|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
180.200.145.221
The design allows for forward and back movement of the WR driver's spherical enclosure. This to allow us to time-align its acoustic centre with the two matching bass enclosures above and below it, using RTA.
I will be listening in the near field as for years now.
Crossover frequency will be 170 Hz (a wavelength / lambda of 2 metres). Either 1st order with passive-line-level input filters to the matching power amps, or 3rd order symmetrical with the addition of a 2nd order active crossover. It has a 170Hz option and in-phase HP/LP outputs.
We can also try the two asymmetrical slope options. HP1&LP3 or HP3&LP1.
The first iteration - using a conventional dynamic 5" plain paper cone allows me to seal the WR driver's sphere when it will roll off around 160 hz 2nd order, giving an effective 3rd order total slope.
But, there's to be a second iteration using Manger drivers in the small spheres. They may need the steeper 3rd order electric (and acoustic) slope running from 170 Hz up.
The spacing between the top and bottom woofers will be less than 1 metre IE less than half the lambda of the crossover frequency. The central driver's centre will be equidistant from the top and bottom woofer's centres.
I'll be listening in the near field, and the amps are matching (same voltage gain) stereo tube amps of 20 watts (P) and 16 watts (UL) respectively.
I have the option of using 1/3rd octave analogue eq. in each bass channel.
Can anyone technical see any big issues with such an array?
Eg.
Am I likely to have ceiling reflection issues?
Room sound FR very different to the direct sound?
Warmest
Tim Bailey
Skeptical Measurer & Audio Scrounger
Follow Ups:
"Room sound FR very different to the direct sound?"
That is likely to happen with a fullrange driver, as most fullrange drivers beam significantly as they go up in frequency. Good for you, for even wondering about it!
The off-axis response will dominate the reverberant field (which is what I presume you mean by "room sound"). As long as the radiation pattern changes significantly with frequency, you cannot eliminate the spectral discrepancy between the first-arrival and reverberant energy via signal processing.
In my opinion the spectral balance of the off-axis energy matters for two reasons: First, the ear/brain system derives timbre from a weighted average of all the energy it hears. The first arrival sound is only part of the story; the spectral balance of the reverberant sound will add to (and either enhance or degrade) the timbre of the first-arrival sound. You aren't conscious of this because it all happens very quickly, but think of singing in the shower as an example. And there is a great deal of peer-reviewed experimental evidence supporting the idea that spectrally correct reflections are desirable (see Floyd Toole's "Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms").
The second issue is listening fatigue. Briefly the ear/brain system classifies incoming sounds as either initial sounds or reflections by comparing their spectra with initial sounds that have recently (within 30 milliseconds or so) been stored in a short-term aural memory. When the off-axis energy is significantly duller than the on-axis energy (because of beaming), there's a major spectral discrepancy between reflections and the initial sound, such that the ear/brain system literally has to work harder to correctly classify these sounds as reflections. I believe that, over time, this extra CPU useage can result in listening fatigue. To the best of my knowledge that conclusion hasn't been tested and proven or disproven, so you can take this paragraph with a grain of salt, but the effects of the reflected energy's spectral balance on timbre and subjective preference (from the previous paragraph) are well documented.
Okay here's my suggested solution: Use a rear-firing tweeter equalized to offset the effects of the fullrange driver's beaming.
On another topic, imo you've made a very good choice by not asking your fullrange driver to handle large bass excursions. That keeps its voice coil in the linear region of the magnetic gap, so clarity will hold up better with loud, complex music.
Best of luck with your project.
Duke
Me being a dealer makes you leery?? It gets worse... I'm a manufacturer too.
A 4" cone begins to beam at 3.5Khz. So, what frequency should we roll the rear tweeter in at? 3.5 kHz or ~5 kHz? At 5k the rear tweeter would be down ~ 6db at 3.5 kHz as against the WR driver. Is that about right?Or is the Eq usually more complex than that? If yes, how to measure for it?
Do we need a level control for each rear tweeter? Constant-Z thingies?
Series or parallel connection with the main driver?
I have a pair of unused/NOS 1980's Foster/x 1" domes, 8ohms, 91db/w ~ the same as the Corals. They'd likely be fine at ~5khz with a 1st order slope, only just at 3.5K.
What about the Mangers to be used in the final build? They are flat bending wave drivers, 21cm across / frame diam? Will they need beaming compensation?
Thanks for your input.
Warmest
Tim Bailey
Skeptical Measurer & Audio Scrounger
Edits: 10/01/14 10/01/14
You'd have to run measurements on your fullrange drivers if you wanted to know what they're actually doing, as the real-world behavior of drivers often departs significantly from theory. In particular, I don't know what the off-axis behavior of the bending-wave Mangers would be like.From here I couldn't predict with any degree of confidence what the target response of your rear-firing tweeter should be, except that it would have a generally rising trend.
Fortunately first order highpass filters are easy to experiment with, you just change the value of a series capacitor. You can probably come up with a worthwhile improvement in an afternoon or so of trial-and-error. Like you might start with 1 uF, and then compare the results with 2 uF, pick the winner, and just keep fine-tuning. Once you've made a clear improvement in one speaker, maybe use that as a reference, and try to beat it by tweaking the other speaker. You don't need to make stereo comparisons at this stage - in fact, that can distract you from hearing small improvements in timbre.
If your fullrange driver has high efficiency you may find that you need a more efficient tweeter to get the best results, but for now, try plowing with the horses you've got and see if the results are encouraging.
Very best of luck with your project!
Duke
Me being a dealer makes you leery?? It gets worse... I'm a manufacturer too.
Edits: 10/02/14
Timbo: If you look at measured response of real drivers, you'll see that most of them have a fairly narrow frequency range where the directivity is more-or-less uniform at the typical measured angles of 30 and 60 degrees off-axis. For the example 4" which you cited, I would expect its off-axis response (say, 30 degrees) to begin dropping by about 1500 Hz, and at 60 degrees off-axis, it could easily be down 10-15 dB by 5000 Hz. This is the kind of thing which Duke is talking about. Obviously, different driver models will behave differently, but this is the sort of thing you can expect. Given your design/cost constraints, what I would do is take a look at the 45 or 60 degree off-axis FR curve, and run the driver up to the frequency where it's not down more than about 6-8 dB or so, if you want decent off-axis sound quality and a nice ambient (room) sound. Be aware that you may find that your driver isn't as wide-range as expected. Duke may disagree with my ballparks, but he does this stuff all the time, and likely can put a finer point on it. But the basic concept is to achieve wide and uniform dispersion over a broad range of the audio spectrum. THEN, you have half a chance at getting great ambient sound AND off-axis direct sound. (The other half a chance is to have a decent room acoustical environment, but that's another topic.)
Duke: Thanks for the mention of Toole's work - I'll have to check that out! Since you're into the whole psychoacoustic thingy, you'll go positively orgasmic with Dr. Eric Heller's "Why You Hear What You Hear".
With regard to your initial post in this subthread - ba-da bing!
:)
Seriously, you need to open a book or two.
The people who designed and built the original speakers believe this will work. Maybe not transient accurate, but worth doing anyway. Especially the last stage. Also see below.One of them still builds spherical speakers. The other who is retired will be helping me with the design, including measuring.
Their advice got us / me this far.
I have Bullock on Boxes and can borrow through/from the Australian National Library, ANU and UC. I read Olson ? (RCA) years ago, through these two guys.
It will not be a fast project, so I have plenty of time.
Warmest
Tim Bailey
Skeptical Measurer & Audio Scrounger
Edits: 09/23/14
I have a DIY/experimenters WMTMW format arrays, and I listen quasi-near field. BUT my woof to woof spacing is a mere 16" and midrange spacing is 6" center to center. (I use truncated 4" mids and a small-flange neo-tweeter that's about 1.6" high, and mid-bass drivers are 7". These driver sizes are all quite small as far as WMTMW designs go, with only the odd mid-dome variations having tighter spacing.
Large-distance arrays are often said to work better for far-field because drivers don't *blend* as *quickly* (not less well, just further away).
But the biggest question of all is: why go through the trouble of time-aligning when you're not going for a transient accurate result anyways? You can only dial in the "correct" distance (delay) for a given frequency.
Big difference between delay and group delay.
Cheers,
Presto
And again, when I put the Manger-spheres in the middle.
Lots of experimenting to do.
Warmest
Tim Bailey
Skeptical Measurer & Audio Scrounger
Sorry, you lost me. The drivers you described can't do 1st order acoustic slopes at 160/170Hz due to their natural rolloffs. Do you mean 1st order electric crossover combined with the natural rolloff?
I find with integrating bass drivers and midbass / fullrange drivers, it's not always as difficult as it sounds like on paper. Then again, I don't shoot for a *naturallly* transient accurate response. When doing digital I can use forward/reverse phase correction (group delay correction) or FIR filters. With analog I usually overlap electrical filters with acoustic rolloffs and shoot for 3rd or 4th order.
To be honest, I used to be very enthusiastic about transient accurate systems, but now it's not so much of a deal breaker for me.
Some people have problems with "absolute polarity" and other such delicacies - and switch polarity to and fro between recordings to figure out which is right. Others just say non-TA speaker systems just don't "sound right" and that only dynamic TA designs or planar speakers will do. Often the bass modules added to planar speakers are not TA. Many say TA is only important in the midband.
I really don't think you'll have too much trouble.
Also - why voltage match the amps? Are the drivers gain matched? I usually null the gain difference between drivers *using* the amps rather than add passive components for padding down drivers. Anyways, I was just curious about that. Gain matched amps didn't seem to be a requirement here so long as the final gain of each "cabinet" is correct.
Cheers,
Presto
Also see my response to Inmate51.My budget is likely tiny compared to yours, it will only dribble in. I am very happy that I will have access to an RTA, laptop, measurement mike combination, and an experienced spkr builder!!!
Phase 1 is one pair of the WR driver spheres and 1 pair of the bass-only spheres close to the floor. Ph. 2 is adding a second pair of bass spheres in a vertical array. Phase 3 is to put Mangers in reduced VB spheres into the array.
Clearly, I'll be doing a lot of learning!
If I find that a transient accurate (TA) crossover is not possible with these items at 170 Hz, I think I'll be able to live with that. I do hope to be able to stick with a 1st order HP for the Mangers, in the hope that they will be TA from the midrange on up, at least, or seem to be!
Can the Manger driver do it? I've corresponded with JC Morrison - he used to post here a lot - about using the Mangers in an 8L enclosure, HP'd 1st order at 150 Hz, AND with a small valve amp of no more than 20 watts. He believes it will work quite well.
There may be a BDS issue but we can compensate for that too, at line level, with two pots a mike and an RTA and ears. It is entirely possible that the final HP acoustic slope will NOT be a pure 1st or 3rd. A good blend and clean behaviour when loud is the target.
The existing WR driver is a Coral Flat5/III. In its small (15L Vb) sphere it is flat to an octave below 170Hz from its BDS up near 300hz. The port tuning is 60 Hz IIRC. So there would be a 1st order HP slope at 170 Hz, kneeing down at 4th order from 60 Hz. If we sealed the encl. we should get a summed slope approximating to 3rd order, when summed with the 1st order passive-line-level PLL input filter on its amp's inputs.
If we find that there is a problem with the smaller sphere's kneed slope blending with the two bigger sphere's slopes, we can try sealing the small spheres.
Using a spare 2nd order active crossover - which is being blue-printed to get a matching 3rd order LP for the bass.
It is possible to use asymmetric 1st/3rd order slopes in such arrays. So we may try them as we move along.
The active crossover can also be used to HP filter the Mangers more steeply. But I hope I don't have to.
The mid-bass driver in the larger spheres (were 2-ways) is a Foster FW 202. These larger spheres are flat from around 40 Hz to 3.3Khz and then roll-off cleanly.
The enclosure's sensitivities do match at the crossover point. At 170 Hz I wouldn't expect there to be a LOT of mismatch for dispersion and room-sound. I also have two, same model, matching voltage gain power amps. rebuilt/modernised LEAK St20s. So, I decided to try matching line-level HP and LP, 1st order input filters.
Reviews of both the 2-way Audiospheres and the Leaks are here at AA.
LBNL - with the 2-way spheres, since the early 1980s I've been consistently detecting polarity, and switching for it. Yes, most of what I listen to is acoustic music.
For me the central take-home concept about music, how we hear it and are affected by it it is the primacy - for us - of starting transients aka attacks. Way more so than the continuous tone - for those instruments that actually have a continuous tone. This fits rather neatly - IMO - with polarity mattering.
I do mostly listen to acoustic music, and prefer simple stereo miking.
Warmest
Tim Bailey
Skeptical Measurer & Audio Scrounger
Edits: 09/23/14 09/23/14
As I said, you need to open a book or two.
Start with Olson's "Music, Physics and Engineering".
You sound like you think you're inventing the wheel. This stuff has been done, decades ago!!
Thank you for the suggestion. There is a response to your first post, predating this second one of yours which included that I have read Olson. A long time ago though, and some of it may not have stuck.:-)!
If I can find or borrow a copy I will read it again.
Warmest
Tim Bailey
Skeptical Measurer & Audio Scrounger
Edits: 09/23/14
Yeah, I saw that, but you have a question mark, so I figured you weren't sure if it actually was Olson or something else. In any case, the book I referenced would be a good place to start, and I believe it's still available. Another excellent book is Vance Dickason's "The Loudspeaker Design Cookbook".
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: