Built my first two-way recently, with an emphasis on aesthetics but with a careful eye on T/S parameters, FR response, et al. I attempted a crossover without measurement and didn't mind the results, but paled to some commercial (NHT) speakers in my living room, so decided to take the jump to measurement and CAD based crossover design. After taking 180 degree horizontal far field measurements of the woofer and tweeter and bringing them into VituixCAD, I have a response I like the look of.
Two things give me pause however, my original crossover modeled in V.C. didn't match my measured real world response, and while I used two ordinary filters: a 2 way LR low pass for the woofer and a 4 way high pass for the tweeter, the only way I could get a curve I liked was with a 22ohm series resistor ahead of the woofer crossover. While I like the response, I don't quite understand what happened with the design with the addition of that resistor, and I can't get a response I'm happy with without it.
Moreover, I worry that while the simulated FR looks good, there is some reason I shouldn't have that resistor there. Drivers are Visaton W100S-8 4" 8ohm Woofer, and Morel MDT 29 1 1/8 soft dome tweeter, also 8ohm.
Question is whether a series resistor on a woofer is odd, or otherwise what I could be doing wrong. Images below are the response curves with and without the series resistor shorted in VituixCAD.
Feel free of course to berate me for whatever other mistakes I may have made here on my first run. I'm happy with the cabinet, and just want to do my best on a passive crossover for non-perfect listening while also trying to learn as much as possible going forward.
Two things give me pause however, my original crossover modeled in V.C. didn't match my measured real world response, and while I used two ordinary filters: a 2 way LR low pass for the woofer and a 4 way high pass for the tweeter, the only way I could get a curve I liked was with a 22ohm series resistor ahead of the woofer crossover. While I like the response, I don't quite understand what happened with the design with the addition of that resistor, and I can't get a response I'm happy with without it.
Moreover, I worry that while the simulated FR looks good, there is some reason I shouldn't have that resistor there. Drivers are Visaton W100S-8 4" 8ohm Woofer, and Morel MDT 29 1 1/8 soft dome tweeter, also 8ohm.
Question is whether a series resistor on a woofer is odd, or otherwise what I could be doing wrong. Images below are the response curves with and without the series resistor shorted in VituixCAD.
Feel free of course to berate me for whatever other mistakes I may have made here on my first run. I'm happy with the cabinet, and just want to do my best on a passive crossover for non-perfect listening while also trying to learn as much as possible going forward.
Attachments
What are L8 and C8 for? That inductor is shorting the amplifier for low frequencies.
A big series resistor only makes sense if the woofer is far more sensitive than the tweeter. In any case, it much increases the Q of the woofer, which is usually not what you want.
How did you end up with second-order Linkwitz-Riley for one branch and fourth-order for the other?
A big series resistor only makes sense if the woofer is far more sensitive than the tweeter. In any case, it much increases the Q of the woofer, which is usually not what you want.
How did you end up with second-order Linkwitz-Riley for one branch and fourth-order for the other?
That's simulating the built-in 6db 100hz highpass in the 2.1 amplifier in my subwoofer. This won't actually end up in the speaker's crossover.
Literally just trying things out. I had originally built a simple 2-order bessel for both drivers but that simulates poorly here.
Revisiting it, (now that I've already ordered components: DOH!) I can get pretty close results with a 3rd or 4th order on both drivers if I adjust the filter impedance. But I confess I don't understand what that's actually doing and what effect it has on performance apart of course for altering the FR curves. I'll try doing some searching.
An LC filter, even back to front like yours, is 12 dB/octave, not 6.
You should be aiming at a 2nd-order Linkwitz-Riley crossover.
You should be aiming at a 2nd-order Linkwitz-Riley crossover.
Still have a question about using the "impedance" tuneable of the library filter. Should I be altering that much? I find it increasing the impedance of the woofer's crossover is very helpful for flattening a big peak near the crossover point.
Okay, so playing some more, two 2nd order Linkwitz-Riley crossovers with a 2500hz crossover point. I added an L-pad after the woofer crossover to bring down the lower frequences to the level of the tweeter's. The tweeter's filter block was tuned with an impedance of 30ohm, which flattened an 8db peak at the crossover point. Visually looks great, but am I missing anything?
Which brings up another question. The woofer sensitivity is lower than the tweeter, I would have expected to have to attenuate the tweeter rather than the woofer. Did I measure wrong, or is there something else about modeling that I could have screwed up? In REW the response graphs have the tweeter around 77db average, with the woofer at 82db average, so if the measurements aren't wrong it does seem to check out.
A speaker that produces 82dB is MORE sensitive than a speaker that produces 77dB from the same input level, which is why you are finding that it needs attenuation.
Part of the problem here is your misuse or misunderstanding of terminology. As further examples, you said 'way' when you meant 'pole', and '6dB' when you meant '6dB/octave'.
That L pad just overloads the amplifier. A series resistor is sufficient. R2 is in parallel with the woofer impedance anyway, so any added precision is illusory.
Part of the problem here is your misuse or misunderstanding of terminology. As further examples, you said 'way' when you meant 'pole', and '6dB' when you meant '6dB/octave'.
That L pad just overloads the amplifier. A series resistor is sufficient. R2 is in parallel with the woofer impedance anyway, so any added precision is illusory.
Last edited:
Thanks on the L-Pad. I only asked because the spec sheet had the sensitivity of the tweeter at 89dB, vs the woofer at 86dB at same voltage and distance, and thus I expected more energy out of the tweeter than the woofer rather than the reverse.. And yes, as a newbie, right in the thread title, I'm likely to get terminology wrong. That said, I have a strong background in electronics, CAD, etc, so I'm trying to get up to speed here for a project but not a career. Bear with me.
Interesting. Can't seem to get anything like the CAD result replacing the L-Pad with a series resistor either before of after the crossover inductor and capacitor. May be my misunderstanding of VituixCAD, but not intuitive based on your advice.
If I add a 150ohm series resistor in front of the woofer crossover, but 150ohms feels like quite a lot. I get closer, but with a ~5db overall reduction in SPL. Which is fine, don't expect this to play loud. But still not as pleasing a flat curve. I probably won't notice it but it's a challenge at this point.
But more to the point, and the initial question is this thread, is that I haven't seen and it feels weird to be heavily damping the woofer with a series resistor esp when the tweeter is speced to have higher sensitivity.
But more to the point, and the initial question is this thread, is that I haven't seen and it feels weird to be heavily damping the woofer with a series resistor esp when the tweeter is speced to have higher sensitivity.
That looks strange, well, all the thread sounds strange
There are basic things that haven't been named. Believing in CAD without knowing the acoustic matters Is bad, being cautious with impedance and load Is good but ignoring amplifier's capability Is silly.
About the quoted, that's silly!
The 1.1 uF & 3.6 mH Cell is something never seen.
Same as thread's title
Yes, all the numbers are wrong
Just consider the piloting of the amplifier.
Take a look to other (millions) crossover schematics: all assume to have a voltage amplifier, right..?!
Definitely missing something in translation of the language here, but appreciate that I am clearly way off base somewhere here. Can someone straighten me out? What I have to work with are some UMIK-1 measurements in REW with the individual drivers amplified with 10 degree measurements from on to 90 degree off axis. Worked through this with the instructions from VituixCAD with REW to import the measurements and went from there. Beyond that I don't have a good background in crossover design obviously, which is as much art as science. One thing I wonder about is whether I should be using the same amplifier as I intend to run the speakers on, but I doubt it, if I've set levels right in REW and the amplifier isn't garbage the sample sweeps should work...
Eventual target mplifier is a Lepai 210PA modest 2.1 plate amplifier in my subwoofer. 30W per channel for the mains. This is for very low play in the bedroom waking up, not looking for power, more for clarity.
My apologies to the entire thread for not having worked at this for long and lacking a ton of theory. In my profession I know all the ins and outs, and my wrong values here are due to the fact that the software gives me all the knobs to adjust a FR curve that looks great without knowing what to the experienced designer knows is obviously wrong. I'm not that experienced designer..
Eventual target mplifier is a Lepai 210PA modest 2.1 plate amplifier in my subwoofer. 30W per channel for the mains. This is for very low play in the bedroom waking up, not looking for power, more for clarity.
My apologies to the entire thread for not having worked at this for long and lacking a ton of theory. In my profession I know all the ins and outs, and my wrong values here are due to the fact that the software gives me all the knobs to adjust a FR curve that looks great without knowing what to the experienced designer knows is obviously wrong. I'm not that experienced designer..
Last edited:
Science tells you that there are coherent sources, as they play all the same amplified signal, that these sources/speakers are mostly aligned on vertical axys and contribute in creating a sound field, the two channels do not contain the same exact signal and (science) tells that human hearing starts from the ears which are positioned laterally (each..!)so the caption happens hortogonally, somehow...