im having a hard time figuring out my crossover design for a 2 way system im building, which includes a 6.5" tangband driver along with a fountek ribbon crossed at 2.8k with a second order butterworth filter. ive read that second order filters have 180 degree phase shift, and i am wondering if this is corrected by switching the polarity(positive and negative connections). im asking this because i read that phase and polarity are different but i do not now why. also one of the calculators that i looked at showed that the wiring for the tweeter was switched but not the woofer, why is this?
http://www.carstereo.com/help/Articles.cfm?id=9
http://www.carstereo.com/help/Articles.cfm?id=9
Hi bondo,
No technical talk: The 180 degree shift means the two drivers are effectively firing at opposite times so to correct that you reverse the leads on the tweeter only. If you were to reverse them both you have canceled the reason for the switch. If you were to take the speakers somewhere and get them tested, you will see a dropout at the XO point if they are connected both positive to positive. Little else changes. There's a lot more to it but I thought you would be able to swallow that pretty easy to start.
No technical talk: The 180 degree shift means the two drivers are effectively firing at opposite times so to correct that you reverse the leads on the tweeter only. If you were to reverse them both you have canceled the reason for the switch. If you were to take the speakers somewhere and get them tested, you will see a dropout at the XO point if they are connected both positive to positive. Little else changes. There's a lot more to it but I thought you would be able to swallow that pretty easy to start.
With 12db its normal to reverse tweeter polarity
But only if drivers are correctly time allignet on front baffle
This is usually not the case
From what I have seen, it seems like 18db is often used on tweeter, to compensate fore time allignment, and still 12db on woofer
Not sure about driver polarity when doing that
Another way is to tilt the box a bit
Mind you, its way more complicated than that, as it is the actual acoustical slopes that matters most
Calculaters wont tell you much about that
To me, THE most critical part of doing crossovers IS the phase issues
Its not very difficult to make a speaker sound right, but phase is a whole different thing
But a 2way should be doable, with proper drivers
And you should be able to hear when phase is ok
Lots of designs around to study
But only if drivers are correctly time allignet on front baffle
This is usually not the case
From what I have seen, it seems like 18db is often used on tweeter, to compensate fore time allignment, and still 12db on woofer
Not sure about driver polarity when doing that
Another way is to tilt the box a bit
Mind you, its way more complicated than that, as it is the actual acoustical slopes that matters most
Calculaters wont tell you much about that
To me, THE most critical part of doing crossovers IS the phase issues
Its not very difficult to make a speaker sound right, but phase is a whole different thing
But a 2way should be doable, with proper drivers
And you should be able to hear when phase is ok
Lots of designs around to study
bondo,
The reason the woofer's polarity stays the same and the tweeter's polarity switches is because the high-pass filter has a 90 degree phase-lead at the crossover point, and the low-pass filter has a 90 degree phase-lag at the crossover point. So the phase-lead with the phase-lag is what causes the 180 degree phase shift.
By wiring the tweeter in reverse your woofer and tweeter are in phase at the crossover point relative to one another but still 90 degrees out of phase relative to the input signal.
Roger
The reason the woofer's polarity stays the same and the tweeter's polarity switches is because the high-pass filter has a 90 degree phase-lead at the crossover point, and the low-pass filter has a 90 degree phase-lag at the crossover point. So the phase-lead with the phase-lag is what causes the 180 degree phase shift.
By wiring the tweeter in reverse your woofer and tweeter are in phase at the crossover point relative to one another but still 90 degrees out of phase relative to the input signal.
Roger
That's the theory.
The reality is not so nice.
You have to test the actual acoustic response of the drivers to see what is really happening. If you are trying to crossover very near to the LF rolloff point of the tweeter, you will find that the actual slopes and phase shifts are quite different.
It is essential to work with the actual acoustic response + xover, and not just the theoretical.
Fwiw, I generally avoid 2nd order and 3rd order xovers if at all possible. Personally I feel that first order (often not practical) or a version of a 4th order seem to end up sounding the best. There are exceptions, but that is what seems to work out generally speaking.
Also, there is some very real difference between the practical implementation of a 2nd order parallel and a 2nd order series network.
_-_-bear
The reality is not so nice.
You have to test the actual acoustic response of the drivers to see what is really happening. If you are trying to crossover very near to the LF rolloff point of the tweeter, you will find that the actual slopes and phase shifts are quite different.
It is essential to work with the actual acoustic response + xover, and not just the theoretical.
Fwiw, I generally avoid 2nd order and 3rd order xovers if at all possible. Personally I feel that first order (often not practical) or a version of a 4th order seem to end up sounding the best. There are exceptions, but that is what seems to work out generally speaking.
Also, there is some very real difference between the practical implementation of a 2nd order parallel and a 2nd order series network.
_-_-bear
Problem is that "6db" doesnt really exist, or at least hard to achieve with ordinary drivers
I have just tried to convert my 3way into a "6db" design, and that didnt work at all...I struggled with severe phase issues
Though, I have done small very simple 2ways with just a coil and cap(6db), and drivers just a 5" and cheap cone tweeter and they worked out very nicely without much effort,, maybe not the last word in "highend", but very easy to listen to, and made the young lad very happy fore a couple of years
Maybe I could have used them now, with a small ribbon on top and a good big woofer on the bottom
I have just tried to convert my 3way into a "6db" design, and that didnt work at all...I struggled with severe phase issues
Though, I have done small very simple 2ways with just a coil and cap(6db), and drivers just a 5" and cheap cone tweeter and they worked out very nicely without much effort,, maybe not the last word in "highend", but very easy to listen to, and made the young lad very happy fore a couple of years
Maybe I could have used them now, with a small ribbon on top and a good big woofer on the bottom
bear said:
Ok, let's not scare him away. I think he had a question that needed a rather elementary answer. Come to think of it he hasn't been back. Maybe he's scared of bears.
tinitus said:
I think that's the league we are in with this poster. He just wants to get what he can right. Bondo you still there?
Sorry... Cal you are right...
Bondo, the fact is that anything you make is better than not making anything.
Even if it comes out wrong, it is a learning experience that will only make it better next time around... The good part is that speakers always make sound and even if the sound doesn't seem quite like what you want, you can change some things around and hear what happens... no big deal.
You can also do something, be unsure and ask questions here. You'll probably get as much information and input as you can handle - just let it wash over you and take out of it whatever you can grab and try to use it...
Have no fear and go for it!
_-_-bear
Bondo, the fact is that anything you make is better than not making anything.
Even if it comes out wrong, it is a learning experience that will only make it better next time around... The good part is that speakers always make sound and even if the sound doesn't seem quite like what you want, you can change some things around and hear what happens... no big deal.
You can also do something, be unsure and ask questions here. You'll probably get as much information and input as you can handle - just let it wash over you and take out of it whatever you can grab and try to use it...
Have no fear and go for it!
_-_-bear
I have used 2nd order passive and active crossovers on many of my recent designs and have never reversed the polarity of the high frequency driver. I always connect the drivers in-phase. Designing a crossover is more then just looking at the textbook filter magnitude and phase as if it were driving a pure resistor. Bear's first post was exactly right, you have to look at the driver's/system's actual acoustic response and impedance to design the crossover.
If you can get a hold of this paper, it will explain a lot:
http://www.aes.org/e-lib/browse.cfm?elib=3978
http://www.aes.org/e-lib/browse.cfm?elib=3978
In a second order crossover you have to switch the polarity of either the tweeter or the woofer i.e. swap the positive and negative terminals. A 180 degrees phase shift means the tweeter and woofer are working in opposition to each other instead of in harmony. By swapping the terminals you get them working in harmony.
Last edited:
When we talk about target slopes - we usually refer to them as follows:
crossover + driver response = acoustic target slope.
A textbook "butterworth" crossover won't give you a textbook acoustic target slope. This is because the driver itself has it's own frequency response peaks, dips, rolloff and impedance changes.
To combine all the above - you might find you need a 3rd order filter (3 components) on the tweeter (plus attenuation padding) with 2nd order on the woofer (2 components) to get a 4th order acoustic XO.
The on-line calculators assume:
A perfect flat frequency response from 1Hz to 20Khz
a perfect flat impedance.
time aligned voice coils (= slanted or stepped baffle putting the tweeter behind / further back than the woofer)
Now look at the FR curve and impedance curve for your drivers....
The impedance curve changes with frequency.
the closer you are to the natural driver roll off (top or bottom end) and /or resonant peak (where the impedance "spikes") - the more worthless the on-line calculators become.
May I suggest you review the sticky in this forum "designing loudspeaker without measurements" or google the same.
Using a XO simulation tool is extremely useful - infact essential.... or you are shooting in the dark.
crossover + driver response = acoustic target slope.
A textbook "butterworth" crossover won't give you a textbook acoustic target slope. This is because the driver itself has it's own frequency response peaks, dips, rolloff and impedance changes.
To combine all the above - you might find you need a 3rd order filter (3 components) on the tweeter (plus attenuation padding) with 2nd order on the woofer (2 components) to get a 4th order acoustic XO.
The on-line calculators assume:
A perfect flat frequency response from 1Hz to 20Khz
a perfect flat impedance.
time aligned voice coils (= slanted or stepped baffle putting the tweeter behind / further back than the woofer)
Now look at the FR curve and impedance curve for your drivers....
The impedance curve changes with frequency.
the closer you are to the natural driver roll off (top or bottom end) and /or resonant peak (where the impedance "spikes") - the more worthless the on-line calculators become.
May I suggest you review the sticky in this forum "designing loudspeaker without measurements" or google the same.
Using a XO simulation tool is extremely useful - infact essential.... or you are shooting in the dark.
1st order only works in a 2 way or a 3 way where the mid has a very broad spectrum (like when using a FR driver as mid). You need to count 2 decades for the slope on both sides of the cr point to make it work.
I use a 1st order at 180Hz in my Mark Audio Alpair 10M gen3/Scanspeak 26W/8534G00 WAW setup, and there it works as the 10M goes rather flat untill +/- 40Hz and the scanspeak is useable to +/- 800Hz, so at both sides i have + 2 decades space.
But on a classic 3 way this does not work as the bandwith of each driver is to small to have this kind of space, a 2nd or higher order is needed.
But when possible to do it right, a 1st order sounds the best to me.
Each order is 45 degrees at the crossover. Second order LP + HP is 4*45=180.
The difference is that LR only comes in even orders. Butterworth is more interesting in odd orders. It is the odd order that makes the polarity work either way.,
LR2 reverse polarity. LR4 same polarity.
Filters aren't always this simple in practice.
The difference is that LR only comes in even orders. Butterworth is more interesting in odd orders. It is the odd order that makes the polarity work either way.,
LR2 reverse polarity. LR4 same polarity.
Filters aren't always this simple in practice.
If the two drivers had the same gain/sensitivity and were placed at the same acoustic distance from the listener then a pair of 2nd order LP+HP filters would cancel at the crossover frequency being +90 deg and -90deg. So 2nd Order crossovers often call for one of the drivers to be wired inverted.
However, the tweeter is usually mounted on the same baffle as the woofer so it is a couple inches closer to the listener giving the tweeter sound a significant phase lead. And that phase lead varies with the listening position. In any case, most drivers have peaks and troughs that are similar amplitude variations.
Of more concern is the different impedance at certain frequencies that different filters have at the driver. Woofers sound much smoother with a single inductor in series than with a shunt capacitor because the undesirable cone resonances are more pronounced when driven by a low impedance. Likewise, the Q of the tweeter resonance is higher when shunted by a choke.
So unless protecting a driver like the tweeter from harmful low frequency energy is a priority, a simple first order filter usually produces the best sound. Third order on the tweeter only, would be my next choice. 2nd order on the tweeter is not usually a problem because the tweeter also requires an attenuator, which dampens the cone by presenting a ~matching impedance.
In the final analysis, just try it and use whatever sound the best.
However, the tweeter is usually mounted on the same baffle as the woofer so it is a couple inches closer to the listener giving the tweeter sound a significant phase lead. And that phase lead varies with the listening position. In any case, most drivers have peaks and troughs that are similar amplitude variations.
Of more concern is the different impedance at certain frequencies that different filters have at the driver. Woofers sound much smoother with a single inductor in series than with a shunt capacitor because the undesirable cone resonances are more pronounced when driven by a low impedance. Likewise, the Q of the tweeter resonance is higher when shunted by a choke.
So unless protecting a driver like the tweeter from harmful low frequency energy is a priority, a simple first order filter usually produces the best sound. Third order on the tweeter only, would be my next choice. 2nd order on the tweeter is not usually a problem because the tweeter also requires an attenuator, which dampens the cone by presenting a ~matching impedance.
In the final analysis, just try it and use whatever sound the best.
That’s right. Gosh it’s been awhile since I’ve had to think that through. I haven’t used BW in so long I forgot what they do. I just remembered never to use them because I never had a reason to.
Something is still telling me the order is based on the number of poles (or reactive elements) I am not entirely sure it’s just 45*4 to get there
I thought each electrical order added 90deg and not 45
As 45deg is as 1rad at cutoff , 45deg on each side of the reactive element not each order
So a 1st order BW Q.707 would be at quadrature at cutoff. 45deg on each respective side of the time constant. The sum is 90deg (or something like that it’s been awhile not quadrature because that is 1/4th of 360 it’s 1/8th so it’s someth else I can’t remember)
Please correct me if I’m wrong
I thought each electrical order added 90deg and not 45
As 45deg is as 1rad at cutoff , 45deg on each side of the reactive element not each order
So a 1st order BW Q.707 would be at quadrature at cutoff. 45deg on each respective side of the time constant. The sum is 90deg (or something like that it’s been awhile not quadrature because that is 1/4th of 360 it’s 1/8th so it’s someth else I can’t remember)
Please correct me if I’m wrong
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.