I think that the 1.25 factor is applied because the impedance we are trying to equalise is on the right-hand side of the impedance peak, and the impedance minimum there is typically a bit greater than the DC resistance of the driver.
It's interesting that 1.25 is around eight ohms divided by the typical coil resistance.
No. What you'll usually find is that the existing impedance minimum isn't as low as the resistance, and at the top end it doesn't reach significantly high enough to need a resistor equal to Re. Higher would generally be better if it can give a good result, however when you design these on a simulator it becomes clear that the ideal value can be different for various reasons. For example how low you need or want to flatten, such as with a 3-way vs a 2-way.
No. What you'll usually find is that the existing impedance minimum isn't as low as the resistance, and at the top end it doesn't reach significantly high enough to need a resistor equal to Re. Higher would generally be better if it can give a good result, however when you design these on a simulator it becomes clear that the ideal value can be different for various reasons. For example how low you need or want to flatten, such as with a 3-way vs a 2-way.
AllenB,
Thanks for the great tutorial. I am building my 2 way crossover, and i would appreciate some help.my compression horn is nominally rated at 8ohms, and i paralleled a 10 ohm resistor, that makes tht tweeter resistance 4.44 ohms. i chose to crossover at 2khz, and my cal value is 10 uf, Inductor value is 0.69mH. The compression driver is 108db sensitive and the woofer is 97db sensitive. i plan to reduce the tweeter to match with woofer by attenuating to 12 or 14 dB. As per your tutorial, 10R series resistor will attenuate 6db. so, to attenuate even further should i increase series rsesitor to 22 ohm, or should i reduce the parallel resistor from 10r to 4.7r? do you recommend a L pad with dpdt rotary switch to maintain the 8 ohm impedance? in such a case, can we skip the parallel resistor altogether?
Thanks for the great tutorial. I am building my 2 way crossover, and i would appreciate some help.my compression horn is nominally rated at 8ohms, and i paralleled a 10 ohm resistor, that makes tht tweeter resistance 4.44 ohms. i chose to crossover at 2khz, and my cal value is 10 uf, Inductor value is 0.69mH. The compression driver is 108db sensitive and the woofer is 97db sensitive. i plan to reduce the tweeter to match with woofer by attenuating to 12 or 14 dB. As per your tutorial, 10R series resistor will attenuate 6db. so, to attenuate even further should i increase series rsesitor to 22 ohm, or should i reduce the parallel resistor from 10r to 4.7r? do you recommend a L pad with dpdt rotary switch to maintain the 8 ohm impedance? in such a case, can we skip the parallel resistor altogether?
Compression drivers tend to have difficult impedance peaks near resonance. For this reason it can help to have the parallel resistor as low as practical, as long as together with the series resistor the total doesn't get too low. By doing this you minimise response peaks when not measuring the impedance. With a compression driver, the horn changes the impedance so you couldn't use factory impedance data even if you wanted to.
As for the variable L-pad, sure. Compression tweeters are usually so sensitive that they are low powered and could be used that way.
Another thing, they are usually able to be crossed lower than recommended when used domestically, provided your horn or waveguide can support it.
As for the variable L-pad, sure. Compression tweeters are usually so sensitive that they are low powered and could be used that way.
Another thing, they are usually able to be crossed lower than recommended when used domestically, provided your horn or waveguide can support it.
4.4 ohms in series with this should bring the tweeter down 6dB with a total impedance of 8.8 ohms.
I agree with all you posted, except for this last line >
This regards the final 'sound' that would be achieved.
That is a very nice information. i indeed use domestically, in a 16x10 ft room.The idea of this 2 way design is to make a high sensitive speaker system to mate with my low powered tube amps. My compression driver is sb audience bianco cd25p.The fs is 1000hz, and the datasheet recommends crossover not less than 2500hz, i did however chose 2000hz, my horn wave guide also from sb acoustics goes down to 600hz.
The woofers sensitivity is 97db and CD is 108db, so i should achieve 14db attaenuation. how can i achieve it?
6.8 series and 2.2 parallel will work in theory, and keep close to 8 ohms total. You may want to adjust them after listening.
You should be able to cross the typical, conventional type 1" compression tweeter near 1kHz and get clean club like levels in a home listening room, YMMV.
This is dependent on having a waveguide that supports it, this will be larger than a horn with that cutoff frequency. Low diffraction is needed for it to remain clean at such high levels because diffraction audibility is level dependent.
AllenB, sorry if it has been answered or explained before....Is there a Guideline to use bsc or skip it? by using the no measurement approach. say for example, i use a 12 inch woofer ,not flush mounted on a 14 inch front baffle?
If you want to skip worrying about the baffle step with a box speaker, I think you'll find it easier when you have a large baffle. 14 inches might be OK, 60cm could be better.
Another way to approach this is to use an equaliser, so you can adjust the baffle step and other tonal issues. It may be desirable to go back and incorporate any EQ into the crossover.. but that's up to you.
What is at stake here is the power balance in the lower midrange. This is a region that gives weight to the sound and adds tactile feel. It's not only going to be lighter but it will have more peaks and dips when it is distributed more around the room.
Some of these won't respond well to being equalised. The smaller the box, the more the upper midrange gets involved and we are more able to hear the spread in addition to the balance.
Another way to approach this is to use an equaliser, so you can adjust the baffle step and other tonal issues. It may be desirable to go back and incorporate any EQ into the crossover.. but that's up to you.
What is at stake here is the power balance in the lower midrange. This is a region that gives weight to the sound and adds tactile feel. It's not only going to be lighter but it will have more peaks and dips when it is distributed more around the room.
Some of these won't respond well to being equalised. The smaller the box, the more the upper midrange gets involved and we are more able to hear the spread in addition to the balance.
AllenB, Here is my update.. The 2.2 ohm parallel resistor and the series resistor 6.8ohms definitely made the tweeter shout to come down.. However I did notice some tonal imbalance.. So I went back to the calculations.. The tweeter impedance is now 1.75ohms, so the capacitor needed change from 10mfd ( previous Tweeter impedance was 4.4R) to 2.2mfd.. Now the tweeter blends well. I also noted that the sound is better without the tweeter inductor 0.26mH. So essentially I am running a first order crossover... So, If more attenuation is needed like 14dB, as in my case, I think setting the resistor Network first and then to calculate the capacitor would be better... Am I right?
For me, this tutorial and crossover exercise is an eye opener, especially the impedance flattening RC network on woofer and the parallel resistor on tweeter.. Those two are the most important things which got the tonal balance right.. Many of the online calculators didn't even mention that a RC network is to be present.. Thanks Again, I will eventually one day start using Xsim.,, But this pen and paper and calculator approach is much more satisfying.
For me, this tutorial and crossover exercise is an eye opener, especially the impedance flattening RC network on woofer and the parallel resistor on tweeter.. Those two are the most important things which got the tonal balance right.. Many of the online calculators didn't even mention that a RC network is to be present.. Thanks Again, I will eventually one day start using Xsim.,, But this pen and paper and calculator approach is much more satisfying.
There are two places you can put the series resistor and it could make a difference. With a conventional measured approach it is often put just before the parallel resistor. This makes your filter load look like 8 ohms.
The no measurement approach has placed the series resistor before the filter because this reduces the filter interaction a little. This is so that changing the impedance the filter is working into doesn't have as much of an effect on the filter peak just before rolloff (where it rises before falling).
Of course, this is no problem if you know which value you will use. In fact, you can achieve the same result both ways if you measure.
Placing the series resistor after the filter might help to reduce the impedance variations of the driver a little further. I can't tell whether they need the help, it would have to be assessed. The main difference to watch for if you alternate between the two places to put the series resistor is that the filter values you need will be very different.
The no measurement approach has placed the series resistor before the filter because this reduces the filter interaction a little. This is so that changing the impedance the filter is working into doesn't have as much of an effect on the filter peak just before rolloff (where it rises before falling).
Of course, this is no problem if you know which value you will use. In fact, you can achieve the same result both ways if you measure.
Placing the series resistor after the filter might help to reduce the impedance variations of the driver a little further. I can't tell whether they need the help, it would have to be assessed. The main difference to watch for if you alternate between the two places to put the series resistor is that the filter values you need will be very different.
Question: for the capacitor, resistor & inductor halving/doubling on 2 woofer connection. does this apply on zobel/woofer impendance flattening as well? because you put them before the speaker connected in paralel (for my case) (1 resistor & 1 inductor per woofer). so I was speculating the speaker impendance r still not yet halved.
Thanks for this question as I'm interested in clarifying the original post.
There are two scenarios. One is the 2.5 way with woofers in parallel only up to the baffle frequency and then one of the woofers goes further, up to the cross.
In the second option, both woofers remain in parallel all the way to the cross. In this second option, the crossover circuit is shared. The woofers can be connected directly in parallel, then the RC circuit and the inductor can all be calculated according to the reduced impedance.
With the 2.5 way option, you can't connect the woofers directly in parallel since they need different circuits. The upper woofer is easy, do that the same as for a single woofer, using the single woofer impedance.
The lower woofer should also be seen as having the single woofer impedance for calculation of any of the components you choose to use. The series inductor is larger since it will be tuned to the baffle frequency, and the RC impedance compensation is the same as before. You could leave the RC out if looking to use fewer components. Since this cross starts lower it is somewhat less critical.
So with the 2.5 way the woofers aren't wired in parallel directly, but the circuits are actually connected in parallel back at the amplifier output. The amplifier still sees them in parallel through the bass frequencies where the circuits are not having a significant effect.
There are two scenarios. One is the 2.5 way with woofers in parallel only up to the baffle frequency and then one of the woofers goes further, up to the cross.
In the second option, both woofers remain in parallel all the way to the cross. In this second option, the crossover circuit is shared. The woofers can be connected directly in parallel, then the RC circuit and the inductor can all be calculated according to the reduced impedance.
With the 2.5 way option, you can't connect the woofers directly in parallel since they need different circuits. The upper woofer is easy, do that the same as for a single woofer, using the single woofer impedance.
The lower woofer should also be seen as having the single woofer impedance for calculation of any of the components you choose to use. The series inductor is larger since it will be tuned to the baffle frequency, and the RC impedance compensation is the same as before. You could leave the RC out if looking to use fewer components. Since this cross starts lower it is somewhat less critical.
So with the 2.5 way the woofers aren't wired in parallel directly, but the circuits are actually connected in parallel back at the amplifier output. The amplifier still sees them in parallel through the bass frequencies where the circuits are not having a significant effect.
Noted. so I can straight connect the woofer in paralel & start the zobel from there. treating them as single speaker with halved impendance.
thanks.
thanks.
NOTE:
Although an R/C zobel network that achieves a totally flat woofer impedance can help select 'textbook' component values,
there are times where deviation is actually required for certain drivers to obtain optimal results.
This is in the realm of where things get 'fine-tuned'.
Although an R/C zobel network that achieves a totally flat woofer impedance can help select 'textbook' component values,
there are times where deviation is actually required for certain drivers to obtain optimal results.
This is in the realm of where things get 'fine-tuned'.
Yes that's right, when considered with the varying response. In addition, there is varying power (on and off-axis total output).
This brings us to what to do when using no measurements. Crossing a two way with a lone inductor used high in the woofer's passband gets close to where the rising impedance offsets the inductor, where it is trying to roll off the upper woofer response. Not only does this mean the RC circuit can make a difference in this position, it also makes it an effective part of the circuit for tweaking that part of the response.
Response and power are usually more consistent with frequency before they begin to fall toward the top end. This can be a matter of choosing a good woofer crossover frequency, and not pushing it too high. Falling power can reduce the apparent output just below the cross, creating a dip in impact.
Falling response can change the shape of the filtering, leading to uncertain phase response which can cause a bumpy response around the crossover.
As for balance issues due to a slowly rising response across the woofer passband, this might be approached similarly to handling the baffle step.
This brings us to what to do when using no measurements. Crossing a two way with a lone inductor used high in the woofer's passband gets close to where the rising impedance offsets the inductor, where it is trying to roll off the upper woofer response. Not only does this mean the RC circuit can make a difference in this position, it also makes it an effective part of the circuit for tweaking that part of the response.
Response and power are usually more consistent with frequency before they begin to fall toward the top end. This can be a matter of choosing a good woofer crossover frequency, and not pushing it too high. Falling power can reduce the apparent output just below the cross, creating a dip in impact.
Falling response can change the shape of the filtering, leading to uncertain phase response which can cause a bumpy response around the crossover.
As for balance issues due to a slowly rising response across the woofer passband, this might be approached similarly to handling the baffle step.