I'm in the final stage of building crossovers for my 2-way full-range plus woofer project that is the "voicing" procedure--no more measurements, tweaking by ears only.
The schematic is shown in the picture above. Crossover point is around 400 Hz. The fixed part is the high-pass section of the 6-Ohm driver because the manufacturer recommends this and the values of components are listed below:
R2 = 2 Ohms, C2 = 33 uF, L2 = 2.6 mH
Component values of the variable part are as follows:
L1 = 3.5 mH, C1 = 100 uF
In voicing (or listening test) procedure, the result was found as described here: female singer's voice is almost clear in its range except the lower notes, it has "Umm" sound coming from the woofer (4-Ohm driver) which I realizes it as the excessive mid-bass issue. So did the male voice.
I'm here to ask for opinions from gurus/experienced DIYers about the "tweaking" methods. However, to not making confusions, I will use number instead of writing long description.
A) L1 = 3 mH, C1 = 68 uF
B) L1 = 4 mH, C1 = 130 uF
C) L1 = 3 mH, C1 = 130 uF
D) L1 = 4 mH, C1 = 68 uF
Which way would you go for--to eliminate the unwanted mid-bass (Umm) problem?
Please let me hear your thoughts.
The schematic is shown in the picture above. Crossover point is around 400 Hz. The fixed part is the high-pass section of the 6-Ohm driver because the manufacturer recommends this and the values of components are listed below:
R2 = 2 Ohms, C2 = 33 uF, L2 = 2.6 mH
Component values of the variable part are as follows:
L1 = 3.5 mH, C1 = 100 uF
In voicing (or listening test) procedure, the result was found as described here: female singer's voice is almost clear in its range except the lower notes, it has "Umm" sound coming from the woofer (4-Ohm driver) which I realizes it as the excessive mid-bass issue. So did the male voice.
I'm here to ask for opinions from gurus/experienced DIYers about the "tweaking" methods. However, to not making confusions, I will use number instead of writing long description.
A) L1 = 3 mH, C1 = 68 uF
B) L1 = 4 mH, C1 = 130 uF
C) L1 = 3 mH, C1 = 130 uF
D) L1 = 4 mH, C1 = 68 uF
Which way would you go for--to eliminate the unwanted mid-bass (Umm) problem?
Please let me hear your thoughts.
The conundrum is that the ear is a terrible measuring device, yet it has to be the final arbitrator. I see no reason to suddenly throw all measurements to the wind when voicing a speaker. Tie what you hear to what you measure and then work out how to correct that region. You may find it's not so simple as tweaking the existing values, but you have to add additional networks to shape what you've got with a bit more precision.
for quick measurements use pink noise signal from youtube and a smartfone app called spectroid works good for the frequency range 100hz to 15khz
Actually, these speakers had been measured for many many times in development. But, as mentioned earlier, this is the final stage of “listening test”.
Hi,
my 2c general tips for the process:
Compare to something, good pair of headphones or good set of speakers! At least in broad sense that it's got as much bass as you like and as much treble as you like, with about same output volume because sound changes if listening at different volume. Ear gets used to the sound pretty quickly, so a referemce is must, or have a process to tweak only at mornings for 30minutes or something like that, to standardize your hearing.
Then, when you think you got it, leave it in and spend time with it, like days, maybe weeks or months and get used to it. Remember to compare to other systems, ask yourself, does it sound right? if not, try to figure out why by what ever means. Don't be afraid to get back to it and change thigns. Take notes of your xo and feelings etc. just to remember what you did a moth ago and why, if you now want to revert back to some old version of the xo.
Basically, it's time to use your listening skills, and you can improve the listening skill while you are at it! Have fun!🙂
my 2c general tips for the process:
Compare to something, good pair of headphones or good set of speakers! At least in broad sense that it's got as much bass as you like and as much treble as you like, with about same output volume because sound changes if listening at different volume. Ear gets used to the sound pretty quickly, so a referemce is must, or have a process to tweak only at mornings for 30minutes or something like that, to standardize your hearing.
Then, when you think you got it, leave it in and spend time with it, like days, maybe weeks or months and get used to it. Remember to compare to other systems, ask yourself, does it sound right? if not, try to figure out why by what ever means. Don't be afraid to get back to it and change thigns. Take notes of your xo and feelings etc. just to remember what you did a moth ago and why, if you now want to revert back to some old version of the xo.
Basically, it's time to use your listening skills, and you can improve the listening skill while you are at it! Have fun!🙂
The Umm is rarely the crossover problem, more likely it comes from the underdampened LF driver (excessive Q, poorly calculated bass reflex alignment, etc).
First thing I would try is increasing the lowpass coil. If not better, then increase the cap as well. If still not better, try a resistor in series with the cap, 1-3 ohms ballpark, 10W is fine.
If none of this above makes any improvement, then try a 25-50W 40 ohm across the woofer directly. What this does is lower the impedance magnitude at resonance but not much anywhere else. If this improves the sound, then an LCR across the woofer will remove the peaky bass. Sometimes a smaller gauge lowpass coil will do the same thing for less money.
If it still does not improve, then you need to likely adjust the 2 ohm resistor on the wideband to 1 ohm or nothing at all.
If none of this above makes any improvement, then try a 25-50W 40 ohm across the woofer directly. What this does is lower the impedance magnitude at resonance but not much anywhere else. If this improves the sound, then an LCR across the woofer will remove the peaky bass. Sometimes a smaller gauge lowpass coil will do the same thing for less money.
If it still does not improve, then you need to likely adjust the 2 ohm resistor on the wideband to 1 ohm or nothing at all.
Yeah measurements and additional information would help... If L1 resistance is high that reduces electrical damping and could be the culprit. Same for adding series resistor potentially makes the sound worse by reducing electrical damping further. Too little damping makes peak in the response. Also the box could have an issue. The system might be close to wall enhancing lows and so on, hard to say. Proper measurements and look at them in vituixCAD could likely help solve the issue if it's in the system balance.
If it's issue with damping you could make the box leak a bit try and find if that helps, loosen the driver mounting screws and put shim between baffle and woofer to introduce some leakage which should increase mechanical damping, like would adding good amount of wadding inside the box. Could be bad sound leaking out from a port too, if there is one. Is it a big or small box? This kind of mechanical problems would be best fixed physically and not in the xo, while xo is perfectly fine to do it if it helps, but first you gotta identify the issue to be able to fix it.
If it's issue with damping you could make the box leak a bit try and find if that helps, loosen the driver mounting screws and put shim between baffle and woofer to introduce some leakage which should increase mechanical damping, like would adding good amount of wadding inside the box. Could be bad sound leaking out from a port too, if there is one. Is it a big or small box? This kind of mechanical problems would be best fixed physically and not in the xo, while xo is perfectly fine to do it if it helps, but first you gotta identify the issue to be able to fix it.
Last edited:
I would add that you should listen to the loudspeakers in the location they'll end being: there and only there you'll have the 'correct' correction for baffle step loss.
As already stated by Tmuikku you'll need a known reference to compare against, headphones are obvious.
The issue is there could be room mode at play and whatever you'll do in your 'voicing' tweaking of crossover it won't help with this as it is room dependent.
That said, correcting for the right amount of BSC in place should make a real step ahead, at least this is my experience.
If you have a computer as source and known headphones it is easy to try:
First define your baffle step center frequency 'coarsely' using the rule of thumb: 115/box width in meter or 380/ box width in feet.
Balance level at a given frequency between headphones and loudspeaker ( set aside the frequency difference for the time being, for this i often use the calculated baffle step filter center frequency to define the 'given frequency' by playing a sinus frequency used for balancing levels between headphone and loudspeaker: you should perceive same loudness from one of the heaphone ear ( use only one ear for headphone) and the loudspeaker set).
From there i use an high shelf filter with a Q parameter of 4 octave ( two octave below and two octave above center frequency)
From there you cut high end until it sounds similar to what you hear in headphones using similar perceived balance ( headphones take away the room coloration from the equation).
It works nicely and will help you voice to your own environnement. Just remember it is ok only for the location defined in room as bsc depend from distances to room's wall.
As already stated by Tmuikku you'll need a known reference to compare against, headphones are obvious.
The issue is there could be room mode at play and whatever you'll do in your 'voicing' tweaking of crossover it won't help with this as it is room dependent.
That said, correcting for the right amount of BSC in place should make a real step ahead, at least this is my experience.
If you have a computer as source and known headphones it is easy to try:
First define your baffle step center frequency 'coarsely' using the rule of thumb: 115/box width in meter or 380/ box width in feet.
Balance level at a given frequency between headphones and loudspeaker ( set aside the frequency difference for the time being, for this i often use the calculated baffle step filter center frequency to define the 'given frequency' by playing a sinus frequency used for balancing levels between headphone and loudspeaker: you should perceive same loudness from one of the heaphone ear ( use only one ear for headphone) and the loudspeaker set).
From there i use an high shelf filter with a Q parameter of 4 octave ( two octave below and two octave above center frequency)
From there you cut high end until it sounds similar to what you hear in headphones using similar perceived balance ( headphones take away the room coloration from the equation).
It works nicely and will help you voice to your own environnement. Just remember it is ok only for the location defined in room as bsc depend from distances to room's wall.