“Loading” of tweeters in large drivers Coaxes

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi

Is this is an issue for most or all coax drivers? I'm wondering how good is the "horn” loading shape of the tweeter in a driver like the P.Audio BM-18CX38, which otherwise seems a great driver (for OB or other)

panomaniac’s experience:
http://www.audiocircle.com/circles/index.php?topic=60599.msg541181#msg541181

While I havent seen it in specs for the driver, http://www.audiocircle.com/circles/index.php?action=search2 mentions it having a dual stage radial phase plug, and an acoustic lens. I could be misunderstanding those terms but in the photo http://www.p-audio.co.uk/products/db_product_1_6_bm-18cx38.htm (or the smaller BM15 CX38 at http://www.loudspeakersplus.com/html/paudiocoax.html), I can't identify an acoustic lens. Might it be sitting behind a dustcap?

Or is the waveguide for the tweeter is the woofer's cone? And if it is, that wouldn’t be an optimal shape, and would cause difficulty getting a really smooth crossover, especially at the lower end of the tweeter’s range.

Might P.Audio's CXHA range (shown at http://www.loudspeakersplus.com/html/paudiocoax.html which clearly have waveguides), *if its "well designed", have a smoother tweeter’s low end?

Hawthorne Audio’s Sterling Silver Iris 15" coaxial uses a Radian high frequency driver 'rated' for an XO of 1000 Hz, but needs to be crossed at about double that, as it's loaded by the woofer's cone.

Are there any 15" coaxials that are known to use a *well designed (if not that large) integral waveguide?
Or for it to be well designed, is a separate waveguide better?

Thanks
 
Hi Pano

If the cone (18”) is the waveguide, its a decent size waveguide, might be good down to about say 600 Hz.

A woofer cone is also relatively “shallow” for a waveguide, which I believe is often better than a high rate of expansion. Maybe that’s why you find the off-axis response is good.

OTOH, quite likely the cone profile is a compromise between what’s good for the woofer and what loads the horn. That would cause the tweeter’s response to be less even.
 
Administrator
Joined 2004
Paid Member
Yeah, 600Hz sounds about right. Maybe the baffle takes it down a little further.

I would assume that the woofer cone isn't the perfect waveguide - but it does seem to do pretty well. Much better than I would ever have thought.

Crossovers can be tricky no matter what. But I think this horn can be tamed. =) It already sounded much better than most, I just wanted to hear "perfection." The insane amount of fast flutter echo in the hotel room didn't help matters any. That flutter may account for a portion of the slight midrange trouble we were having. Only time and hard work will tell!
 
I've been very interested in the dual concentric type coaxial loudspeakers like these for quite some time now, but still haven't taken the plunge, for a number of reasons.. mainly just leery/skeptical about several things regarding the design of these drivers, and the impact these things probably have on the sound. Some of these issues have already been mentioned, such as the "snout" or tube referred to by Duke on the OB forum at AC, which panomaniac linked to above..

I'm a bit leery about the long "snout" that, in most cases, the compression driver's output must pass through to reach the woofer cone. Big Tannoys and several B&C drivers use the same magnet for both woofer and compression driver, which significantly reduces the snoutlength - but that format may be more prone to flux modulation. Which is the more audibly significant issue, I do not know."

For what must have been the last couple of years now, I've often pondered all the Tannoy style coaxial drivers from several pro loudspeaker manufacturers, including: these P.Audio's, Tannoy, BMS, B&C, Beyma, Radian, Hemp Acoustics, JBL, Eminence, Hawthorne Audio, SammiSound, Atlas Sound.. and probably tons more that I can't remember right now off the top of my head.

One thing I think pretty much all these types of drivers have in common is that the frequency response of the high frequency driver is very ragged. Think along the lines of a large fullrange driver with a whizzer.. sometimes even worse, and probably less predictable. From what little I've been able to observe during all this time, I think this is most likely due to a combination of things, all of which are detrimental to the resulting output from the high frequency driver.. Like the long "snout" mentioned earlier that any sound from the high freq driver must pass through. Then, there's a far from optimal cone shaped "thing" attached to the end of that "snout" that sits just behind the coaxial driver's dust cap (if it has one, and most of these pro coaxial units do, unfortunately). From there, there's a nasty transition from that little round metal cone shaped thing to the woofer's cone, and often over the woofer's voice coil former in between (yuck!). Then of course, you have a suboptimal shape for a large waveguide, which just so happens to be whatever woofer cone is used.. and then there's the bumpy surround, rubber/foam/cork gasket thing, screws etc, and then any small gaps between the woofer's frame and the baffle that it's mounted to. From what I can tell, I've always assumed that most (if not all) of these loudspeaker manufacturers haven't put much (if any) effort into custom designing woofer cones to be used as waveguides for these drivers. I might be wrong here, but it seems to me as though the vast majority just use whatever woofer cone for T/S parameters etc (or whatever), and whatever the high frequency response ends up being is basically what you get, without much effort put into changing it. I might be wrong about that though, but that's just what it looks like to me. Tannoy, on the other hand, is probably a different story, although I still question some of their pro models and in-ceiling products.

Some other things that might be worth mentioning.. Correct me if I'm wrong, but I always just assumed that any conical waveguide that was anywhere close to being constant directivity would have the typical falling frequency response, following the increased output from loading of the horn. Strangely, this isn't seen on spec sheets for most of the coaxial drivers from the manufacturers listed above, except for a couple of the smaller ones where it can be seen a bit.. like the 8" P.Audio BM series coaxial, as well as a nice looking 8" coax without a dustcap from BMS, if I remember correctly. As for pretty much all the others, I don't think this is the case. The response just keeps going everywhere, really ragged. With those smaller coaxials where the HF response droops down, it's ragged too.. but maybe not quite as bad. This is mostly all observed just from looking at manufacturer spec sheets BTW, as I haven't seen very many measurements from 3rd parties, maybe a couple, don't remember.

Also, I've seen a couple models from B&C and Beyma where they somehow integrated the HF driver into the same motor assembly as the woofer. Seems like this might eliminate some problems caused by traveling through that long "snout", but from the looks of the spec sheets, the response still looks pretty awful. I wonder what's going on here?

What's that funky looking metal thing that Tannoy uses in the "snout" of so many of their drivers? I came across something about it a while back, but I can't seem to find it now. I'm sure it was something to improve the HF response, and I'm guessing it must have worked, for the most part.

Anyway, it's really coincidental to me that this thread was posted.. not only because of the topic of coaxial drivers, but also because the P.Audio BM-18CX38 has been the main driver of discussion. For the last several days, I've been looking at that particular driver (again) very closely. If I remember correctly, I was modeling T/S params for it again real quick in WinISD Pro a few nights back, and realized I could probably get a pretty good amount of output from it in a sealed enclosure for use with subwoofers.. without having to use a highpass filter on it. Of course, it'd only go down to 80Hz 2nd order though. But at "normal" listening levels, some EQ could probably be used to get it a little lower if I wanted to cross to subs at say 60 or maybe even 50 Hz instead of 80. Without the highpass filter, group delay would be much lower, and I figure it'd just sound a lot better overall in that range.. as long as there's a good way for me to have full control over the phase alignment of the subwoofers. I dunno.. was just a thought.

I'm sure I can think of more to post, but I gotta get back to some other stuff for now. I might have some PDF's and/or images to add that might be of interest too, I'll have to look.

Anyhoo, if there's any way someone could get ahold of some good measurements of that 18" P.Audio coax (or any of them for that matter) and post them, it'd really make my day. :)
 
MisterTwister said:
how about converting this beast into coaxial
http://www.pyleaudio.com/itemdetail.asp?model=PDW21250
Remove dustcap and attach compresion driver with horn to pole piece.

..haha, I've often thought about that too. I'm just not sure what I'd use for the horn tho. I can hardly make enclosures, much less fabricate anything like that, so that's probably out of the question for me. Any ideas? What are you planning to use for the horn? Whatever it is, a 2" exit HF driver would have to be able to go pretty low, probably crossed at 600-650 Hz at the highest, and I like at least an octave of flat response below that. That might be asking too much though, I dunno.

And yes, I agree.. the big P.Audio coaxes are expensive. It's tough to take such a risky plunge into the unknown, especially after you've done it far too many times in the past.

edit: For some reason I was thinking that 21" Pyle had a 6" voice coil, but it's only 4". In that case, I wonder how you'd find a small enough driver that'd go low enough to cross to it. Probably impossible without mounting it some other way, too far from the 21" driver, and then you'd have a difficult time trying to use the 21" cone as a waveguide too. Hrmmmm..
 
BHTX said:
One thing I think pretty much all these types of drivers have in common is that the frequency response of the high frequency driver is very ragged. Think along the lines of a large fullrange driver with a whizzer.. sometimes even worse, and probably less predictable.


While I'll grant you that the top octave isn't generally as clean as one finds from better domes flush-mounted, in the better designed ones it's not that bad. Here's Stereophile's 30deg averaged response of the Tannoy TD12, for instance.
tantd12FIG4.jpg


IMO, the benefits in midband power response are worth the less-than-stellar top octave.

BHTX said:
What's that funky looking metal thing that Tannoy uses in the "snout" of so many of their drivers? I came across something about it a while back, but I can't seem to find it now. I'm sure it was something to improve the HF response, and I'm guessing it must have worked, for the most part.


There's some propaganda about it in this paper.

Here's a color cutaway:
dmtcross.jpg


I can't say anything about other coaxes, because except for a brief interlude with the 8" B&C I really only have experience with the Tannoys.
 
Hello All....

Interesting conversation on the P. Audio BM18CX38.


OK... this will likely be a long thread... much to talk about. Alot of what I will say is pretty well known... so I hope I do not bore too many with this. Plz forgive spelling-syntax errors!!

After fooling with... or should I say experimenting with many coaxials over the past 40 years... mainly the large 12-18" pro versions... including EV, Altec, Selenium, Tannoy (Very long ago) and now P.Audio, I would like to share the following.

One thing I have learned is that any coax that has a decent design can be tamed... it comes down to mostly crossover and of coarse tuning that crossover to suit the users intended application.

Most manufactures that supply or advise a recommended crossover assume the main use of these drivers is high output pro... commonly for PA and or monitor work. So the main crossover design compromise is balancing power handling (durability) with sonic signature. This usually means taking the LF section up as high as possible. Of coarse this is not desirable in a small room application... at low to modest play back volumes in small room near field listening conditions....

And ... in the past... nearly all manufacture supplied crossovers did not fully correct for the horizontal offset of the voice coils... so the outcome was not phase correct through the cossover region... The last of the Altec crossovers got close to getting it... Jeff Markwart does a great job of explaining this and his phase correct crossovers work VERY well.

My interest in the BM18X38 is in OB use... so my crossover solution is geared toward the special considerations OB requires. I do not want to get into all the design aspects of this kind of crossover at this time. But one of the main considerations is the performance of the HF section.... what it naturally does and does not do well. And key to that is the loading effects of the woofer cone and baffle combination. The BM18's I have worked with have HF FS in the low 500 hz area. And from crude measurements they really get up to over 18Khz on axis before a quick death. And with out any crossover they can do well up to about 12 Khz or so at 30 degrees off axis. This is better than a lot of dedicated dome and cone midranges mounted on a typical baffle. So... it was obvious to me that with a proper crossover combined with the wave guide effects of the large woofer cone and the output ability of the HF driver were conducive to... and could deliver excellent power response in a near field envionment.

Add the nearly 16" diameter of the cone and some baffle area ... say 800-1000 sq. inches... in a home hifi situation the HF will work from a practicle sense down to under 600 HZ. And is fully loaded from about 780 HZ up. This makes getting the crossover right much easier. Maybe more on this in another post.

As for the HF section... it... by modern considerations is pretty straight forward. Most if not all large format compression drivers use a radial (or similar) phase plug... with notched sides and exits. It appears the BM18 also has some sort of (fiberglass??) mesh at or near the exit of the HF that acts.... along with the dust cap as an acoustic lense of sorts. Once you establish a phase correct crossover.... this seems to happent around 950 Hz acoustic...Give or take... all of this come together to make for a speaker that can image with the best of them... regardless of design or type.

As for the drooping repsonse of the HF up above 10K... this is typical for most compression HF drivers and I think is more indicative of a mid broad mid band peak than a high end roll off. This also can be cured with crossover tricks... JM uses a voltage divider... I try to solve it with a staggard third order BW. Or a combo of the two. The important thing is to preserve the phase correctness of the crossover while taming the mid band HF rise. Maybe more on that another time to. This is already WAY too long.

Best...

JBSpeakerman.:smash:
 
Administrator
Joined 2004
Paid Member
Thanks JB!

I was just about to post when your post popped up.
You pretty much explained it.

If you were to take into account all the truly bad things that BHTX posts, you'd think these crazy things could never sound good - but they do! Must be something strange going on....
 
John Bsuch said:
As for the drooping repsonse of the HF up above 10K... this is typical for most compression HF drivers and I think is more indicative of a mid broad mid band peak than a high end roll off. This also can be cured with crossover tricks... JM uses a voltage divider... I try to solve it with a staggard third order BW. Or a combo of the two. The important thing is to preserve the phase correctness of the crossover while taming the mid band HF rise. Maybe more on that another time to. This is already WAY too long.

Best...

JBSpeakerman.:smash:


Actually, falling response (with frequency) is a fundamental characteristic of ANY constant-dispersion (i.e, the dispersion angle is relatively constant WRT frequency) horn. In fact, any horn that does NOT exhibit this characteristic can be assumed to NOT be constant dispersion.

And yes... it's most effectively dealt with in the crossover. Usually, some combination of low-Q highpass filters, crossover point staggering (making the highpass electrical frequency a bit higher in frequency than textbook, to "gap" the response at crossover), and using shelving filters to increase the top-octave relative output. Combinations of these can produce remarkably good results... in fact, for one example, I've gotten a Radian 5208-based design to be within +- 2.5 dB over the range of the horn (1500 Hz to 21KHZ, in fact). There are other ways of doing this (Wayne at Pi Speakers has an interesting tack... which on the face of it seems almost OPPOSITE of what I mentioned above... but it WORKS, because I'be HEARD it)... but this is probably the most common and effective method, at least IME.

Regards,
Gordon.
 
Gordon..

You make some very good points.... in my post I was approaching the the HF driver response shaping issue from a cross over mind set.. thinking of the mid range excess relative the very high end deficency of out put... because that is what I need to match to get a realizable... usably flat acoustic out put.

No doubt about it.... All compression drivers will exhibit a fall off on the top end... it is the shape and rate that count and have to be dealt with.

And using staggard third order and or voltage dividers is what I have found works for me... but by all means... there are a lot of other good... sucessful ways around the barn. What has amazed me over the years is how many designers just do not get this solved.... Wayne and others of coarse have figured it out.

JB
 
With regards to the Tannoy paper linked to by Pallas, above, I am puzzled by page 8 which purports to show the advantages of co-axial dispersion by showing dispersion patterns at various frequencies of coax and non-coax drivers - a total of six graphs - all of which are identical unless my eyes have finally failed me!
 
Russell Dawkins said:
With regards to the Tannoy paper linked to by Pallas, above, I am puzzled by page 8 which purports to show the advantages of co-axial dispersion by showing dispersion patterns at various frequencies of coax and non-coax drivers - a total of six graphs - all of which are identical unless my eyes have finally failed me!

Nope, they haven't failed you. I did the same thing, stared at it for probably 3 minutes scratching my head saying "What the hell?" several times before finally giving up and going on reading.

Thanks Pallas for that paper, it is the one I was referring to in my previous post.
 
Pallas said:
While I'll grant you that the top octave isn't generally as clean as one finds from better domes flush-mounted, in the better designed ones it's not that bad. Here's Stereophile's 30deg averaged response of the Tannoy TD12, for instance.

BHTX said:
Tannoy, on the other hand, is probably a different story, although I still question some of their pro models and in-ceiling products.

As far as Tannoy goes, I agree. Most of the graphs I've seen use a bit too much smoothing though, and it's difficult to tell what's really going on.. and they usually have the finished crossover on them too. I want to see some REAL measurements of the individual drivers, and their off-axis response. Anyway, I really like what little bit I've heard from Tannoy (some DMT 15's some years back.. and I think that's about it as far as the DC's go). I only wish more could follow in their footsteps, rather than ignoring their design principles and accomplishments from years ago.

John Bsuch said:
This usually means taking the LF section up as high as possible.
John Bsuch said:
And ... in the past... nearly all manufacture supplied crossovers did not fully correct for the horizontal offset of the voice coils... so the outcome was not phase correct through the cossover region...
John Bsuch said:
could deliver excellent power response in a near field envionment.

..I strongly agree with the first two quotes above, and I always find this disappointing as well, altho I don't know why, as I'm not sure I've ever come across a great xo from the manufacturer. I guess it just couldn't be that simple, that'd be too easy for us all, lol. As for the last quote, I agree.. and we've obviously shared very similar thoughts regarding this. Excellent power response, imaging, and tons of headroom.. not to mention all the wow factor as a side benefit. :D

John Bsuch said:
As for the drooping repsonse of the HF up above 10K...

I wasn't referring to the roll off above 10k though..

BHTX said:
but I always just assumed that any conical waveguide that was anywhere close to being constant directivity would have the typical falling frequency response following the increased output from loading of the horn. Strangely, this isn't seen on spec sheets for most of the coaxial drivers from the manufacturers listed above, except for a couple of the smaller ones where it can be seen a bit.. like the 8" P.Audio BM series coaxial, as well as a nice looking 8" coax without a dustcap from BMS, if I remember correctly. As for pretty much all the others, I don't think this is the case.

GordonW said:
Actually, falling response (with frequency) is a fundamental characteristic of ANY constant-dispersion (i.e, the dispersion angle is relatively constant WRT frequency) horn. In fact, any horn that does NOT exhibit this characteristic can be assumed to NOT be constant dispersion.

..Exactly what I was trying to say. I don't understand why so many of the raw responses I've seen of these dual concentric style drivers don't represent a constant directivity device, and yet a few of them clearly "try" to do it.. altho usually with a lot of raggedness. It just doesn't make much sense to me.

panomaniac said:
Thanks JB!

I was just about to post when your post popped up.
You pretty much explained it.

If you were to take into account all the truly bad things that BHTX posts, you'd think these crazy things could never sound good - but they do! Must be something strange going on....

..Jeez. You almost make it sound like I'm being ganged up on for a smack-down or something. How lame. ;)

Was just sharing my own thoughts about them is all, because I have a lot of questions after staring at each and every coaxial I can find, day in and day out for a couple years or more now. Obviously, I know they CAN sound good, or else I wouldn't be so obsessed with them, and I certainly wouldn't bother to post in this thread.

John Bsuch said:
in my post I was approaching the the HF driver response shaping issue from a cross over mind set.. thinking of the mid range excess relative the very high end deficency of out put... because that is what I need to match to get a realizable... usably flat acoustic out put.

I'm curious as to whether or not you've measured and experimented with the off-axis response of these drivers at all? If so, what did you find? Seems to me as though a perfectly flat on-axis response from a constant directivity device might seem a little bright? Or maybe I'm wrong. Or, maybe these aren't constant directivity, as previously mentioned.. or else freq response from the HF driver would fall as frequency rises, as previously mentioned. But if that's the case, then why do they claim to be constant directivity?
 
I read the Tannoy "white paper", and while it's an extended repetitive ad, and imo in the odd case contradictory, avoiding either a "snout" (tube), or a flare projecting into the woofer cone definitely appears to me to be a better route.

While the specifics of "dual concentric" are presumably patented or the term copyrighted, are any other large (say 15 inch) coaxes single point source? (The P.Audio BM-18CX38 or 15 inch P.Audio BM-15CX38??)

BHTX,
you mentioned some models from B&C and Beyma with the HF driver integrated the into the same motor assembly as the woofer. But the response still looks pretty awful . . sounds like single point source - poor execution?

Amongst other things, single point source would avoid any need to correct phase in the crossover.

What sort of money do a pair of large Tannoy dual concentric drivers go for (preparing for sticker shock)? P.Audio’s BM-15CX38 are each USD 450 . .

If large single point coaxes aren't affordable, JB thanks for the name Jeff Markwart re correcting phase in the crossover. I'll read this later http://home.earthlink.net/~jmarkwart/

Pallas,
Agreed the midband is much more important than the top octave.
The benefits in midband power response you referred to – do you mean single point source Tannoy vs the rest, off axis?

Which make me think, has anyone quantified or verified the advantages of a single point source?

JB,
Thanks for joining in. Yes, much to talk about . .

The Tannoy paper claims many things, which probably warrant a separate post. One assertion is that "the Tannoy Dual Concentric covers broader areas of the listening environment", with graphs (p 14) showing typical coaxials having about 30 degs coverage, vs Tannoys having about 50 degs.
I'd think that while 50 degs would be good for a wide sweet spot; but depending on the room's size, a "typical coaxial's" vertical dispersion of 30 degs would have the advantage of less floor and ceiling interaction.

More importantly, the Tannoy paper also says
- A true Point Source gives the same sound from seat to seat.
- Dual Concentrics can use simpler, better sounding and more efficient crossover
- "Harmonic relationships" are better preserved using a single point source (I think is the same as saying phase response is better)

with your wide experience, do you agree?

Btw (apart from less VD = less extension) the smaller 15 inch BM-15CX38 is presumably equally as good as the 18 incher?

My interest in coaxes is also in OB use. I'm planning an open baffle about 1400 sq. inches. Could you elaborate on how baffle area helps the HF play lower?

Do you know what the acoustic lens in these P Audios does?

Sorry for all the questions

Cheers
 
BHTX said:
I'm a bit leery about the long "snout" that, in most cases, the compression driver's output must pass through to reach the woofer cone. Big Tannoys and several B&C drivers use the same magnet for both woofer and compression driver, which significantly reduces the snoutlength - but that format may be more prone to flux modulation. Which is the more audibly significant issue, I do not know."
the tulip waveguide was the favourite for a while.
Tannoy have reverted to the pepper pot treble horn for their top of the range speakers. This combined with going back to Alnico for the Westminster HE, has pushed the price even higher.

The 385hpd often gets to around £200 to £400 a pair in Ebay for a variety of condition.
Recently a refurbished pair hit £700.
These hpd drivers are Alnico & pepper pot. The last of this type before the big fire forced a complete redesign to Ferrite & Tulip. However they do not last forever due to the use of foam for the surround and cannot be repaired cheaply due to the non standard size & shape of the required surround. Spares are available at about 3times the price of equivalent 15inch surrounds. A complete cone and surround and VC is available direct from Tannoy.
 
I googled . . single point source speaker Coax driver

A couple of hits:

http://www.tnt-audio.com/casse/ha_si_part1_e.html suggests that
"if a tweeter is necessary(!), we can do one thing to preserve one feature of a single driver speaker, mount it in the centre of the larger driver, thus preserving the single point-source advantage", which on face value ignores the issue for a single point source of depth.
But then it adds
"Having discovered that there wasn't such a driver, (the designer) Darrel Hawthorne approached Eminence and they agreed to make one for him."

Anyone know if Hawthorne Audio's Silver Iris drivers are true single point sources?

Common Sense Audio http://www.paudiospeakers.com/paudio.html implies that the PAudio BM..CX38 series are single point sources.

I emailed both Common Sense and Hawthorne to clarify if they have with zero difference in the source alignment front - back.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.