Ideally I'd like to use a Kef XQ10 coaxial driver I bought without the crossover/enclosure, and I'm a little unsure if there's a different design path for drivers that are already time and perhaps, acoustic center/phase coherent.
It's a 5.25" poly mid with .75" aluminum dome tweeter and I'm thinking about using it in a low-ish power center channel installation. (45 watts RMS, Fosgate mono amp)
on the driver, it's marked "4 ohm" and the statistics by Kef for their speaker has it crossed at 2.0Khz. The impedance profile is minimum 3.2 ohms, according to the literature. I don't know the sensitivity of the individual driver portions but I believe the system is rated at 6 ohms impedance and the sensitivity is 87 db. Other systems using this driver cross at 2.6Khz. I don't have any information on the slopes or OEM design specs.
Ideally I'd just put a cap on the tweeter/coil on the woofer, and hope for the best, but after reading about cone shapes and ideal horn profiles I can see that there is a chance this driver combination might need a little shaping if I'm going to have it right. Kef is known for complicated crossovers but they tout a simpler network for this particular driver combo.
anyone who has designed a crossover for, or has a working design for, or best of all, a schematic of this particular speaker's crossover, would be appreciated for their time in helping.
It's a 5.25" poly mid with .75" aluminum dome tweeter and I'm thinking about using it in a low-ish power center channel installation. (45 watts RMS, Fosgate mono amp)
on the driver, it's marked "4 ohm" and the statistics by Kef for their speaker has it crossed at 2.0Khz. The impedance profile is minimum 3.2 ohms, according to the literature. I don't know the sensitivity of the individual driver portions but I believe the system is rated at 6 ohms impedance and the sensitivity is 87 db. Other systems using this driver cross at 2.6Khz. I don't have any information on the slopes or OEM design specs.
Ideally I'd just put a cap on the tweeter/coil on the woofer, and hope for the best, but after reading about cone shapes and ideal horn profiles I can see that there is a chance this driver combination might need a little shaping if I'm going to have it right. Kef is known for complicated crossovers but they tout a simpler network for this particular driver combo.
anyone who has designed a crossover for, or has a working design for, or best of all, a schematic of this particular speaker's crossover, would be appreciated for their time in helping.
I will point out to you that with 2nd order filters the stop band is eventually 180 degrees out of phase....designers on the cheap con't care or worry about this as the drivers are not time aligned. But with a driver such as a coicident source Seas or Kef they are.
So at some point in the out put of the stop band there is potential for cancellations.
createing jagged response from phase cancelllations...I would look at 4th order crossovers so as to limit output and interaction of your tweeter and cone horn and minimize phase errors as on it's way to a complete phase reversal it is going to be 180 degrees out of phase for very briefly....also with a fourth order you lighten the load and reduce the harsh low frequency output of the tweeter and limit the higher frequency cone breakup....for example B&W and Focal nearly exclusively use 4th order crossovers to keep their drivers sounding clean. . . component tolerance is critical in 4th order crossovers....and depending on your calculation sources you may need to compensate for peaks and lulls at the crossover point with crossover offset....I think butterworths or known for a 2-3 db peak at crossover point and linkwitz 4th orders or cascaded butterworths....one of the filter caluculations causes a lull...you can compensate for it or use it to eq the frequency response of a driver combination .. . .look it up and or use simulators that acknowledge or compensate for these effects or use the effects to your advantage as I just noted... . there is some heady freeware stuff out there that is difficult to use but rather informative.....like I learned from one simulator that the best transient response comes from a bessel filter and the gentle roll off as it approachs the crossover point can be used to eq a hot spot in a driver . . . measure your drivers for loaded impeadence curves and utilize factor response curves to layout a goal for your desing....gausian seems attractive for a peak but has real bad transient response...have a nice day....
So at some point in the out put of the stop band there is potential for cancellations.
createing jagged response from phase cancelllations...I would look at 4th order crossovers so as to limit output and interaction of your tweeter and cone horn and minimize phase errors as on it's way to a complete phase reversal it is going to be 180 degrees out of phase for very briefly....also with a fourth order you lighten the load and reduce the harsh low frequency output of the tweeter and limit the higher frequency cone breakup....for example B&W and Focal nearly exclusively use 4th order crossovers to keep their drivers sounding clean. . . component tolerance is critical in 4th order crossovers....and depending on your calculation sources you may need to compensate for peaks and lulls at the crossover point with crossover offset....I think butterworths or known for a 2-3 db peak at crossover point and linkwitz 4th orders or cascaded butterworths....one of the filter caluculations causes a lull...you can compensate for it or use it to eq the frequency response of a driver combination .. . .look it up and or use simulators that acknowledge or compensate for these effects or use the effects to your advantage as I just noted... . there is some heady freeware stuff out there that is difficult to use but rather informative.....like I learned from one simulator that the best transient response comes from a bessel filter and the gentle roll off as it approachs the crossover point can be used to eq a hot spot in a driver . . . measure your drivers for loaded impeadence curves and utilize factor response curves to layout a goal for your desing....gausian seems attractive for a peak but has real bad transient response...have a nice day....
Greetings.
I have no knowledge of the KEF driver of which you speak.
High order crossovers, as mentioned above, have their place and are often mandatory for drivers with rugged out-of-band responses. They are also comparitively easy to design by 'text book', using any number of free or commercial software programs. Second order crossovers are plagued with issues.
If you have test and measurement facilities, and a good understanding of crossover topology interactions, and a heap of inductors and capacitors to hand, then by all means embark on that project.
Before you do though, I urge you to try your system with a simple cap in series with the tweeter, and nothing at all on the bass/mid. Try a 2.2 or 3.3uF, try it with the tweeter phase normal and inverted, and see what you think. Lots of dual concentrics work a treat like this, although some end up with a broad midrange hump. You are lucky there, because your use of a dedicated amp means it is quite easy to implement an active (or line level passive) filter, or even a passive low-Q tuned circuit to pull back the midrange.
Don't be blinded by the inevitable Himalayas in the top end above about 10k. Learn to look through the curves. Either take average measurements across a few angles off axis or try one of those horrid L-Pad variable attenuators to establish what static L-Pad resistor values will set the treble level to your liking. If you have the equipment, the time and the patience to mess around, you can get great results. Otherwise, use your ears, They are better than most people think.
I once designed a commercial Dual Concentric speaker (still in production 13 years later!) that was almost perfect with a single cap crossover. It had a broad, shallow midrange hump. I loved it, and would have left it that way. The company's ethos though was flat frequency response, and an 11 element crossover (of the highest quality components) was needed to get the response within +/- 1dB through the midband. Took 2 weeks to design the crossover, and it didn't look like a 'text book' anything. It took hundreds of measurements and tweaks to get it there.
So always try simple first - you might be surprised and save yourself a lot of work.
As always, like a broken record, I suggest you investigate the MiniDSP either as an active crossover of selectable slope, or as an EQ for a simply crossed over but lumpy-ish system.
Regards,
Keith Arnold
blakkshepeaudio
I have no knowledge of the KEF driver of which you speak.
High order crossovers, as mentioned above, have their place and are often mandatory for drivers with rugged out-of-band responses. They are also comparitively easy to design by 'text book', using any number of free or commercial software programs. Second order crossovers are plagued with issues.
If you have test and measurement facilities, and a good understanding of crossover topology interactions, and a heap of inductors and capacitors to hand, then by all means embark on that project.
Before you do though, I urge you to try your system with a simple cap in series with the tweeter, and nothing at all on the bass/mid. Try a 2.2 or 3.3uF, try it with the tweeter phase normal and inverted, and see what you think. Lots of dual concentrics work a treat like this, although some end up with a broad midrange hump. You are lucky there, because your use of a dedicated amp means it is quite easy to implement an active (or line level passive) filter, or even a passive low-Q tuned circuit to pull back the midrange.
Don't be blinded by the inevitable Himalayas in the top end above about 10k. Learn to look through the curves. Either take average measurements across a few angles off axis or try one of those horrid L-Pad variable attenuators to establish what static L-Pad resistor values will set the treble level to your liking. If you have the equipment, the time and the patience to mess around, you can get great results. Otherwise, use your ears, They are better than most people think.
I once designed a commercial Dual Concentric speaker (still in production 13 years later!) that was almost perfect with a single cap crossover. It had a broad, shallow midrange hump. I loved it, and would have left it that way. The company's ethos though was flat frequency response, and an 11 element crossover (of the highest quality components) was needed to get the response within +/- 1dB through the midband. Took 2 weeks to design the crossover, and it didn't look like a 'text book' anything. It took hundreds of measurements and tweaks to get it there.
So always try simple first - you might be surprised and save yourself a lot of work.
As always, like a broken record, I suggest you investigate the MiniDSP either as an active crossover of selectable slope, or as an EQ for a simply crossed over but lumpy-ish system.
Regards,
Keith Arnold
blakkshepeaudio
thank you for the responses, both.
Keith Arnold, your thought process is parallel with my own, I am unwilling to make the measurements that describe the various unique points this driver has in it's response, but going with generalities, I would say that the deep curvilinear cone, designed to function as a perfunctory horn loading for the tweeter outside of the tangerine waveguide's stationary axis, is also well suited for the response obtained where there are very little break-up modes that are large in proximity, there should be no Himalayas, more a gentle Dover cliff to the sea..
if the cone was more straight-sided and made of stiffer materials than a polymer with titanium deposition, like that of a metal construction, I'd see the Kilimanjaro somewhere out on the 4-9Khz plains.
So I believe leaving the coil off of the midrange will be effectually a matter of taste, but one of my primary concerns was the way the null from the tweeter response, being time-coherent and apparently nearly in phase, would be an extended depressive run for the 3-6Khz region.
Low powered as this design will remain, I don't know if such shallow slope crossing the .75" aluminum dome at 6K or even 8Khz, will affect it's long-term survival or not, I know many car audio tweeters are designed this way, and they live so I don't see why not?
from what I understand about 6 db/oct. or first order electric crossovers, the woofer and tweeter are supposed to be in phase at the crossover point, because the phase rotates just 90 degrees each way, leaving a unity gain situation... I didn't say I knew a lot about these things, lol...
now, the question is whether or not using a coil on the woofer is going to (theoretically) achieve a flatter response, if the sensitivity of both drivers is similar throughout the crossover region, or...
just doing without a coil and letting the possibly smooth woofer response roll off naturally ends up the better bet?
I've got plenty of poly caps to try the cap-only approach, but would have to order a few coils to figure this out, so that's more of where I'm coming from. How will I know if the response is valid, or correct, or just plain wrong, because it's a center channel and there's no way to test it against it's brother?
I've got Polk LSi 7's, to either side so maybe I'll just weigh out how good it sounds against those and if that's close enough, leave good enough alone..
Keith Arnold, your thought process is parallel with my own, I am unwilling to make the measurements that describe the various unique points this driver has in it's response, but going with generalities, I would say that the deep curvilinear cone, designed to function as a perfunctory horn loading for the tweeter outside of the tangerine waveguide's stationary axis, is also well suited for the response obtained where there are very little break-up modes that are large in proximity, there should be no Himalayas, more a gentle Dover cliff to the sea..
if the cone was more straight-sided and made of stiffer materials than a polymer with titanium deposition, like that of a metal construction, I'd see the Kilimanjaro somewhere out on the 4-9Khz plains.
So I believe leaving the coil off of the midrange will be effectually a matter of taste, but one of my primary concerns was the way the null from the tweeter response, being time-coherent and apparently nearly in phase, would be an extended depressive run for the 3-6Khz region.
Low powered as this design will remain, I don't know if such shallow slope crossing the .75" aluminum dome at 6K or even 8Khz, will affect it's long-term survival or not, I know many car audio tweeters are designed this way, and they live so I don't see why not?
from what I understand about 6 db/oct. or first order electric crossovers, the woofer and tweeter are supposed to be in phase at the crossover point, because the phase rotates just 90 degrees each way, leaving a unity gain situation... I didn't say I knew a lot about these things, lol...
now, the question is whether or not using a coil on the woofer is going to (theoretically) achieve a flatter response, if the sensitivity of both drivers is similar throughout the crossover region, or...
just doing without a coil and letting the possibly smooth woofer response roll off naturally ends up the better bet?
I've got plenty of poly caps to try the cap-only approach, but would have to order a few coils to figure this out, so that's more of where I'm coming from. How will I know if the response is valid, or correct, or just plain wrong, because it's a center channel and there's no way to test it against it's brother?
I've got Polk LSi 7's, to either side so maybe I'll just weigh out how good it sounds against those and if that's close enough, leave good enough alone..
Last edited:
Greets.
It only takes simple and inexpensive test equipment to design passive crossovers.
If you plan on more speaker projects it might be a good investment, and is the only way to really know what you are getting.
There are lots of ways to design without test equipment and I wouldn't presume to describe them all as you seem to have done your homework.
I would simply make the following points.
1-Good luck getting a flat HF response from a DC driver.
2-The more complex a crossover you attempt, the more likely you are to come unstuck without test equipment.
3-Text book crossovers seldom work as desired. Some here claim they have made successful designs using simulations and spreadsheets. I do not dispute this but my experience is to the contrary.
4-Phase response of crossovers is less important than the phase response of complete systems. Think about the phase relationships between a naked bass/mid driver and a tweeter with a cap on it.
Good luck with your adventure.
blakk
It only takes simple and inexpensive test equipment to design passive crossovers.
If you plan on more speaker projects it might be a good investment, and is the only way to really know what you are getting.
There are lots of ways to design without test equipment and I wouldn't presume to describe them all as you seem to have done your homework.
I would simply make the following points.
1-Good luck getting a flat HF response from a DC driver.
2-The more complex a crossover you attempt, the more likely you are to come unstuck without test equipment.
3-Text book crossovers seldom work as desired. Some here claim they have made successful designs using simulations and spreadsheets. I do not dispute this but my experience is to the contrary.
4-Phase response of crossovers is less important than the phase response of complete systems. Think about the phase relationships between a naked bass/mid driver and a tweeter with a cap on it.
Good luck with your adventure.
blakk
that does seem to be the issue, I know there have been several car audio DC (dual concentric, I hope) drivers in the past, the Memphis MSYNC and the Aria, Orion had one, of course, Kef with the Uni-Q, and I'm coming to the realization that using a woofer cone for a tweeter horn loading, probably means a complex crossover is necessary. These car audio products never seemed to capture the market, and I am pretty sure their relative crossover complexity was hemmed in by bean counters and actual space limitations, but in order to produce a reference level product, the coincident source driver may need more massaging in the signal gymnastics than the separated source counterpart.
I would have thought that the acoustic centers being aligned would have mitigated the crossover complexity issue but it apparently exacerbates it. I would guess that had this been untrue, the market for coincident source drivers would be greater today, with a larger selection and a competitive marketplace filled with them.
I'm going to put a cap on it and let it ride, for a while. While I contemplate my crossover options.
I would have thought that the acoustic centers being aligned would have mitigated the crossover complexity issue but it apparently exacerbates it. I would guess that had this been untrue, the market for coincident source drivers would be greater today, with a larger selection and a competitive marketplace filled with them.
I'm going to put a cap on it and let it ride, for a while. While I contemplate my crossover options.
- Status
- Not open for further replies.