We gave done a number of speakers with XO in this region, usually 1st order. There are some very good wide-rangers that excel when used with helper woofers. Getting the XO this low also means you have a good chance of getting the drivers to or closer than the magical 1/4 wl C-C spacing.
dave
Given that no crossover can ever have a perfect polar response, aren't the two statements above contradictions? The polar response is indeed the "main thing", but there is no solution, so isn't that a problem? And it is made better or worse by various choices, mostly of the frequency location.
I have to agree with Earl here.....phase interaction at the XO point can never be perfect....or as nice as say a wideband driver that avoids the issue entirely.
That being said, i certainly respect your point as to what's likely audible and what isn't where a properly done passive network is concerned......but then again, it can be avoided along with the exampled pear shaped polars. Worth the effort IMO and experiences which are based on subjective listening.
We all know we can't measure everything we hear and it's precisely the reason we have such a wide variety of drivers and materials to choose from. Whether it's in a CSD or HD plot where a driver's magic tone comes from is up the designer to decide. People LOVE ribbons regardless of their ragged HD smack in the XO range......go figure?
That being said, i certainly respect your point as to what's likely audible and what isn't where a properly done passive network is concerned......but then again, it can be avoided along with the exampled pear shaped polars. Worth the effort IMO and experiences which are based on subjective listening.
We all know we can't measure everything we hear and it's precisely the reason we have such a wide variety of drivers and materials to choose from. Whether it's in a CSD or HD plot where a driver's magic tone comes from is up the designer to decide. People LOVE ribbons regardless of their ragged HD smack in the XO range......go figure?
Yes it is a contradiction in that it can never be 'perfect', but I don't think that any loudspeaker is ever going to be perfect. Designs are all about trade offs and compromises and choosing to place the xover, for example, at say 5kHz, to get it out of that critical band, whilst heavily sacrificing the off axis, is not a satisfactory one. Especially if crossing at 2.5kHz would solve the majority of the off axis issues without sacrificing anything other than the fact that the loudspeaker is no longer crossed at 5kHz.
I think that a critical point is being missed here.
I can make a "perfect" crossover in both phase and magnitude, but only at a single point in space. As long as the two drivers being crossed are at different spatial locations there will always be off-axis aberrations and these are never a "good" thing. It is almost ubiquitous that people ignore the off axis effects when they talk about crossovers, but, to me, these are the problems that need to be resolved - they are not insignificant.
The more crossover points there are the more problems there are. The higher in frequency they are the bigger the problems (because of driver spacing). So lower and fewer crossovers are the rule as far as I am concerned, and they must be designed for optimal off axis performance over the performance at some arbitrary spatial angle. To my knowledge this is virtually never done, but it is the key to a superior sound quality system.
I can make a "perfect" crossover in both phase and magnitude, but only at a single point in space. As long as the two drivers being crossed are at different spatial locations there will always be off-axis aberrations and these are never a "good" thing. It is almost ubiquitous that people ignore the off axis effects when they talk about crossovers, but, to me, these are the problems that need to be resolved - they are not insignificant.
The more crossover points there are the more problems there are. The higher in frequency they are the bigger the problems (because of driver spacing). So lower and fewer crossovers are the rule as far as I am concerned, and they must be designed for optimal off axis performance over the performance at some arbitrary spatial angle. To my knowledge this is virtually never done, but it is the key to a superior sound quality system.
True, everything that we are talking about here are degrees of problems, greater or lessor. But when one requires narrow directivity control, which requires large areas, the 1/4 wl point is very hard to meet. If you don't care about a narrow directivity or about controlled directivity at all then things get much easier, but that's throwing the baby out with the bathwater.
If there's a properly designed one suitable for home use available to DIYers, I'm all over it. The new one from SEAS seems to be just that but awefully expensive at $429.00
I am not sure what "properly designed" means, but I have tried the B&C coaxes. They certainly know what they are doing so I assume that they are "properly designed". The lobbing problem is alleviated but a whole new set of problems ensues. The waveguides are, by necessity, too small and this causes poor polar control and internal standing waves.
I have to say that for simplicity I would just buy a pair of B&C coaxes and a miniDSP in an open baffle, with subs. Its not a great solution but it is certainly a high-value one. I posted some results here at DIY some time ago on this study and I offered to make the design available, but there wasn't any interest. As I said, the results weren't great - by my standards - but I'll bet they were better than about 80% of the stuff that I see around here.
By properly designed I meant a coax done with the intention of using the cone of the mid section specifically as a wave-guide. More specifically, with proper effort put into making the throat and cone edge termination (what would be the surround) work well in conjunction with the tweeter to keep the response looking decent.
SEAS latest coax drivers do this very well, as do some of KEFs latest offerings, as do some of TAD/Pioneers stuff. The only objection I can see from your point of view is that they aren't on a large enough scale to work anywhere near as effectively as what you typically go for.
Given your preferences though, it would almost seem worthwhile in getting a manufacturer to custom build a suitable driver. One where the cone profile, edge termination and throat were designed specifically to give you the kind of directivity control you're after, but without compromising them for the low frequency and raw SPL capabilities of the mid/bass section.
All of the aforementioned coax drivers are mid/tweeters only because the edge termination/surround required doesn't provide the kind of travel necessary for decent bass. I would have thought though that if you were to use a 12 or 15" driver, that only having a limited amount of linear excursion wouldn't really be a huge limitation. Given the domestic environment I bet you don't really use more than a mm or two with your 15" designs.
SEAS latest coax drivers do this very well, as do some of KEFs latest offerings, as do some of TAD/Pioneers stuff. The only objection I can see from your point of view is that they aren't on a large enough scale to work anywhere near as effectively as what you typically go for.
Given your preferences though, it would almost seem worthwhile in getting a manufacturer to custom build a suitable driver. One where the cone profile, edge termination and throat were designed specifically to give you the kind of directivity control you're after, but without compromising them for the low frequency and raw SPL capabilities of the mid/bass section.
All of the aforementioned coax drivers are mid/tweeters only because the edge termination/surround required doesn't provide the kind of travel necessary for decent bass. I would have thought though that if you were to use a 12 or 15" driver, that only having a limited amount of linear excursion wouldn't really be a huge limitation. Given the domestic environment I bet you don't really use more than a mm or two with your 15" designs.
Really!? I love the totally impractical recommendations that people give around here.
At any rate, your comments sound pretty hypothetical IMO. I am interested in practical solutions to achieving a superior result. The KEF coax was not really that good which is why it never went anywhere. The ones that I tried didn't work all that well and I don't see a small dome tweeter being a very good option either. I'll stick with the solutions that have been proven to work, but thanks for the thoughts.
It's possible to match directivity with a smaller woofer by using a passive resistive enclosure for cardioid like response.
Seems that KEF's and Pioneers latest offerings appear to be the best right now.
You must not have tried their latest generation 8" coax driver with a 1.5" dome: Apple Desktop Speakers Using Light Weight Cabinet Design - Page 2
Although i'm not a fan of plastic membranes in bass or midrange loudspeakers, i must say that they look great.
Are they available at some store ?
I've heard the recent KeF coax drivers in their systems and every one has a nasty bit of spit in the 4-5khz range that's unbearable to me. That might be their crossover work.....who knows. But the real point is these drivers aren't available for sale and given the nature of a moving waveguide, I wouldn't use them below 500hz in any design to keep cone motion minimal.
I'm optimistic with the most recent driver options and offerings that will allow for systems who's crossover can be placed outside of the critical range of hearing. While I'm not as critical of the passband as Earl, I'd certainly push the range as far as possible without having to trade off performance elsewhere. Small format dedicated midranges and fullrangers seem to be addressing the problem in a not so direct way but I'll take what I can get. The B&G Neo10 also looks like a viable candidate if properly implemented.
Troels has a write up,of the new SEAS premium Coax for those interested. He promises a revisit with a complete design soon. When I re evaluate my price comments earlier, the value isn't too bad considering a mid AND tweeter if the performance is there.
I'm optimistic with the most recent driver options and offerings that will allow for systems who's crossover can be placed outside of the critical range of hearing. While I'm not as critical of the passband as Earl, I'd certainly push the range as far as possible without having to trade off performance elsewhere. Small format dedicated midranges and fullrangers seem to be addressing the problem in a not so direct way but I'll take what I can get. The B&G Neo10 also looks like a viable candidate if properly implemented.
Troels has a write up,of the new SEAS premium Coax for those interested. He promises a revisit with a complete design soon. When I re evaluate my price comments earlier, the value isn't too bad considering a mid AND tweeter if the performance is there.
I have tried a 3 way electronic crossover and I can't believe what I heard. Changing the crossover points when listening to different drivers, is very interesting, no fixed crossover point. Depending on what I am listening to(jazz,blues,rock), I can change some crossover points and get a different listening perspective for each cut. Try it and see.
PeterC.
PeterC.
I remember emailing you some polars a couple of months ago. A resistance box most certainly works, although it requires quite a bit of tinkering for the best results. Have a look at these two sonograms. One is a closed box with an 8" woofer. The other is of a box of the same size and with the same woofer, yet it's a resistance box. Measured under identical conditions.
Have a look at Kimmo Saunisto's website too:
DIY Loudspeakers Kimmo Saunisto
Attachments
- Status
- Not open for further replies.