Lowest f3 is usually obtained where Qtc=0.7
Your crossover may give unexpected results where the impedance peaks, so having that peak low in level and low in frequency is what I meant. It really depends on how well you can craft a filter for it.
Hi,
for what its worth depending on what information you feed
it Xover Pro will likely give you completely wrong results.
You would need to feed it the exact measured responses
of the drivers installed in the cabinet measured correctly.
and the impedance profiles of the drivers in situ.
I don't like xover Pro. it does loads of electrical textbook
filters that are totally irrelevant in practise, all that does
matter is the final acoustic response of the drivers.
It allows you to completely get a design totally wrong.
And nearly all beginners get 3 ways very wrong.
For why read some of the links below, particularly the
stuff on simulating the design of speakers. Xover Pro
just lets you ignore stuff that has to be accounted for
in the overall design.
rgds, sreten.
undefinition
(see if nothing else, the excellent FAQs)
The Speaker Building Bible - Techtalk Speaker Building, Audio, Video, and Electronics Customer Discussion Forum From Parts-Express.com
Zaph|Audio
FRD Consortium tools guide
Designing Crossovers with Software Only
RJB Audio Projects
Jay's DIY Loudspeaker Projects
Speaker Design Works
HTGuide Forum - A Guide to HTguide.com Completed Speaker Designs.
DIY Loudspeaker Projects Troels Gravesen
Humble Homemade Hifi
Quarter Wavelength Loudspeaker Design
The Frugal-Horns Site -- High Performance, Low Cost DIY Horn Designs
Linkwitz Lab - Loudspeaker Design
Music and Design
Great free SPICE Emulator : SPICE-Based Analog Simulation Program - TINA-TI - TI Tool Folder
for what its worth depending on what information you feed
it Xover Pro will likely give you completely wrong results.
You would need to feed it the exact measured responses
of the drivers installed in the cabinet measured correctly.
and the impedance profiles of the drivers in situ.
I don't like xover Pro. it does loads of electrical textbook
filters that are totally irrelevant in practise, all that does
matter is the final acoustic response of the drivers.
It allows you to completely get a design totally wrong.
And nearly all beginners get 3 ways very wrong.
For why read some of the links below, particularly the
stuff on simulating the design of speakers. Xover Pro
just lets you ignore stuff that has to be accounted for
in the overall design.
rgds, sreten.
undefinition
(see if nothing else, the excellent FAQs)
The Speaker Building Bible - Techtalk Speaker Building, Audio, Video, and Electronics Customer Discussion Forum From Parts-Express.com
Zaph|Audio
FRD Consortium tools guide
Designing Crossovers with Software Only
RJB Audio Projects
Jay's DIY Loudspeaker Projects
Speaker Design Works
HTGuide Forum - A Guide to HTguide.com Completed Speaker Designs.
DIY Loudspeaker Projects Troels Gravesen
Humble Homemade Hifi
Quarter Wavelength Loudspeaker Design
The Frugal-Horns Site -- High Performance, Low Cost DIY Horn Designs
Linkwitz Lab - Loudspeaker Design
Music and Design
Great free SPICE Emulator : SPICE-Based Analog Simulation Program - TINA-TI - TI Tool Folder
This is what I was trying to get at. My intuition says there is a difference between the value of low end response and having a flat response when used in a 3-way. Of course, my girlfriend says my intuition will probably get me arrested.
It is clear I have large gaps in my understanding.
Sreten, that is good advice. For me, this entire process reenforces the need for both patience and proper technique. I'm getting ahead of myself, it is obvious. I'll take actual measurements after I've modeled and built the enclosure.
A friend has a copy of Leap Crossover Shop. I'll work with that once I have real data.
Thanks everyone for your patience and desire to help. I'll be back!
Hmm,
There are so many considerations other than Qtc being glossed over in this thread, and there are also two basic approaches to crossing over a midrange driver with a woofer that need to be contrasted.
Approach one is to put the mechanical resonance of the driver well below the intended cut-off frequency so that the electrical filter largely dictates the response through the crossover region.
So for example if your desired cut-off frequency was 300Hz, you would want the mechanical resonance of the driver to be below 150Hz or so, and with a modest Qtc. (0.5 < Qtc < 1)
I see a suggestion in the thread that the resonant frequency of the driver in the enclosure should be 1/4 of the cut-off frequency - 75Hz in this case, but I think that's very unrealistic and unnecessary.
If you have a driver that is large enough to have a free air resonance well below 75Hz, it's likely going to have a large Vas and need a very large enclosure to keep it below 75Hz, so a lot of space will be wasted from the bass enclosure. On the other hand a smaller driver with a small Vas that is going to work well in a small midrange enclosure is unlikely to have even a free air resonance below 75Hz, let alone in the enclosure.
Besides, a driver with a resonance of 75Hz in a small enclosure is hardly a midrange driver, its a woofer.
Advantages of this approach are that if you can get the mechanical resonance sufficiently low, you have plenty of flexibility in adjusting your cut-off frequency and roll off slope purely with the electrical filter, you're not operating the driver too close to its resonant frequency, (lower distortion) exact Qtc of the driver in its enclosure is not too important, and by choosing a driver large enough to have a sufficiently low resonance frequency you're probably selecting a driver that has performance to spare at the bottom end, in terms of excursion, power handling and so on.
Disadvantages of this method are that to get a given acoustical roll off slope you need more components, (all of the initial roll-off is coming from the crossover filter) it generally results in a lot of wasted space by needing quite a large midrange enclosure, (That can be good or necessary depending on the design for other reasons though) and unless your crossover is 18dB/oct or greater, excursion will remain constant (12dB/oct) or actually increase (6dB/oct) as you go down below the crossover frequency, until the mechanical resonance is reached.
So if your cut-off was 300Hz at 12dB/oct and the driver resonance was 150Hz the excursion at 150Hz would still be the same as 300Hz, which could be a problem for drivers with limited excursion.
Instead of trying to push the resonant frequency of the driver well below the crossover point, the second approach is to incorporate the 12dB/oct mechanical roll-off that occurs at resonance into your crossover slope, using a much simpler crossover, possibly 1st order.
This requires the resonant frequency of the driver in its enclosure to be at or very close to the cut-off frequency you want, depending on the exact configuration. At minimum you would want an electrical 1st order filter which would give you 18dB/oct acoustic roll off. A 2nd order filter would give you an acoustic rolloff of 24dB.
Advantages of this approach are less components for a given acoustic roll off slope, excursion begins dropping immediately below the cut-off frequency, a much smaller midrange enclosure becomes practical, and/or a driver with a much higher resonance frequency.
Disadvantages are the driver is operated at and below resonance (potentially higher distortion) slopes of less than 18dB/oct aren't possible while still having some minimal electrical filter, (possible difficulty in mating to the slope of the woofer low pass, or need for asymmetric crossover slopes) whilst the driver will be subjected to less excursion demands, it will be subject to greater power dissipation demands from bass frequencies since only part of the roll-off is electrical, the exact resonance frequency and Qtc become a critical part of the crossover alignment, so must be carefully adjusted, obtaining a high enough resonance frequency whilst simultaneously keeping Qtc low enough may be difficult with many drivers, requiring a resonance compensator.
Of the two I prefer the first method, although I do admit that it tends to require a relatively large midrange enclosure, and if you're putting a driver with a large Vas in a small enclosure the rise in Qtc would typically be such that you'd need to correct it somewhat with a resonance compensator, which is 3 large extra components.
Many midbass/midrange drivers when placed into modest sized enclosures will have a Qtc of greater that 0.7, so will have some peaking at the bottom end. Depending on where this resonance is in relation to the crossover frequency and baffle step frequency of the cabinet this may actually be an advantage, helping to lift the bottom end response and providing some of the baffle step correction required for that driver. If it's at an undesired frequency you might need a resonance compensator instead.
At the end of the day what matters is the acoustic transfer function with baffle step loss taken into account - how you achieve it doesn't really matter, but one thing is for sure, taking a midrange driver, throwing it in a sealed enclosure and applying a text book high pass passive filter to it will never work properly.
Crossing over a midrange driver with a woofer is actually quite tricky to get right especially with a passive network, and is a lot more challenging than applying baffle step correction in a 2 way.
When considering what volume midrange enclosure to use there is also more than just the resonance frequency and Qtc of the driver to consider - the size of the enclosure also dictates how much stuffing you can put in there to absorb standing waves.
If you make the enclosure too small then you simply can't put enough stuffing in and you will have standing wave problems throughout the midrange leading to re-radiated sound from the internal reflections in the cabinet, which will give a "boxy" sound. From this perspective its better to err on the side of making the enclosure a bit bigger rather than smaller, so you can have a generous amount of stuffing.
There are so many considerations other than Qtc being glossed over in this thread, and there are also two basic approaches to crossing over a midrange driver with a woofer that need to be contrasted.
Approach one is to put the mechanical resonance of the driver well below the intended cut-off frequency so that the electrical filter largely dictates the response through the crossover region.
So for example if your desired cut-off frequency was 300Hz, you would want the mechanical resonance of the driver to be below 150Hz or so, and with a modest Qtc. (0.5 < Qtc < 1)
I see a suggestion in the thread that the resonant frequency of the driver in the enclosure should be 1/4 of the cut-off frequency - 75Hz in this case, but I think that's very unrealistic and unnecessary.
If you have a driver that is large enough to have a free air resonance well below 75Hz, it's likely going to have a large Vas and need a very large enclosure to keep it below 75Hz, so a lot of space will be wasted from the bass enclosure. On the other hand a smaller driver with a small Vas that is going to work well in a small midrange enclosure is unlikely to have even a free air resonance below 75Hz, let alone in the enclosure.
Besides, a driver with a resonance of 75Hz in a small enclosure is hardly a midrange driver, its a woofer.
Advantages of this approach are that if you can get the mechanical resonance sufficiently low, you have plenty of flexibility in adjusting your cut-off frequency and roll off slope purely with the electrical filter, you're not operating the driver too close to its resonant frequency, (lower distortion) exact Qtc of the driver in its enclosure is not too important, and by choosing a driver large enough to have a sufficiently low resonance frequency you're probably selecting a driver that has performance to spare at the bottom end, in terms of excursion, power handling and so on.
Disadvantages of this method are that to get a given acoustical roll off slope you need more components, (all of the initial roll-off is coming from the crossover filter) it generally results in a lot of wasted space by needing quite a large midrange enclosure, (That can be good or necessary depending on the design for other reasons though) and unless your crossover is 18dB/oct or greater, excursion will remain constant (12dB/oct) or actually increase (6dB/oct) as you go down below the crossover frequency, until the mechanical resonance is reached.
So if your cut-off was 300Hz at 12dB/oct and the driver resonance was 150Hz the excursion at 150Hz would still be the same as 300Hz, which could be a problem for drivers with limited excursion.
Instead of trying to push the resonant frequency of the driver well below the crossover point, the second approach is to incorporate the 12dB/oct mechanical roll-off that occurs at resonance into your crossover slope, using a much simpler crossover, possibly 1st order.
This requires the resonant frequency of the driver in its enclosure to be at or very close to the cut-off frequency you want, depending on the exact configuration. At minimum you would want an electrical 1st order filter which would give you 18dB/oct acoustic roll off. A 2nd order filter would give you an acoustic rolloff of 24dB.
Advantages of this approach are less components for a given acoustic roll off slope, excursion begins dropping immediately below the cut-off frequency, a much smaller midrange enclosure becomes practical, and/or a driver with a much higher resonance frequency.
Disadvantages are the driver is operated at and below resonance (potentially higher distortion) slopes of less than 18dB/oct aren't possible while still having some minimal electrical filter, (possible difficulty in mating to the slope of the woofer low pass, or need for asymmetric crossover slopes) whilst the driver will be subjected to less excursion demands, it will be subject to greater power dissipation demands from bass frequencies since only part of the roll-off is electrical, the exact resonance frequency and Qtc become a critical part of the crossover alignment, so must be carefully adjusted, obtaining a high enough resonance frequency whilst simultaneously keeping Qtc low enough may be difficult with many drivers, requiring a resonance compensator.
Of the two I prefer the first method, although I do admit that it tends to require a relatively large midrange enclosure, and if you're putting a driver with a large Vas in a small enclosure the rise in Qtc would typically be such that you'd need to correct it somewhat with a resonance compensator, which is 3 large extra components.
Many midbass/midrange drivers when placed into modest sized enclosures will have a Qtc of greater that 0.7, so will have some peaking at the bottom end. Depending on where this resonance is in relation to the crossover frequency and baffle step frequency of the cabinet this may actually be an advantage, helping to lift the bottom end response and providing some of the baffle step correction required for that driver. If it's at an undesired frequency you might need a resonance compensator instead.
At the end of the day what matters is the acoustic transfer function with baffle step loss taken into account - how you achieve it doesn't really matter, but one thing is for sure, taking a midrange driver, throwing it in a sealed enclosure and applying a text book high pass passive filter to it will never work properly.
Crossing over a midrange driver with a woofer is actually quite tricky to get right especially with a passive network, and is a lot more challenging than applying baffle step correction in a 2 way.
When considering what volume midrange enclosure to use there is also more than just the resonance frequency and Qtc of the driver to consider - the size of the enclosure also dictates how much stuffing you can put in there to absorb standing waves.
If you make the enclosure too small then you simply can't put enough stuffing in and you will have standing wave problems throughout the midrange leading to re-radiated sound from the internal reflections in the cabinet, which will give a "boxy" sound. From this perspective its better to err on the side of making the enclosure a bit bigger rather than smaller, so you can have a generous amount of stuffing.
Last edited:
Thanks Simon for that detailed response! I really appreciate it!!
I'm going to have to digest this more thoroughly later, but I think I grasp the basic idea (but perhaps I don't! Please critique if possible!)
I chose the 2nd order roll off in order not to drive the mid too low, but to flatten the curve out I added a Zoebel to compensate. Pretty much your first scenario, yes? As you say, and as I am beginning to understand, being far from the Fs makes the enclosure less important. Still, the enclosure is meaningful and getting too high or too low a Qtc means less accuracy as a rule.
I'm going to test these drivers in the real world instead of relying on the data sheets. Gotta study much more to get a better grasp of what is what and then go at it again. I'll post as I can for those interested.
I'm going to have to digest this more thoroughly later, but I think I grasp the basic idea (but perhaps I don't! Please critique if possible!)
I chose the 2nd order roll off in order not to drive the mid too low, but to flatten the curve out I added a Zoebel to compensate. Pretty much your first scenario, yes? As you say, and as I am beginning to understand, being far from the Fs makes the enclosure less important. Still, the enclosure is meaningful and getting too high or too low a Qtc means less accuracy as a rule.
I'm going to test these drivers in the real world instead of relying on the data sheets. Gotta study much more to get a better grasp of what is what and then go at it again. I'll post as I can for those interested.
If you get a microphone and start measuring your actual drivers try these programs also
jbagby
But you will need the FRD files for each driver.
Sreten takes the experts view and wile he is right in many respects i have always advised people starting out to "just do it" as that is the easy way to get your mistakes out of the way, 3-Ways are hard I have several abandoned projects waiting for my skills to improve but I also have several successful ones.
One rule of thumb useful in building 3-ways is to have the midrange driver take up as much of the signal as possible and using the "telephony bandwidth" of 300 Hz to 3000Hz as a starting point is usually good advice.
As I bi-amp I can usually cross a little lower and maximum benefit seems to come from crossing around 120/ 150 Hz which some folk consider sub-woofer territory, but even 300Hz needs big inductor and capacitor values which is why some designs use a much higher XO point.
300Hz is where a lot of woofers happen to be at or near their nominal impedance too and in a region of flat response which makes XO design a little easier.
Any peaks in FR need to be at least 2 octaves away from the cross-over point and explains why some woofers and midranges are easier to work with than others.
Knowing which woofer and tweeter you are using may make suggestions easier
jbagby
But you will need the FRD files for each driver.
Sreten takes the experts view and wile he is right in many respects i have always advised people starting out to "just do it" as that is the easy way to get your mistakes out of the way, 3-Ways are hard I have several abandoned projects waiting for my skills to improve but I also have several successful ones.
One rule of thumb useful in building 3-ways is to have the midrange driver take up as much of the signal as possible and using the "telephony bandwidth" of 300 Hz to 3000Hz as a starting point is usually good advice.
As I bi-amp I can usually cross a little lower and maximum benefit seems to come from crossing around 120/ 150 Hz which some folk consider sub-woofer territory, but even 300Hz needs big inductor and capacitor values which is why some designs use a much higher XO point.
300Hz is where a lot of woofers happen to be at or near their nominal impedance too and in a region of flat response which makes XO design a little easier.
Any peaks in FR need to be at least 2 octaves away from the cross-over point and explains why some woofers and midranges are easier to work with than others.
Knowing which woofer and tweeter you are using may make suggestions easier
Thanks for the links! The more the better!
Yes, I started at 300 to 3K. Ended up just a tad closer. I was intending on just getting in the ballpark with X-Over Pro and tweak it with real world testing/analysis. But it is all tossed now as I just used the Dayton WT3 tester and my results were significantly different than the data sheet.
I'm also a "just do it" guy, trial by fire, etc. But I am also quite successful with many of my creations and I have fun no matter what the outcome. I knew I was biting off a heck of a lot with the 3-way and I am getting lots of lessons here!
I have built guitar and bass speakers for a while with good success, but they are a very different animal with somewhat different requirements than a a stereo speaker for general listening.
So I decided to give it a try and spent a lot of time reading on my own.
Noticing this, my girl decides to be so ultra sweet and asked some guy on the net what brand of driver to buy, etc. Next thing I know I have 2 CAW-938s, 2 EM-428s and 2 CAT-408s. A great thought and a great gift!! But the Qts of the woofer is rated at .6!! Could I make the best of it? After just measuring it, it is actually .8......feh. Best for car doors and OB (another huge can of worms for a far off day). But the CAW-938 is listed in Morel's audiophile speaker section, not car/auto. Do they expect this to be tri-amped and driven by some monstrous damping machine? Something like that......
My girl is very proud of the gift so I really want to make something worthwhile and say, "Sweetie, you done good! Aren't we having fun listening to music on the towers WE made together?!"
Maybe I can get different woofers and just not tell her the truth. Or just come clean?
Is there a way to salvage this situation? Dunno, but I'm gonna keep at it!
As ever, many thanks for all the input and great suggestions!!
Yes, I started at 300 to 3K. Ended up just a tad closer. I was intending on just getting in the ballpark with X-Over Pro and tweak it with real world testing/analysis. But it is all tossed now as I just used the Dayton WT3 tester and my results were significantly different than the data sheet.
I'm also a "just do it" guy, trial by fire, etc. But I am also quite successful with many of my creations and I have fun no matter what the outcome. I knew I was biting off a heck of a lot with the 3-way and I am getting lots of lessons here!
I have built guitar and bass speakers for a while with good success, but they are a very different animal with somewhat different requirements than a a stereo speaker for general listening.
So I decided to give it a try and spent a lot of time reading on my own.
Noticing this, my girl decides to be so ultra sweet and asked some guy on the net what brand of driver to buy, etc. Next thing I know I have 2 CAW-938s, 2 EM-428s and 2 CAT-408s. A great thought and a great gift!! But the Qts of the woofer is rated at .6!! Could I make the best of it? After just measuring it, it is actually .8......feh. Best for car doors and OB (another huge can of worms for a far off day). But the CAW-938 is listed in Morel's audiophile speaker section, not car/auto. Do they expect this to be tri-amped and driven by some monstrous damping machine? Something like that......
My girl is very proud of the gift so I really want to make something worthwhile and say, "Sweetie, you done good! Aren't we having fun listening to music on the towers WE made together?!"
Maybe I can get different woofers and just not tell her the truth. Or just come clean?
Is there a way to salvage this situation? Dunno, but I'm gonna keep at it!
As ever, many thanks for all the input and great suggestions!!
Did you pulse the woofer with LF for a while before using WT?
i'm told it helps to see the true values
Don't forget that 20% variance on nominal in normal and the inbox performance is probably the same, nice mid-woofers and the recommended 60 litre boxes sound reasonable.
Bottom line in this situation? GF/SO gave you a gift- you just HAVE to use them ( I got a quad of P-13s for my birthday and I still haven't used them ) Just tell her that they are wonderful woofer s and now you really reallly must buy the big 15/18 inch subs to go with them to get that bottom octave and a half.
That will need an electronic XO and another amp set of course
i'm told it helps to see the true values
Don't forget that 20% variance on nominal in normal and the inbox performance is probably the same, nice mid-woofers and the recommended 60 litre boxes sound reasonable.
Bottom line in this situation? GF/SO gave you a gift- you just HAVE to use them ( I got a quad of P-13s for my birthday and I still haven't used them ) Just tell her that they are wonderful woofer s and now you really reallly must buy the big 15/18 inch subs to go with them to get that bottom octave and a half.
That will need an electronic XO and another amp set of course
Yes, gotta break these babies in. But thought I'd check them before, just as a learning thing. But I doubt they will end up with the stats Morel has published. Just goes to show that one cannot do a good job by hurrying.
The sub thing is a good idea. Just sent an old BGW 750m off to Russ Sherwood at Eagle amps in Kansas City. It will have plenty to drive a solid sub. Perhaps a small plate amp in each cab just to drive the woofers?
But this thread is digressing. I believe I get the gist of how to look at Qts and its relationship to Qtc, box size and driver function. Many thanks for the fine input!
The sub thing is a good idea. Just sent an old BGW 750m off to Russ Sherwood at Eagle amps in Kansas City. It will have plenty to drive a solid sub. Perhaps a small plate amp in each cab just to drive the woofers?
But this thread is digressing. I believe I get the gist of how to look at Qts and its relationship to Qtc, box size and driver function. Many thanks for the fine input!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.