Hi Ian and Adam
Thanks. I'm not really in agreement with what he says there. I think that there are a lot of mistakes. When I look at his data, I think that it proves my point about HOM and contradicts his point about phase plugs. But alas, this is all complex stuff and people are looking at it in different ways. In the end it will all shake out as to what's what, but in the mean time we are likely to see a lot of contradictory data and opinions, and a multitude of different explainations of the same thing in different ways.
Thanks. I'm not really in agreement with what he says there. I think that there are a lot of mistakes. When I look at his data, I think that it proves my point about HOM and contradicts his point about phase plugs. But alas, this is all complex stuff and people are looking at it in different ways. In the end it will all shake out as to what's what, but in the mean time we are likely to see a lot of contradictory data and opinions, and a multitude of different explainations of the same thing in different ways.
What occurred to me was that compression driver models, such as that used by Keele. Depict the acoustic compliance of the front chamber as an inductance, this would seem to have a low pass effect and increase the apparent diaphragm mass in this region, this being consistent with the fact that conical horns for instance are measured to go down lower than simple theory would have it.
What seems to be occurring in Gunns data is energy storage in a second order system.
rcw

What seems to be occurring in Gunns data is energy storage in a second order system.
rcw

rcw said:
Weather the front compliance is an inductor or a capaciator depends on the variables in the model. But the front chamebr is a LP filter, that much is clear.
I think that Gunness data is not simply a second order system, not in the lumped parameter sense. It is second order in the wave equation, which is really the only way to explain it.
Very cool! We await the results.
Are these 400Hz profiles for an 800Hz crossover point? There is a wide choice of good drivers to do that. Nice.
Are these 400Hz profiles for an 800Hz crossover point? There is a wide choice of good drivers to do that. Nice.
On looking at the Gunness data it struck me that if you hit a bunch of Helmholtz oscillators with a pulse it might well look like that.
Then if this is so you need two distinct coupled air masses to make a Helmholtz oscillator, and the typical phase plug looks like a bunch of coupled air masses.
I had the idea that if you made a phase plug that had no coupled air masses but consisted of for instance an annular space in front of a ring, with perhaps a few radial vanes to suppress circumferential modes, this might well be of benefit.
Rcw.
Then if this is so you need two distinct coupled air masses to make a Helmholtz oscillator, and the typical phase plug looks like a bunch of coupled air masses.
I had the idea that if you made a phase plug that had no coupled air masses but consisted of for instance an annular space in front of a ring, with perhaps a few radial vanes to suppress circumferential modes, this might well be of benefit.
Rcw.
We have chosen Le Cléac'h because it's bit more universal so that you can cut it a bit for a variety of ordinary dome tweeters with exit port 30-35mm but still have 1" for compression drivers. On the picture 2" throat is marked with vertical line nearly in the middle of the horn. These small horns we can make from one piece. Later on we will try 100Hz horns with flat drivers.
I don't have any decent and cheap compression driver so can't design for particular model. Anyone tried <a href="http://www.tb-speaker.com/detail/1230_04/w1-1828s.htm">Tangband W1-1828S</a>?
OK, I see the 2" cut point. So you could try them with 1", 1.4" and 2" drivers. It's going to be interesting.
That's a strange little TB driver you linked to - never seen it. (BTW, your WWW button link isn't working - check the last part of the URL)
@ WT. Man! You are really doing it the hard way. Bravo!
What is your intended use?
That's a strange little TB driver you linked to - never seen it. (BTW, your WWW button link isn't working - check the last part of the URL)
@ WT. Man! You are really doing it the hard way. Bravo!
What is your intended use?
rcw said:
To me Gunness "time smear" is audio-speak for group delay. I prefer the generally accepted engineering term. That the impulse response has group delay at higher frequencies is what I have been saying for at least a decade now, long before Gunness showed it and gave it a new code name. The group delay is from two sources, one is the internal reflections that he also acknowledges and the other from the HOM which he does not acknowledge.
Now he claims to be able to use DSP to correct this, but his own data shows that if he corrects it on-axis, then it is no better and maybe even worse at off axis locations. Again, this is exactly what I have been saying all along, that the group delay aspects of the sound field cannot be globally corrected with EQ because they are different at every field point.
So what he says that I agree with is not new to me, and the rest, I don't really agree with - that DSP CAN correct these problems.
With 2" will be too short.
Here it is:
http://www.tb-speaker.com/detail/1230_04/w1-1828s.htm
Very interesting how they manage 120Hz from 1" 🙂
It's only available directly from TB. Waiting for Ms. Diana calculation.
Do you know anything about Wharfedale compression drivers from EVPX series? Can buy a pair with waveguides cheaply.
http://tinyurl.com/m4b2q5
Notice - it's 10dB down at 120Hz. That's how they get it. 😉
Interesting waveguid on the Wharfdales. See that they are crossing from the 15" at 2.2KHz. Right were that 15" should be screaming.
Interesting waveguid on the Wharfdales. See that they are crossing from the 15" at 2.2KHz. Right were that 15" should be screaming.
gedlee said:
Exactly right. If the group-delay is different at different spatial points, DSP will degrade off-axis time and frequency response, not improve it.
Lynn Olson said:
Basically this will lead to constant directivity designs - or to how John Kreskovsky put it - to loudspeaker designs acting "min phase" over a wide room angle.
😉
Soongsc's contour is spot on in this reagard (up to the point where all horns / wave guides fail due to throat dimension).
I am outlining this in detail in my paper to come, titled:
"Dipole Horn - never heard ?
(introducing "Dipole Constant Directivity Device” (DCDD) to the audio community)"
🙂
Michael
I was watching that baby gradually appear yesterday, will be picking it up in a few days.mige0 said:
jzagaja said:
Guys, I know that everyone thinks that you can take a cheap TangBand or Hi-Vi speaker and slap it on a waveguide, and get something that will play full range with directivity.
Unfortunately, it doesn't work.
All of the horn gurus told me the same thing three years ago, but I charged ahead anyways, and found out the hard way.
GM and Tom Danley clued me into the fact that you have to work the other way. Basically you need to figure out what frequency range your waveguide will cover, and then choose a driver that works on the waveguide.
That's the opposite way that most people build speakers. I think that most people see a driver that looks "promising", and start scheming about building a speaker with it.
The two best free places to learn the math are here on diyaudio, and Martin King's documentation for his software program is an excellent free reference on horn math.
http://www.quarter-wave.com/Horns/Horn_Theory.html
Based on the parameters of the tang band speaker you linked to, you'd be luck to get it to play to 600hz in a waveguide with a compression chamber. That's why compression drivers have incredibly light and fragile diaphragms, coupled with enormous motors. Those two features give compression drivers the ability to play to 5 or even 10khz.
Admittedly, you could forego the compression chamber, but then you run into a whole 'nother set of problems. (lack of excursion, bad polar response due to the diaphragm shape, cancellation, etc...)
John
Quite correct.
Designing a good waveguide driver systems is not all that easy - thats why there are so few arround. But if I say that I'm being "arrogant" and talking down to people. Let them learn the hard way I guess.
Quite correct.
Designing a good waveguide driver systems is not all that easy - thats why there are so few arround. But if I say that I'm being "arrogant" and talking down to people. Let them learn the hard way I guess.
panomaniac said:
I probably sound like a broken record at this point, but you really need a compression driver when you're building a horn or a waveguide.
If you use a conventional woofer, it usually rolls off way too soon, sometimes as early as 500hz.
In the process of evaluating woofers for a Unity horn, I stumbled across one very unique loudspeaker. It's a paper-coned woofer with a fairly small diameter, and a freaken' huge motor on it. There are plenty of other woofers like that, but this one is neodymium/ferrite 😀
http://www.miscospeakers.com/SpeakerDetail.cfm?SpeakerID=319
Here's a pic of mine, making a bunch of other midranges jealous 😉
Because of this goofy motor, the specs show that it can cover 350hz to 2khz.
That's pretty unique - I've never seen a woofer that could do this, normally you'd have to spend the big bucks on a compression driver to get that kind of performance.
It's only $30 too. Four of these at the apex of a 30" waveguide would cover the entire midrange, and they'd give you about 110DB with one watt!
Of course there are compression drivers that can do that too:
http://www.parts-express.com/pe/showdetl.cfm?Partnumber=294-584
An externally hosted image should be here but it was not working when we last tested it.
What sets the KCN5FD apart is that it has a ton of displacement to play low with authority, whereas a compression driver is struggling to play 400hz. The fact that you can buy ten of them for the price of one compression driver is icing on the cake
