Beyma CP25 questions

[Shock-Late]

What's the correct CP25 orientation? should the "cheeks" be horizontal or vertical? What gives out the widest dispersion?

[Al Klappenberger]

Shock,

I asked the sales rep a Beyma of America that very question when I bought my first two. His response was that it depends on the dispersion you want. That didn't help me very much!

Anyhow, look at "my stereo stuff" with (View profile - System Profile) buttons.

That will show you how I mounted mine.

Al K.

[J.4knee]

Their web site lists the dispersion as 100 H by 60 V so I would assume from that, the "baby cheeks" should go horizontal. I could be wrong though.

[WMcD]

I continue to noodle with my favorite theory. It is that horns should be aimed not so much at the listener, but away from nearby reflective surfaces. This is to prevent side wall echos.

So my thought is that corner speakers, toed in, are best canted directly at 45 degrees from the wall. If we have a narrow dispursion, it should be horizontal. K-Horn mids do this by design.

The other part of the noodle is that if you have a center speaker, the first bad reflection is going to be from the floor or ceiling. So the narrow dispursion should be vertical.

The latter may be why the tractrix horn is flipped over in some Klipsch center channel designs; or why people rave about the vertical CW.

In any case, it is darn difficult to make valid measurements and people's individual rooms vary quite a bit.

Gil

[Al Klappenberger]

Guys,

Ok.. A picture is worth a thousqand words and a math equation or two. Here is a series of screen shots from my filter design program (PCFILT). The Fourier section was designed to analyse what a filter does in the time domain. Here I have done the analysis with NO filter, just the harmonics of a 1 KHz square wave. I have the fundamental (no harmonics), fundamental and 3rd harmonic, 3rd and 5th harmonic, and with all up to the 19th, which represents 19 KHz or the limits of human hearing. The bottom shot is including all harmonics up the 101th. This represents frequency response to 101 KHz. Note that it takes extremely high numbers of harmonics before the "ringing" goes away.

BTW: This is a graphic plot of exactly the equation John Warren posted earlier.

Al K.

[Al Klappenberger]

Guys,

Ok.. To add insult to injury, here's another mind bender!

Here is a computer study of a 6 KHz square wave through two netwrok tweeter filters. The first is the tweeter filter in my "ALK" type "A" network. It' 3rd order or 18 dB / octave. The second and third are through my ES5800 extreme-slope tweeter filter. That's about 120 dB / octave. The first two include harmonics to beyond 100 KHz. The bottom plot includes only the harmonics the human ear can hear (only ONE, at 18 KHz).

I'll bet'cha it's not what you expected to see!

Al K.

[Elso Kwak]

Hi, the Beyma CP21/F has even better dispersion and frequency response. Any experience with that one?

Al, a Bessel filter has no ringing or overshoot.

[Al Klappenberger]

ELSO,

There is a very important point you are missing. A Bessel lowpass has

linear phase and will not distort a waveform. A Gaussian filter is the best

filter type to use for transient response, not Bessle. BUT-- A highpass

filter can never be linear phase! Therefore, using anything but a Butterworth

or Chebyshev highpass filter becomes a waste of parts! When you "invert" a

linear phase lowpass prototype into a highpass filter the phase

characteristics are lost leaving you with only it's very poor amplitude

response.

Attached is a computer analysis of a 4th order Bessle 6 KHz highpass filter.

As you see, the square wave response is virtually the same as with the

Butterworth filter.

Al K.

[Al Klappenberger]

Guys,

Here's some more variations on a theme.. I set up a fast-rise square wave

generator (Wavetek 178) set to 2000 Hz into a Krohn-Hite 3500 tunable filter

into my SD375 FFT analyzer. The attached plots are the time response and

spectrum of the 2000 Hz square wave through two different filter settings.

One is a 2 KHz 4th order (24 dB / octave) highpass filter. The other is with

the filter switched to a 20Hz highpass. 20 Hz is so low that it can be

considered no filter at all to a 2 KHz square wave.

Note that the time domain response looks just like the computer predicted

response for the 6 KHz highpass I posted earlier. Also, the spectrum is

unchanged by the switching the filter from 20Hz to 2 Khz. The change in the

time domain is caused by the nonlinear phase response of the filter. I

believe that the human ear is totally deaf to phase errors between the

harmonics of a complex waveform like a square wave. Your brain can only

perceive the presence and amplitude of the harmonics. All it knows is the

"spectrum". It only hears the spectrum since there is no reference for phase

between different frequencies. They are changing far to fast for the brain to

follow. It takes an oscilloscope to do that! Since you have two ears, it can,

however, perceive the difference between the phase of a single component

(harmonic or fundamental) in both ears simultaneously. I believe that this is

how the brain constructs a stereo image.

I should be honest here and point out that it is only MY OPINION that the

human ear is deaf to phase errors between components of a complex waveform.

This also says that stuff like phase linearity and group delay is simply

stuff to scare the uninformed and means nothing! I have seen technical

articles that support this point of view, so it's not just my opinion. I do

have some support!

Another interesting point. A 6KHz square wave has it first spectrum component

(harmonic) at 3 times 6 KHz. That's 18 Khz! Most people (including me) cannot

hear 18Khz. Therefore, you can't hear any difference if you switch from a 6

KHz square wave to a 6 KHz sine wave! I have tried this, and I can't hear

anything but a slight loudness change.

Al K.

[rigma]

Al Klappenberger,

I have your crossovers in my Khorns and have JBL 2404H tweeters I would like to use in this setup. Is there anything I need to do to the crossover to use these tweeters?

Rigma

[Al Klappenberger]

Rigma,

No.. I think those tweeters will drop right in. All you might want to add an attenuator (L pad) if they seem to hot. I think they are an honist 105 dB sensitivity. I'm not sure the speaker is that sensitive. Also the loss in my network tweeter filter is very low. This might all add up to a very hot tweeter compared to the rest of the speaker. Let your ears decide after the new tweeters are installed.

Al K.

[dubai2000]

Rigma,

like Al suggested, I simply put the JBLs in without any crossover mod and I am very happy with the result....and no, I do not find the tweeter too hot, it's actually a lot smoother than the K-77 (IMO)...but of course that's personal taste.

Wolfram

[rigma]

Thanks AL & Wolfram, I am working on the mounting baffle as we speak!

Rigma

[Elso Kwak]

Taming the Klipschorn

Hi Al(l),

I just put a 1kHz square wave to my Klipschorns from a test CD through a NON-OS DAC.

I measured on the mid driver. Of course the tilt on the top is high due to the limited low frequency response but I DO NOT have the wild ringing and overshoot you pictured.

I found a Bessel filter was slightly better than a Butterworth sonically. My Bessel high-pass filter is not an "inverted" low-pass. Also I found that a third order Butterworth high-pass filter for the squawker was better than the first order as in the Klipsch AA or A network, especially at high volume. This is possibly due to an intermodulation effect of unwanted low frequencies on the voice coil.

As the Bessel has less steep initial fall-off I went four a fourth order. I tried your filters too. The whole exercise is all about sound quality. Do I need to say more?

<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />

[Al Klappenberger]

Elso,

First off, you said that your Bessel filter was not an inverted lowpass.

Well, actually IT IS! I was able to duplicate your exact tweeter highpass

filter by inverting and scaling the following normalized lowpass to 6 Ohms

and 6000 Hz. The normalized lowpass "G" values are: 1.5012 .9781 .6127 .2114.

You probably designed it from a formula in a book for a Bessel crossover

where you get values for both a lowpass and highpass. This method still

derives it basis from normalized lowpass element values! The main point is

that a linear phase lowpass filter (maximally flat delay) will NOT remain

linear phase when inverted to a highpass (as yours is). All you get when you

do that is the poor stopband rejection of the Bessel response withOUT its

linear phase advantage. It's a waste of parts! Besides, your ear is deaf to

phase linearity anyhow!

"I measured on the mid driver. Of course the tilt on the top is high due to

the limited low frequency response but I DO NOT have the wild ringing and

overshoot you pictured."

Of course you didn't see the ringing at the midrange driver. My plots posted

earlier were all of the TWEETER filter, not the midrange! The bandpass filter

(high-lowpass combination) in your squawker section removes the harmonics

that causes the ringing. Put a 6KHz square way into your tweeter and you will

see them!

You keep saying stuff like "all about sound quality". This is true, but

everybody's hearing and interpretation of sound quality is different. You can

only say what sounds good to YOU! This is why I will almost never say "this

sounds better than that" unless the difference is like night and day! Each

person hears what he wants to hear (including me). When you get down to

Bessel versus Butterworth filters, you can only quantify the difference by

instrument readings and theory. The sonic difference as you perceive it is to

small to be absolute about one over the other EXCEPT for your own use. Even

then, be sure you are not kidding yourself. It can be effected by anything

from what pills you are taking to your last walk in the park!

Al K.

[Elso Kwak]

Al,

My midrange filter is basically the same as for the tweeter.

But if you don't want to listen and only measure....

These loudspeakers are made to listen to NOT to measure endlessly.

[Al Klappenberger]

Elso,

Of course you need to listen! just don't kid yourself into thinking that you ear can hear everything on an absolute basis! An instrument won't lie to you but your ears can! On the other hand, your instrument may not be measureing the correct thing. Here's where you ears are needed. Hear it, then ask how can I measure that? If you can find a way, do it! If you can't find a way, then the entire subject becomes subjective.

Al K.

[dhsettim]

I have tried the Beymas bought mine from DIYparadiso for a very good price. I have them sitting on a couple of books ontop of my 1977 Khorns and use a type A crossover with Hovland and russkie teflon caps (no woofer inductor). I'm sure they are not as good as the more expensive JVC but I think they sound great, but then I admit that I have cloth ears. Thanks Al for the suggestion.

tim

[Al Klappenberger]

Guys,

Getting back to the original subject ....

Elso suggested the Beyma CP21 tweeter. I looked it up on the Beyma web site and checked the specks. It looks very good on paper but it is not a horn. It looks to be a "slot". Slots have super dispersion but also have a reputation of sounding like "frying pans". I am very happy with the CP25, so I don't think I would experiment further.

Al K.

[Elso Kwak]

----------------

On 5/18/2004 6:01:24 AM Al Klappenberger wrote:

Guys,

Getting back to the original subject ....

Elso suggested the Beyma CP21 tweeter. I looked it up on the Beyma web site and checked the specks. It looks very good on paper but it is not a horn. It looks to be a "slot". Slots have super dispersion but also have a reputation of sounding like "frying pans". I am very happy with the CP25, so I don't think I would experiment further.

Al K.

----------------

Baby-Butts or Frying Pans

Al,

Now you are doing it yourself. Though measurements speak against you, you condemn the BeymaCP21/F slot tweeter. As Paul Klipsch pointed out distortion is inverse to efficiency I doubt this is a bad tweeter. No measurements to corrobate your comments? Why would a slot tweeter be worse than a horn tweeter?

And how would you describe the sound of the Electrovoice T350? Backing eggs?

Al you have become totally inveracious to me.