Hello,
Reading abundant material on crossover design I can find no guide as to how to select an octave or decade relationship between lower and upper crossover frequencies. Anyone please advise?
Also, given the spec sheet for a certain driver and response graph I see that a specific 10" bass driver has a natural resonance of 23Hz. Should I avoid a crossover frequency equal to a multiple of this number or should I be choosing one which is a multiple?
Without a 'recommended crossover frequency' from the manufactuer, what specification values do I need to determine the 'best' crossover frequency, or at least a good one to start with....?
Many thanks in advance.
Reading abundant material on crossover design I can find no guide as to how to select an octave or decade relationship between lower and upper crossover frequencies. Anyone please advise?
Also, given the spec sheet for a certain driver and response graph I see that a specific 10" bass driver has a natural resonance of 23Hz. Should I avoid a crossover frequency equal to a multiple of this number or should I be choosing one which is a multiple?
Without a 'recommended crossover frequency' from the manufactuer, what specification values do I need to determine the 'best' crossover frequency, or at least a good one to start with....?
Many thanks in advance.
A decade is the fundamental frequency by a number of 10. So, 1Khz to 10Khz for example. An octave is doubling of the frequency. First order function would have a 6db/octave or 20db/decade slope (idealistic). Do the math and you`d see how they`re interrelated.
Fs of the driver has zero impact on the desired lowpass point (well, unless it is at or around Fs where magnitude would affect the transfer function but this is unrelated to low frequency drivers). Where to crossover depends on many factors, but often the drivers would point you to the right direction. Don`t use online calculators - they`re pretty useless. Better read something like the loudspeaker cook book and aim to understand 2nd order crossovers, then try a 2-way comprised of paper/poly cone drivers, these are more forgiving.
Fs of the driver has zero impact on the desired lowpass point (well, unless it is at or around Fs where magnitude would affect the transfer function but this is unrelated to low frequency drivers). Where to crossover depends on many factors, but often the drivers would point you to the right direction. Don`t use online calculators - they`re pretty useless. Better read something like the loudspeaker cook book and aim to understand 2nd order crossovers, then try a 2-way comprised of paper/poly cone drivers, these are more forgiving.
Last edited:
To answer your question, there is the "rule" of doubling the frequency from the resonant frequency for a HP filter. So for a tweeter with a Fs of 1KHz, the crossover point shouldn't be less than 2KHz, else you'll limit the power the driver can sustain, so limiting the SPL. The same is true for a midrange.
Another "rule" for the woofer-tweeter crossover point is that you should not go higher than the frequency at which the woofer is -3dB on the 30 degree off axis FR compared to on axis. So you'll avoid the beaming problem on the woofer, considering the fact that at low frequencies (for a tweeter), the tweeter response is omnidirectional.
But I urge to say you, learn to walk before running. Thinking to design a 3 way without any base is foolish. Buy the Loudspeaker Design Cookbook by V. Dickason, and then if you still want to design a speaker, go for a 2 way.
Ralf
Quite often the manufacturer recommended frequency range is overly optimistic.
Another "rule" for the woofer-tweeter crossover point is that you should not go higher than the frequency at which the woofer is -3dB on the 30 degree off axis FR compared to on axis. So you'll avoid the beaming problem on the woofer, considering the fact that at low frequencies (for a tweeter), the tweeter response is omnidirectional.
But I urge to say you, learn to walk before running. Thinking to design a 3 way without any base is foolish. Buy the Loudspeaker Design Cookbook by V. Dickason, and then if you still want to design a speaker, go for a 2 way.
Ralf
Quite often the manufacturer recommended frequency range is overly optimistic.
Mostly, it depends on the mass ( moving mass, Mms ) of the cone, and its ability to "move quick". So, for a 10" woofer having very low Fs ( 23 Hz), it's possible that its moving mass ( which is the weight of the cone+ the coil+former+some part of the suspension ) is ~ 100 g.
Also the diameter of the cone dictates something, as the cone should behave like a perfect piston but it doesn't ( the peripheal part of it responds differently to solicitation respect to the center, that's why the break-up occurs ) so the larger the cone, the less a pistonic behavior is expected; aluminum cones are very rigid but ringing happens because of the metal nature, that's why the manufacturers make multi-layers with different materials.
Also the diameter of the cone dictates something, as the cone should behave like a perfect piston but it doesn't ( the peripheal part of it responds differently to solicitation respect to the center, that's why the break-up occurs ) so the larger the cone, the less a pistonic behavior is expected; aluminum cones are very rigid but ringing happens because of the metal nature, that's why the manufacturers make multi-layers with different materials.
For a 3-way system, it is often useful to cross over at the baffle step frequency if possible (depending on your choice of woofer and mid).
You can also cross over somewhere else and use global baffle step, or one built into the individual transfer functions.
For low-mid crossover below 300Hz, the phase transition appears to not be too critical, but the impedance could be an issue. A few large drivers typically have 6 or 4 ohm nominal impedance, so the amplifier will be working fairly hard in a range with lots of musical energy and at some points, two coils in parallel.
I've had a few issues mating two 8 ohm drive units in that respect, the resulting impedance curve makes it suitable for 4 ohm drive only. I can't even begin to imagine what a 4 ohm woofer would look like.
You can also cross over somewhere else and use global baffle step, or one built into the individual transfer functions.
For low-mid crossover below 300Hz, the phase transition appears to not be too critical, but the impedance could be an issue. A few large drivers typically have 6 or 4 ohm nominal impedance, so the amplifier will be working fairly hard in a range with lots of musical energy and at some points, two coils in parallel.
I've had a few issues mating two 8 ohm drive units in that respect, the resulting impedance curve makes it suitable for 4 ohm drive only. I can't even begin to imagine what a 4 ohm woofer would look like.
Guys - many thanks for the deliberations and quick responses... In answer to some of the queries above, I have been given a few drivers for a project, th bass driver is a Monacor SPH-250KE with Fs=23Hz and Mms of 45g.
I understand the meaning of Octave and Decade but was wondering under what circumstances you choose the relationshipo between crossover frequencies, or just ignore and go for totally unrelated frequencies based on driver data.
The mid I have is Monacor again, being SPH-102KEP with an Fs of 95Hz, and range of fx - 10kHz, but fz is not specified... Mms is 3g. Both drivers are 8 Ohms...
An SEAS 1" dome tweeter is also being considered.
The suggestion is a 2nd-order Linkwitz-Reilly passive design...
Hopefully this is a bit more info to steer me in the right direstion.
Many thanks in advance 🙂
I understand the meaning of Octave and Decade but was wondering under what circumstances you choose the relationshipo between crossover frequencies, or just ignore and go for totally unrelated frequencies based on driver data.
The mid I have is Monacor again, being SPH-102KEP with an Fs of 95Hz, and range of fx - 10kHz, but fz is not specified... Mms is 3g. Both drivers are 8 Ohms...
An SEAS 1" dome tweeter is also being considered.
The suggestion is a 2nd-order Linkwitz-Reilly passive design...
Hopefully this is a bit more info to steer me in the right direstion.
Many thanks in advance 🙂
If you already have the drivers on hand, I would suggest getting their data into a simulator along with a few popular tweeters and see what works best before committing to a design. This will save you a lot of heartache. You can do this with actual measurements, but even using manufacturer data will be helpful. You need the .frd and .zma files (if not available, you'll have to trace them manually), and don't forget to enter driver offsets in the simulator.
In the end you may land up with a 3rd order slope on one driver and a 5th order on the other (or maybe not), the thing to optimise is the target response and phase blending. You also need to consider beaming in the mid-tweeter crossover, as well as the individual driver's distortion characteristics. Sticking a standard topology on any driver combination is at best an initial wish, in the end it may be that your final XO differs from this quite a bit.
Edit: A useful method to start a design (with a dome tweeter, at least) is to look at the highest frequency you can cross at without getting the mid driver to beam. For a 4" driver that would be roughly 2.5KHz, maybe 3kHz. That would be the high crossover, I would choose the lower end of the range. A 10" bass driver would typically mount on a 12.5" wide baffle and the baffle step frequency would be around 360Hz. Those two XO points would be a nice place to start, IMO. Rest you can optimise as you go along.
In the end you may land up with a 3rd order slope on one driver and a 5th order on the other (or maybe not), the thing to optimise is the target response and phase blending. You also need to consider beaming in the mid-tweeter crossover, as well as the individual driver's distortion characteristics. Sticking a standard topology on any driver combination is at best an initial wish, in the end it may be that your final XO differs from this quite a bit.
Edit: A useful method to start a design (with a dome tweeter, at least) is to look at the highest frequency you can cross at without getting the mid driver to beam. For a 4" driver that would be roughly 2.5KHz, maybe 3kHz. That would be the high crossover, I would choose the lower end of the range. A 10" bass driver would typically mount on a 12.5" wide baffle and the baffle step frequency would be around 360Hz. Those two XO points would be a nice place to start, IMO. Rest you can optimise as you go along.
Last edited:
STOP! [insert screeching sounds here]. Just hang fire a second.
Suggestion from whom?
A 'real' LR2 would almost certainly need some form of stepped or tilted baffle to align the acoustic centres or you'll end up with a ruddy great null at the crossover frequency.
Generally there isn't a rule. There are many things that will serve as guides (Mms can be one, but is a little esoteric). Generally you select based on the on & off-axis frequency responses, distortion performance, power-handling etc. The fact that you don't seem to know this (that's not a reflection or suggesting you're an idiot, it's just an observation) suggests you're new to this. In which case:
It should be stated right now that on-line calculators, textbook formulas etc. do not work in practice. They provide electrical slopes, but do not account for the driver's own acoustical slopes & behaviour, phase etc. For e.g., say you have a tweeter rolling off 2nd order at 2KHz. If you then put a 2nd order electrical high pass filter on it, this will become an acoustic 4th order slope: electrical and natural driver slopes are additive. Nor will the drivers be 8ohms -that's a nominal only. Re is also a nominal: impedance changes with frequency & this needs to be accounted for in the filter design, either with flattening (RC Zobel shunted across the driver) or in the topology and design values of the filter itself.
If you're not familiar with any of this, my immediate advice: if you just want to build something, build a proven, well-documented design. If you're wanting to learn about designing loudspeakers, stop right now and don't even think about buying drivers yet: you'll only regret it, because there's a high chance you'll just be pouring money down the drain. For a start, Paul's site here is an excellent introduction: https://sites.google.com/site/undefinition/diy He has some quality, reasonably priced projects there, with links to more. Then move to the textbooks if you want to go into things in much more detail, e.g. The Loudspeaker Design Cookbook.
Other sites with quality 3-way design projects include Zaph|Audio and DIY Loudspeaker Projects Troels Gravesen Troels has moved more toward providing kits, but the older projects are still there, and very good. Likewise Tony's site here in the Archive section: Humble Homemade Hifi
Suggestion from whom?
A 'real' LR2 would almost certainly need some form of stepped or tilted baffle to align the acoustic centres or you'll end up with a ruddy great null at the crossover frequency.
Generally there isn't a rule. There are many things that will serve as guides (Mms can be one, but is a little esoteric). Generally you select based on the on & off-axis frequency responses, distortion performance, power-handling etc. The fact that you don't seem to know this (that's not a reflection or suggesting you're an idiot, it's just an observation) suggests you're new to this. In which case:
It should be stated right now that on-line calculators, textbook formulas etc. do not work in practice. They provide electrical slopes, but do not account for the driver's own acoustical slopes & behaviour, phase etc. For e.g., say you have a tweeter rolling off 2nd order at 2KHz. If you then put a 2nd order electrical high pass filter on it, this will become an acoustic 4th order slope: electrical and natural driver slopes are additive. Nor will the drivers be 8ohms -that's a nominal only. Re is also a nominal: impedance changes with frequency & this needs to be accounted for in the filter design, either with flattening (RC Zobel shunted across the driver) or in the topology and design values of the filter itself.
If you're not familiar with any of this, my immediate advice: if you just want to build something, build a proven, well-documented design. If you're wanting to learn about designing loudspeakers, stop right now and don't even think about buying drivers yet: you'll only regret it, because there's a high chance you'll just be pouring money down the drain. For a start, Paul's site here is an excellent introduction: https://sites.google.com/site/undefinition/diy He has some quality, reasonably priced projects there, with links to more. Then move to the textbooks if you want to go into things in much more detail, e.g. The Loudspeaker Design Cookbook.
Other sites with quality 3-way design projects include Zaph|Audio and DIY Loudspeaker Projects Troels Gravesen Troels has moved more toward providing kits, but the older projects are still there, and very good. Likewise Tony's site here in the Archive section: Humble Homemade Hifi
Last edited:
Hi,
Already having the drivers = a really steep learning
curve to do any sort of a good job. See here :
FRD Consortium tools guide
and :
FRD tools example 2
Your problem is getting decent data on your drivers.
The bass unit, in a 3 way you could probably fudge, but
the mid unit you absolutely need to know its response
at the top end to design a proper x/o to the tweeter.
rgds, sreten.
Already having the drivers = a really steep learning
curve to do any sort of a good job. See here :
FRD Consortium tools guide
and :
FRD tools example 2
Your problem is getting decent data on your drivers.
The bass unit, in a 3 way you could probably fudge, but
the mid unit you absolutely need to know its response
at the top end to design a proper x/o to the tweeter.
rgds, sreten.
- Status
- Not open for further replies.