eliminate any ringing in this situation.
You could design an active filter that has a pair of complex zeros covering the poles of the midrange loudspeaker, two zeros in the origin and four poles at the desired Linkwitz-Riley locations. You then get a second-order asymptotic electrical roll-off with a bit of a wiggle that ensures that the combined response with the loudspeaker is fourth-order Linkwitz-Riley.
I have another driver that gives a fc of 70 ish with a qtc of 0.9 , I guess that will be a lot better for my application , I cannot move the crossover any higher as my sub is integrated in a side firing manner I need to keep the cross over point as low as possible...120 is still high but I have read at many places people are happy even with 150-180hz
If the midrange HP filter has a Q of 1.2 and the driver is placed on a typical 6" wide baffle, the system is not going to ring. The excess energy in the lower end of the midrange passband is going to shunt around the sides of the baffle and it will have flat response, not in the nearfield, where it matters. If you design the system by looking at the driver, enclose and filters, but discount the baffle behavior, the response will be very droopy.
i have a 200mm baffle and intend to use a baffle step correction with 3db at around 550hz ....
IMO the concern is not so much about "excess energy" but rather the time domain response of a peaking high pass alignment. There will be some ringing if the driver is used without a low-Q HP filter that mitigates the behavior at resonance.
Good idea, or if you desire a second order rolloff and can use an active filter, this is a perfect application for a Linkwitz Transform. It can be used to modify both the Q and also the frequency where the rolloff begins, which you can move up in frequency (in which can it is like a shelving filter), down in frequency (the typical usage of the LT for subwoofers, not recommended in this situation), or just keep it the same.
My proposal would boil down to using a cascade of a second-order Butterworth high-pass at 120 Hz and some sort of biquad that modifies the loudspeaker response into Q = 1/2 sqrt(2), fcut-off = 120 Hz. The whole thing would then be fourth-order Linkwitz-Riley from electrical input to acoustical output.
A Linkwitz Transform circuit would be perfectly suitable for the biquad, but if I can find the time, I'll work out another circuit that does the same with a smaller number of accurate capacitors.
A Linkwitz Transform circuit would be perfectly suitable for the biquad, but if I can find the time, I'll work out another circuit that does the same with a smaller number of accurate capacitors.
I wanted to propose this circuit (but with the op-amp supplies properly connected and decoupled, of course). The first section is a second-order Sallen and Key high-pass, the second section covers the poles of the loudspeaker with zeros and places two new poles.
To get from Q = 1.2 and a natural frequency of 90 Hz to fourth-order 120 Hz Linkwitz-Riley, I chose and calculated these values:
C1 = 22 nF
C2 = 150 nF
C3 = 22 nF
C4 = 22 nF
R1 = 22.1 kohm (or 22 kohm)
R2 = 70111.02469 ohm
R3 = 30965.45225 ohm
R4 = 32893.24758 ohm
R5 = 10 kohm
R6 = 2893.147721 ohm (or 2880.056555 ohm)
R7 = 42628.61345 ohm
R8 = 85257.22691 ohm
Resistances rounded to E12 or E24 values:
R1 = 22 kohm
R2 = 68 kohm
R3 = 33 kohm (or 30 kohm E24)
R4 = 33 kohm
R5 = 10 kohm
R6 = 2.7 kohm (or 3 kohm E24)
R7 = 47 kohm (or 43 kohm E24)
R8 = 82 kohm
With E96 values:
R1 = 22.1 kohm
R2 = 69.8 kohm
R3 = 30.9 kohm
R4 = 33.2 kohm
R5 = 10 kohm
R6 = 2.87 kohm
R7 = 43.2 kohm
R8 = 84.5 kohm
Mind you, I still need to run the circuit through a pole-zero extraction program to check if my calculations are correct.
To get from Q = 1.2 and a natural frequency of 90 Hz to fourth-order 120 Hz Linkwitz-Riley, I chose and calculated these values:
C1 = 22 nF
C2 = 150 nF
C3 = 22 nF
C4 = 22 nF
R1 = 22.1 kohm (or 22 kohm)
R2 = 70111.02469 ohm
R3 = 30965.45225 ohm
R4 = 32893.24758 ohm
R5 = 10 kohm
R6 = 2893.147721 ohm (or 2880.056555 ohm)
R7 = 42628.61345 ohm
R8 = 85257.22691 ohm
Resistances rounded to E12 or E24 values:
R1 = 22 kohm
R2 = 68 kohm
R3 = 33 kohm (or 30 kohm E24)
R4 = 33 kohm
R5 = 10 kohm
R6 = 2.7 kohm (or 3 kohm E24)
R7 = 47 kohm (or 43 kohm E24)
R8 = 82 kohm
With E96 values:
R1 = 22.1 kohm
R2 = 69.8 kohm
R3 = 30.9 kohm
R4 = 33.2 kohm
R5 = 10 kohm
R6 = 2.87 kohm
R7 = 43.2 kohm
R8 = 84.5 kohm
Mind you, I still need to run the circuit through a pole-zero extraction program to check if my calculations are correct.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.