The hole directly "on-axis" sounds to me like diffraction from the mouth. I see this ever day. You won't see it in many horns because the wavefronts are not coherent enough to create it. But if the wavefront is very coherent (i.e. very nearly purely spherical) then there will be a cancellation on axis at some frequency (there has to be) because of the diffraction from the edges of the mouth. Rounding and foam help, but even rounding does not elliminate it in my waveguides because they are axi-symmetric and the wavefront is very coherent.
It is true that this hole is very small in spatial extent and as such has very little effect on the sound power and virtually no effect if one is not on-axis when listening.
It is true that this hole is very small in spatial extent and as such has very little effect on the sound power and virtually no effect if one is not on-axis when listening.
Confining this to the bandwidth below 100Hz ( a bass horn ) how significant is the effect of mouth edge diffraction?
Theoretically speaking, would a continuous horn expansion flare ( say like a Cornu Spiral ) suffer less effect of mouth edge diffraction than the sub horns that are built from plywood, that have disruptions in the flare rate?
Or is it of no significant with the limited bandwidth used?
Theoretically speaking, would a continuous horn expansion flare ( say like a Cornu Spiral ) suffer less effect of mouth edge diffraction than the sub horns that are built from plywood, that have disruptions in the flare rate?
Or is it of no significant with the limited bandwidth used?
Just a pet pieve of mine - I've expressed it often, most recently in the Audio Express article on horns.
At very LFs the shape of the walls is irrelavent, its only the inlet and outlet areas that matter and any connection between them is going to do approximately the same thing. This position was objected to by Bjorn, but the next Audio Express has some simultations that prove my point.
Horn shape is only a factor when the changes in this shape are comparable to a wavelength. Thus for a 40 Hz signal - about 25 foott wavelength - the horn would have to be long enough for a change in the surface of about 5 feet - or lets say a ten foot diameter mouth. It would take one heck of a horn to have a mouth this size! Any size less than that and basically the shape doesn't matter because the wavlengths don't even see the shape.
At very LFs the shape of the walls is irrelavent, its only the inlet and outlet areas that matter and any connection between them is going to do approximately the same thing. This position was objected to by Bjorn, but the next Audio Express has some simultations that prove my point.
Horn shape is only a factor when the changes in this shape are comparable to a wavelength. Thus for a 40 Hz signal - about 25 foott wavelength - the horn would have to be long enough for a change in the surface of about 5 feet - or lets say a ten foot diameter mouth. It would take one heck of a horn to have a mouth this size! Any size less than that and basically the shape doesn't matter because the wavlengths don't even see the shape.
Thanks for the clarification.
I have read the material and your reply.
I look forward to the next installment.
Sometimes in articles like Bjorn's, for the sake of brevity the detail of when phenomena is significant is lost.
I understand the need sometimes to describe in general terms and the practical numbers you just described help.
Ever since the subject of diffraction came up in those articles I have attempted to get a functional or better understanding of it - how and when it applies in audio.
Frankly I have found very little.
I don't want to draw an incorrect inference; At what point ( frequency wise ) or under what circumstances is diffraction significant?
I have been told that the bends that are inevitable in folding a horn act as "low pass filters" and was one of the factors that simulations don't account for. I have not seen anything on this. I wondered if this was accurate?
On a side note: the larger horns I heard were 8'x8' at a concert in the old Cleveland Indians Ball Park in the 70's. As I recall there were 8 total.
Thanks again
The problem with diffraction is that it is a very wide ranging phenomina. Basically it almost always occurs, but is often insignificant. Think of it this way, whenever a wavefront has to go around a discontinuity in a surface there has to be diffraction. So if you think about it is ever present. Mathematically it is so hard to deal with that its all but ignored in the literature. There are only a very few situations where it can be handled mathematically and those give us some idea of what goes on, but real world situations tend to be so complex that they defy analysis and so the phenomina is generally ignored. The folded horn issue is a classic example. Its often done and typical examples "suggest" that a LP function is evident, but the reality is that every situation is different. Its so complex and there are so many ways that it can go wrong that I simply don't do it.
gedlee said:
I have personally tried a lot of exotic subwoofers, from tapped horns to hyperbolic horns to dual-reflex bandpass.
I grudgingly tried using multiple subs this year. I say "grudgingly" because I *enjoy* building crazy speaker projects, and the thought of settling for a simple bandpass was something that I wasn't excited about.
But I gotta admit, multiple subs sound amazing.
In fact, they sound so amazing that I'm going from three subs to eight.
The reason I bring this up is that four or eight subwoofers can easily match the efficiency and maximum SPL of a horn sub, AND it offers many advantages in a real room.
In other words, horn subs are fun to talk about and build, but in the real world they're trumped by multiple small subs.
Admittedly, multiple unconventional subs have advantages in prosound, where the SPL and power levels are dramatically higher.
gedlee said:
When I moved the Summas into a different room of the house, I only moved one sub. As soon as I fired it up I noticed a difference immediately -
I could tell where the subwoofer was located. It was obviously in front of me.
As soon as I had three subs going again, my ability to localize the subs disappeared.
The main reason I'm doing eight is that they're very cheap and small. I suppose three mega-subs might keep up with the Summas. The subs I'm using are uber-cheap, about $25 each.
Patrick Bateman said:Here's some new (and better) models of my Tang Band mids for a Unity horn.
If you want to build this horn, here's the details:
#1 - This is a sixty degree conical horn. The mouth is 30" in diameter (twice as big as my Summas.)
#2 - The throat has a diameter of one inch.
#3 - There are four midrange holes. Each hole is located two inches forward of the apex, as measure on the EDGE of the waveguide. In other words, just get a ruler, measure out two inches from the side, and DRILL THE HOLE!
#4 - The waveguide holes are half an inch deep and half an inch in diameter. This is a huge deal, seriously. If you change the depth or the width, the whole design will go to hell. Modifying the diameter OR the depth modifies the shape of the midrange response. In other words, modify the holes at your own peril
#5 - The most important parameter of all is the impedance curve. We're dealing with VERY small volumes of air here kids; it's literally impossible to build this thing properly unless you can measure the impedance curve. The impedance curve will show you if you have leaks, it will show you if the volume of your chambers is wrong, and it will show you if the holes are sized properly. Basically I recommend that you modify the air volumes and the hole size until the impedance curve looks correct.
Attached is a pic of the predicted frequency response.
The synergy horn uses a frustrum instead of a cylinder for the midrange ports.
It looks like this increases the high frequency output of the mids quite a bit. Here's a horn response model comparing the difference between holes that are 0.5" deep and holes that are 0.04" deep (virtually no depth whatsoever.)
In the graph you can see that the output at 1.5hz increases by five db, which is quite a bit.
Attachments
In 2006 I cloned a Unity horn, and threw it in my car. The sound was excellent, but it looked ridiculous. One of my friends said that "it looked like there was a bomb on my dash."
I can only imagine how the highway patrol would react if I was pulled over for speeding.
Of all the crazy speaker projects I've tried in the car, this was the best, but it was short lived. At the time I was commuting to another state for work, so I had the pleasure of listening to it for twenty hours a week, while I was stuck in the car. IMHO, the most notable aspect of the unity was it's intelligibility. I constantly found myself hearing things in songs that I hadn't noticed before. For instance, background singers, the "texture" of certain instruments, words to lyrics that were previously unintelligible, etc...
I believe this is because the Unity excels in the time domain, as the midrange and tweeters are very close together.
Last year I finally had a chance to listen to a "real" Unity horn at a friend's house, and I heard a lot of the same thing. The unity made real drums sound different than synthesized drums and could articulate the human voice in a way that only headphones can rival.
Yesterday afternoon I hooked up my my new tapped horn for the first time, but there was one little problem. I have no front stage in the car. All I have are the stock speakers. Putting the old unities back together would be a lot of work, and the new unities described here aren't finished.
So I went in the garage, found a box of drivers from the original project in 2006, and picked a good candidate. (When I built the Unity clone in 2006 I measured close to a dozen candidates to find the best. These were the drivers that didn't make the cut.) I put a single Aurasound 3" woofer in a baffle and mounted it to the kick panels of my Accord. No crossover, not even a box. Just a simple dipole. I cut the baffle so the response would go down to 250hz or so, and I offset the woofers to smooth the response.
And this crummy $15 speaker on a baffle sounded a LOT like the Unity I slaved over for months. WTF
There was an excellent soundstage that exceeded the boundaries of the car, excellent impact from the tapped horn, lots of dynamics, etc...
Best of all, the clarity was back; this cheap little aurasound woofer articulated the way that the Unity did.
At this point, all the fans of single-driver speakers are going to be saying "duhhhh." But I've NEVER tried a full range in the car. I've tried line arrays, horn loaded compression drivers, Unity horns and two ways.
I once tried a Monsoon planar in the car also. I'm guessing that the Monsoon didn't sound as good as the Aurasound because the baffle of the Monsson is way too small. (The panel measures 6" x 9" IIRC.)
All is not well in Paradise, unfortunately. The Auras have gobs and gobs of distortion at anything but polite levels. They're rated for 10watts IIRC, and I was running them at 35. There's only so much power you can pour into a 3" woofer.
But now I'm wondering if I'll ever get around to finishing this Unity:
When it might make more sense to just horn load a pair of 2" or 3" woofers.
I can only imagine how the highway patrol would react if I was pulled over for speeding.
Of all the crazy speaker projects I've tried in the car, this was the best, but it was short lived. At the time I was commuting to another state for work, so I had the pleasure of listening to it for twenty hours a week, while I was stuck in the car. IMHO, the most notable aspect of the unity was it's intelligibility. I constantly found myself hearing things in songs that I hadn't noticed before. For instance, background singers, the "texture" of certain instruments, words to lyrics that were previously unintelligible, etc...
I believe this is because the Unity excels in the time domain, as the midrange and tweeters are very close together.
Last year I finally had a chance to listen to a "real" Unity horn at a friend's house, and I heard a lot of the same thing. The unity made real drums sound different than synthesized drums and could articulate the human voice in a way that only headphones can rival.
Yesterday afternoon I hooked up my my new tapped horn for the first time, but there was one little problem. I have no front stage in the car. All I have are the stock speakers. Putting the old unities back together would be a lot of work, and the new unities described here aren't finished.
So I went in the garage, found a box of drivers from the original project in 2006, and picked a good candidate. (When I built the Unity clone in 2006 I measured close to a dozen candidates to find the best. These were the drivers that didn't make the cut.) I put a single Aurasound 3" woofer in a baffle and mounted it to the kick panels of my Accord. No crossover, not even a box. Just a simple dipole. I cut the baffle so the response would go down to 250hz or so, and I offset the woofers to smooth the response.
And this crummy $15 speaker on a baffle sounded a LOT like the Unity I slaved over for months. WTF
There was an excellent soundstage that exceeded the boundaries of the car, excellent impact from the tapped horn, lots of dynamics, etc...
Best of all, the clarity was back; this cheap little aurasound woofer articulated the way that the Unity did.
At this point, all the fans of single-driver speakers are going to be saying "duhhhh." But I've NEVER tried a full range in the car. I've tried line arrays, horn loaded compression drivers, Unity horns and two ways.
I once tried a Monsoon planar in the car also. I'm guessing that the Monsoon didn't sound as good as the Aurasound because the baffle of the Monsson is way too small. (The panel measures 6" x 9" IIRC.)
An externally hosted image should be here but it was not working when we last tested it.
All is not well in Paradise, unfortunately. The Auras have gobs and gobs of distortion at anything but polite levels. They're rated for 10watts IIRC, and I was running them at 35. There's only so much power you can pour into a 3" woofer.
But now I'm wondering if I'll ever get around to finishing this Unity:
When it might make more sense to just horn load a pair of 2" or 3" woofers.
I have one of the last pairs of the Lambda Unity's. I've got them internally and externally crossed at line level with a DEQX. I've contemplated going to a passive crossover in the horn, but would want to improve on the version that was originally supplied. To do that, I would want to extend the HF response of the mids to have more flexibility on the crossover point. This is not much of a problem with the DEQX. With moderately steep filters (even 48dB) I can roll off the compression driver to keep distortion low. Not so easy with passive unless I can move up the response range of the mids.
I can reduce the volume of the mid chamber some, and that does shift the response up a little. I've considered getting a couple of the Aurasound drivers to play with. It would be easier to play with the bandpass parameters. Also, I've wondered about the effect of having the exit holes well off center for the 5"ers. It would seem that this has to affect phasing some. With the smaller drivers, I could center the diaphragm on the exit, such that the path length is close to the same for all points. Could even use 8 per side. Any thoughts?
Sheldon
I can reduce the volume of the mid chamber some, and that does shift the response up a little. I've considered getting a couple of the Aurasound drivers to play with. It would be easier to play with the bandpass parameters. Also, I've wondered about the effect of having the exit holes well off center for the 5"ers. It would seem that this has to affect phasing some. With the smaller drivers, I could center the diaphragm on the exit, such that the path length is close to the same for all points. Could even use 8 per side. Any thoughts?
Sheldon
Patrick Bateman said:
Double "duhhhh.........."
RS use to sell a little 3.5" 'FR' driver (wide BW tweeter actually) which with WG and TL rear chamber mounted under the dash of a '69 El Camino did a marvelous job combined with the 'stereo' woofers (inductor rolled off Sony 6.5" 'FR' drivers actually) behind the seat. Hardly a dB racer, but loud/clean enough to play DSOTM, etc. over open windows filled with the sound of a much hotrodded 350 'singing' through tuned tunnel induction/pipe headers/'Cherry Bombs' at freeway speeds. What more do you need? Gees, I miss those days....... Don't miss the smog though.
GM
nickmckinney said:
Nick,
A few months back I had an opportunity to listen to one of your Unities, using the TAD 2001. The TAD reminded me a lot of the BMS 4540ND, with an airy, delicate presentation in the top octave.
On page ten of this thread I argued that the BMS is an excellent match for two-way horns, because it can play out to 20khz, while conventional compression drivers cannot.
Yesterday I was suprised to find that JBL is now using a BMS ring radiator in the newest version of their M-Pro line. This gives us an opportunity to compare the response of a BMS compression driver and a conventional compression driver on the same waveguide.
This is because the predecessor to the MP 415 used a conventional compression driver, which has now been replaced with a ring radiator. The baffle is the same, and both designs use a 15in woofer. The BMS is crossed over a bit lower, at 1.5khz. (The old design used a 1.7khz xover.)
The crossover, schematics, dimensions etc are all available from JBL. The woofer is $179 and the compression driver is $140. The waveguide isn't for sale
The top pic is the conventional compression driver; the bottom pic is using a BMS.*
* See this thread for an explanation of JBL's relationship to BMS:
http://www.audioheritage.org/vbulletin/showthread.php?t=6368&page=7
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.