PDA

View Full Version : Crossoverpoint for horn drivers



Niklas Nord
06-15-2005, 01:07 AM
If we take for example one JBL 2451 and cross this one at some 1600hz
instead of 600hz. Anybody knows how much power it will hande then?

That is, a higher crossowerpoint..

yggdrasil
06-15-2005, 03:19 AM
Power rating:

- 100w continuous program above 500Hz
- 150w continuous program above 1kHz

http://www.jblpro.com/pages/components/cmp_drvs.htm#2451HJ

Niklas Nord
06-15-2005, 04:20 AM
maybe somewhere 200watts then

spkrman57
06-15-2005, 08:55 AM
My 2 cents worth here:

1" driver @ 1.6khz

1.4" and 2" @ 800hz

These points allow plenty of safeguarding of the drivers and I have had success with those points.

Ron

whgeiger
06-15-2005, 09:07 AM
maybe somewhere 200watts then

NN,

Due to the large crest factor found in musical and other typical program signals, displacement limit set by a RMS power specification is optimistically unfortunate. Also, at 200 watts, a 'pressing' of the driver's thermal dissipation envelope will occur; not to mention, that an unacceptable level of throat induced distortion will be observed.

If more acoustic output is needed, then use a horn/driver array for reliable results. In the P.A. setting, reliability is everything except that intelligibility and inaudible distortion are expected givens.

Regards,

WHG

grumpy
06-15-2005, 09:57 AM
200W ? ... you'd be pushing well over 130dB SPL at meter distance. Instantaneously, the voice coil might not become a blown fuse (I recall old Dynaudio tweeter ads showing very brief bursts of 1000W+) , but I hope you're wearing hearing protection. :blink:

Niklas Nord
06-15-2005, 02:12 PM
Yes but music can well have peaks up to 120db, when listening to
classical music, or some percussion..
Not direcly dangerous to the ear if it´s just peaks and the overall
pressure is some 95db.

Do you mean that 200watt´s would blow the driver?

The thing with throat distorsion is interesting.
I have been asking alot about distorsion in horn drivers, but no one
has been able to give any detailed information about distorsion, and
how it compares to conventional tweeters..

yggdrasil
06-15-2005, 03:42 PM
Where are you going to use this driver?

whgeiger
06-15-2005, 04:33 PM
NN,

Some comments follow your text.

Regards,

WHG


Yes but music can well have peaks up to 120db, when listening to classical music, or some percussion.. Not directly dangerous to the ear if it’s just peaks and the overall pressure is some 95db.
A high crest factor, but underneath a powerful low frequency component hopefully seen by a robust woofer.

Do you mean that 200watts would blow the driver?
a) At lower frequencies, diaphragm fractured when slammed into phase plug.
b) At higher frequencies, voice coil meltdown marks termination of thermal run-away.

The thing with throat distortion is interesting. I have been asking allot about distortion in horn drivers, but no one has been able to give any detailed information about distortion, and how it compares to conventional tweeters.
Distortion occurring in a horn throat is attributable to the elastic non-linearity of air that is emphasized by
a) its entrapment in the small-enclosed spaces of phase plug passages and
b) the relativity high intensity of acoustic energy that passing through them.

The following formula may be used to characterize 2nd. harmonic distortion.


[D2h] = [k]*[f]/[fc]*(([Wa]/([St]*[N]))^(1/2)) – Second Harmonic Distortion


where


[k] – Empirically Derived Constant
= 0.4 - for conventional exponential horns that exhibit some beaming at high frequencies and thus a relatively flat on axis response.
= 0.2 – for constant directivity horns that exhibit a declining (un-equalized) on axis response.

[fc] – Flare Cut-Off Frequency of the Phase Plug Passages (Hz)

[f] > [fc] - Signal Frequency of Interest (Hz)

[St] – Area of Phase Plug Slits (Adjacent to Diaphragm) (in^2)

[N] – Quantity of Drivers Used

[Wa] – Total Acoustic Output (Ave. Watts)

For direct radiating tweeters, that are less efficient, principal distortion products are related to the non-linearities of diaphragm excursion that are emphasized by the prerequisite increase in volume displacement.

edgewound
06-15-2005, 07:19 PM
Here's my take on it and it kinda needs to be qualified....here goes:applaud:


I'll assume you're using it for hifi or studio monitor use...

if it's a two way system, a good frequency would be in the 1100-1200Hz neighborhood. Why? Because of a thing called "power response". How high you cross it over will determine how great the off axis response is in the upper mid frequencies that come out of the horn and don't beam too much like a flashlight.
You might want to consider an 18db or 24dB/oct hi-pass to keep the lower frequencies from upsetting the diaphragm too much. Now with that said...the driver will be able to try and reach 18-20KHz with a higher xover point since it would be relieved of lower freq duties.

As far as power handling goes...the more power the better...within reason, of course. You want plenty of headroom available to prevent the amplifier from clipping, and as we all know (we should anyway) amp clipping is public enemy number one with with HF drivers...doesn't matter what flavor they are...clipping burns the voice coil...and also sounds really bad:barf: .

You want your system to sound clean...so you need pleanty of clean power.

Correct me if I'm wrong, please:)

Thanks for reading my thesis,

Edgewound

Niklas Nord
06-16-2005, 12:01 AM
My questions come from some ideas I have, to build some speakers..
My Auroratowers will be finished this summer, så I must have someting
else to work with, I want a project :D

Loudspeaker built for hifipurpose, the goal would be as low distorsion as
possible. Then I would like to use a horn driver, and maybe a tractrix horn
couse theose seems to have better caresteristics than conventional horn
as in my K2 s9500.
A two way design maybe is hard to go with, three way if it´s better..

Is there anybody that knows any facts about conventional tweeters, do they
have lower distorsion than horn drivers? If let´s say the horn driver is crossed
somewhere ower 1khz, and the conventional tweeter is crossed ower 2khz.

Mr. Widget
06-16-2005, 12:17 AM
Is there anybody that knows any facts about conventional tweeters, do they
have lower distorsion than horn drivers? If let´s say the horn driver is crossed
somewhere ower 1khz, and the conventional tweeter is crossed ower 2khz.

You mention crossover frequencies but not SPLs... based on your earlier posts I believe you also want high SPLs. A conventional dome tweeter with a sensitivity of 91dB/watt will have quite a bit of distortion at 125dB. A TAD 4003 with a sensitivity of 112dB/watt will have the lowest distortion, a JBL 2451 will also have very low distortion.

WIdget

edgewound
06-16-2005, 12:32 AM
You've got several variables to deal with as far as what you're looking to accomplish, and what kind of distortion your willing to put up with. It's all a compromise somewhere. The 2451 is going to give you greater efficiency than ANY dome radiating tweeter, such as a Morel or Dynaudio, (if you can find any)or Ceratec, et al... The least distortion inducing horn type available, I believe, would be a conical style waveguide. Since sound waves are spherical in nature, the circular shape of the waveguide will be less prone to reshaping...or distorting, if you will...the physical shape of the sound waves radiating from the horn. The ideal approach, some would advocate,would be a coaxial arrangement, or Duplex as Lansing called it, so all frequencies are emanating from the same acoustic plane or "window", thereby reducing comb filtering or lobing effects of overlapping radiation patterns of separated drivers. Of course the drivers should be time/phase aligned in the crossover network to reduce the effects of phase distortion. Wow...lots of distortion issues to address:blink: .


Anyway, I hope these brief thoughts are helpful to you.

Anyone have any other thoughts here?

Thanks, Nik

Edgewound

Mr. Widget
06-16-2005, 12:40 AM
...lots of distortion issues to address:blink: .

Very good point. I was just thinking of simple harmonic distortion, but there are many forms of distortion. NN, you do pose some interesting questions!

To really get into it, I suppose we need to know more about your requirements, system, room etc.

Widget

Niklas Nord
06-16-2005, 02:16 AM
Well, If we think about the K2 s9500, one 375nd and two 1400nd.
I don´t think we will need any ultra high frequency drivers couse we
can´t hear well beneath 16khz, and I´m very happy with the high
frequency in the s9500.

I was whinking of a loudspeaker that would be working from 80hz,
80hz couse there is 16 SUB1500 drivers from 0hz to 80hz in the auroratowers.

http://www.pal.pp.se/~nord/projektaurora/prod/prod014.jpg
(statement, and crazy.. and just right)

well, from 80hz.
And more capable than the s9500 when it comes to SPL. And offcourse
lower distorsion (the s9500 has no noticeable distorsion).

I dont know if the 2451 is the best driver to work with. I would like it to be
crossed higher than the K2, to have lower distorsion.
Maybe the driver in the new K2 s9800 will have better measurements than
the older drivers. Or perhaps some TAD will be better..

I have been thinking of two 1500AL, and maybe two 5" - 10" midranges.

I´m just brainstorming here.

Niklas Nord
06-16-2005, 02:20 AM
You mention crossover frequencies but not SPLs... based on your earlier posts I believe you also want high SPLs. A conventional dome tweeter with a sensitivity of 91dB/watt will have quite a bit of distortion at 125dB. A TAD 4003 with a sensitivity of 112dB/watt will have the lowest distortion, a JBL 2451 will also have very low distortion.

WIdget

But, is it not true that the higher the sensitivity, the higher the horn distorsion will be, couse to air compression?! Is it realy nice to have as high sensitivity
in the horn driver?

whgeiger
06-16-2005, 03:19 AM
NN,

This brief, by its very nature, only highlights major design issues. Further investigation is required to perfect a loudspeaker system design.

First, there is no such thing as a ‘full range driver’. Those purporting to be “full range” use an acoustical crossover mechanism to perform the frequency separation between smaller and larger radiating portions of the driver diaphragm. At best they are just another version of a multiple radiator two- or three-way system

The audio spectrum spans approximately three decades of frequency; e.g., (1) 20-200 Hz., (2) 200-2000 Hz. and (3) 2,000-20,000 Hz. To expect a single driver to materially cover more than one of them is simply unrealistic, if achieving audio reproduction at low distortion is the principal design objective.

The second decade (2) is critical as it encompasses most of the speech range (SR, 300-3,000 Hz) where the human ears are most sensitive. For this reason carving up SR into two separate pieces to facilitate use of ‘common’ drivers is ill advised. Therefore, it is this region that should be addressed first by any regimen used for quality loudspeaker system design. Finding drivers for this range will be extremely difficult, for surprisingly only a few exist and some are quite expensive. For the loudspeaker designer, this state of affairs leaves much to be desired in the driver marketplace.

Once the midrange issues have been satisfactorily resolved, then the special needs of spectrum extremes may be addressed with some modicum of rationality. Even though SR carve-up is avoided, the crossover overlap regions must be addressed with care in regards to phasing/delay, matching radiating patterns and protection of the HF unit from LF signal overload. High pass filters with 12dB/Oct slopes provide barely sufficient filtering for this purpose.

At the low end (1), getting ‘big air’ cleanly and effectively overcoming room acoustics to achieve uniform frequency response will be the key design challenges for this decade.

At the upper end (3), maintaining a uniform radiating pattern without compromising frequency response will be the challenge. Pattern control here is essential for

a) minimizing early reflections from nearby walls that tend to smear the acoustical image and

b) maintaining ‘sweet spot’ dimensions of reasonable proportions so that more than one person can fully enjoy the reproduced performance.

These objectives are of course also applicable to the speech range (SR) as well; but there, such control is relatively easy to achieve.

Regards,

WHG

Niklas Nord
06-16-2005, 03:32 AM
OK, what drivers do we have that would suit something like this?

loudspeaker working range 80hz to 16khz or somewhere 22khz is better !
2 way or 3 way !

ultra low distorsion

grumpy
06-16-2005, 10:27 AM
2012H might be a current/JBL way to start...

ralphs99
06-19-2005, 05:52 AM
Hi WHG,

I read your comments with interest. Not having a crossover in the 300-3kHz range sounds like a good idea. It seems logical. But what evidence is there that having crossover points outside this range is the best engineering compromise? Certainly for horn loaded systems covering this decade with a single driver is a difficult challenge. I believe that due to the paucity of quality midrange direct radiator drivers on the market it's also a challenge there.

Why 300Hz? Why 3kHz? Many of the fundamentals of instuments lie below 300Hz. The ear is more sensitive at 4kHz than at 500Hz. So it seems to me that the midrange driver really needs to cover the range of 200Hz to 5kHz or more to satisfy the requirements of not having a crossover where the ear is sensitive and not having a crossover in the region of the fundamental of the male voice and the majority of musical instruments.

This seems impossible to achieve while at the same time satisfying the desirables of adequate SPL, low distortion, controlled directivity, low stored energy etc. Maybe a carefully placed crossover point in the region of 1kHz is a better compromise. JBL and many other manufacturers seem to think so.

What do you think? I'm interested in hearing peoples opinions on the matter. Because as you point out WHG, the midrange is where speaker system needs to begin.

Cheers, Ralph.

whgeiger
06-22-2005, 04:50 AM
RS99,

Response to your comments follows them.

Regards,

WHG


I read your comments with interest. Not having a crossover in the 300-3kHz range sounds like a good idea. It seems logical. But what evidence is there that having crossover points outside this range is the best engineering compromise?

Engineering metrics depend on the weights applied to the parameters in opposition (whose values must be traded). In a DIY project, cost takes a back seat to reduction of all forms of distortion. Here, ‘sweat equity’ is applied that allows component expenses to be elevated (somewhat) within a fixed project budget.


Certainly for horn loaded systems covering this decade with a single driver is a difficult challenge. I believe that due to the paucity of quality midrange direct radiator drivers on the market it's also a challenge there.

There is no material difficulty in getting slightly more than a decade of quality bandwidth from well designed drivers. For the mid-range unit, it is getting coverage at the low end that presents a problem. A ‘beefy’ driver, with a cost to match, is required to get the job done. For typical commercial systems, use of such drivers is avoided for obvious reasons. Acoustical subtitles of a performance in a P.A. setting are simply masked-out by the attendant elevation of the ambient noise level. At-home, the opposite is true, except while a ‘heavy’ dance party is in progress.


Why 300Hz? Why 3kHz?

These boundaries are certainly not sacred; however, the span between them is important so that the acoustic ‘load’ is ‘fairly’ shared amongst the drivers used. 300 Hz – 5,000 Hz. might be a better choice given the latest release of ISO 226 (:2003) [1].


Many of the fundamentals of instuments lie below 300Hz.

The ear is more sensitive at 4kHz than at 500Hz. So it seems to me that the midrange driver really needs to cover the range of 200Hz to 5kHz or more to satisfy the requirements of not having a crossover where the ear is sensitive and not having a crossover in the region of the fundamental of the male voice and the majority of musical instruments.

The fundamentals generated by, pipe organs, synthesizers and percussion instruments collectively span the entire audio spectrum; so, the only constant is the sensitivity of human hearing. The fact remains that at the frequency extremes, ear sensitivity is in decline.

If you look at the PWT power response of a typical compression driver you will note that, except for the displacement limitation, frequency response extends well below the recommended crosover point. What happens here is, as the space in front of the phase plug is increased to releive the displacement limitation, the concomitantly increased front chamber volume reduces the driver’s high frequency response. At the upper end, extension to or slightly above 5,000 Hz. is certainly desirable if the added cost can be tolerated. ALE [2] and GoTo Unit [3] make compression drivers to do this.


This seems impossible to achieve while at the same time satisfying the desirables of adequate SPL, low distortion, controlled directivity, low stored energy etc.

While seemingly true, typical manufacturers of loudspeaker system have motives materially different than ours, and in many cases, are not even designing products for the ‘niche market’ we represent. Remember, we are using horn drivers in an ‘at-home’ setting that were designed for use in P.A. venues. Tradeoffs made here extend well beyond matters that are simply acoustic.


Maybe a carefully placed crossover point in the region of 1kHz is a better compromise. JBL and many other manufacturers seem to think so.

This action requires use of a multiplicity of small [Sd] bass drivers so that their operation remains materially within the ‘piston range’, while collectively providing the required volume displacement. Alternatively, what emerges is a four-way system design. If a horn enclosure is used, it is typically folded up and placed in a room corner to reduce room occupancy. Irrespective of prevailing marketing hype, performance of such designs above 400 Hz. is at best questionable.


What do you think? I'm interested in hearing people’s opinions on the matter. Because as you point out WHG, the midrange is where speaker system needs to begin.

That is the precisely the reason for the post: getting that point across and placing design focus where it needs to be. While "the devil is in the details”, they still remain ‘window dressing’ for the design approach outlined here and certainly do not form a basis for its refutation.

References:

[1] Full Revision of International Standards for Equal-Loudness Level Contours (ISO 226)http://www.aist.go.jp/aist_e/latest_research/2003/20031114/20031114.html (http://www.aist.go.jp/aist_e/latest_research/2003/20031114/20031114.html)

[2] ALE A7550 Mid-Range Compression Driver (100-10,000 Hz)
http://www.tachyon.co.jp/%7Esichoya2/ale/drivers.html (http://www.tachyon.co.jp/%7Esichoya2/ale/drivers.html)

[3] Goto Unit SG-570 Mid Range Compression Driver (200-8,000 Hz)
http://www.eifl.co.jp/index/goto/goto.htm (http://www.eifl.co.jp/index/goto/goto.htm)