I have a pair of QSC HPR152i clones along with B&C DE250 drivers. The horns are 14"x10" horns. I also have a pair of JBL 2226H.
I'm curious to know if there are any obvious warning signs before I start making sawdust. Specifically, I am concerned about the center-to-center distance between the CD+Horn and the Woofer. I've read before that the CtC distance should not exceed the crossover frequency.
The closest I can get the center to center distance is 12 inches, and that's with overlapping the horn and rear mounting the woofer.
12in = 304.8mm, which corresponds to a wavelength of 1,132Hz. I'm not sure if that horn+tweeter combo will play that low.
As I was perusing through Troels website trying to find an implementation of a waveguide in a speaker design, I came across his Quattro 2-way speaker.
In that specific design, the crossover frequency between the waveguide loaded tweeter and the woofer is 2.5Khz. The wavelength of 2.5khz is 137.6mm. The center-to-center distance per Troels website was 170mm.
Obviously Mr. Troels knows what he is doing and I'm sure he is well aware of this general rule of thumb. So there is clearly something at play that allows flexibility (maybe trading one benefit for another).
Thoughts?
I'm curious to know if there are any obvious warning signs before I start making sawdust. Specifically, I am concerned about the center-to-center distance between the CD+Horn and the Woofer. I've read before that the CtC distance should not exceed the crossover frequency.
The closest I can get the center to center distance is 12 inches, and that's with overlapping the horn and rear mounting the woofer.
12in = 304.8mm, which corresponds to a wavelength of 1,132Hz. I'm not sure if that horn+tweeter combo will play that low.
As I was perusing through Troels website trying to find an implementation of a waveguide in a speaker design, I came across his Quattro 2-way speaker.
In that specific design, the crossover frequency between the waveguide loaded tweeter and the woofer is 2.5Khz. The wavelength of 2.5khz is 137.6mm. The center-to-center distance per Troels website was 170mm.
Obviously Mr. Troels knows what he is doing and I'm sure he is well aware of this general rule of thumb. So there is clearly something at play that allows flexibility (maybe trading one benefit for another).
Thoughts?
Last edited:
"How much" - that depends on you. If you model your situation as I am learning to do, you can gain insight.
While modelling my TMW my spacing is around 370mm my cross over from M to T in the example is around 500hz....it gave me a cancellation around xover, slightly above 0 axis maybe 20-25 + degrees before it really kicked in...and then below 0 axis, the crossover being so low CTC no longer an issue so no noticeable lobing until far off like 60 (-)degrees.....
The above axis cancellation is probably common in a lot of designs with a horn/waveguide...I think when CTC starts getting pushed too far, you start to have a second cancellation creep up from below, and it resembles the narrow vertical polar seen with some MTM designs....the directivity of the horn doesn't help, I don't think...I say I don't think because, I am still learning but, the directivity of the horn, to me, resembles the beaming of a woofer being used higher than its diameter to wavelength relationship. A horn looses directivity past the width or length dimension so the polar can be seen as strongly influenced by source size and redundant source separation....
AS you can see above in the charts, its not an uncommon thing for a major manufacturer to have this issue on the vertical axis...
Attachments
Last edited:
Centre to centre distance makes most sense when applied to 2 theoretical point sources. In the real world, the large radiating surface areas surrounding the centres should mitigate the problem to some degree. 2.5kHz is pretty high though – you should experience some off-axis issues. Whether that's an issue is up to you.
Last edited:
twin columns line array interference?
James, I have a question re multiple sources/centre to centre interference, hope you can help?
Will the following bass/mid driver arrays have any obvious sonic issues I need to worry about?
I have built (and love the sound of !) various on wall line arrays using full range drivers and never had a problem with driver interference when listening at 3 meters away or more.
I would like to build an on wall array where each speaker is 1550mm tall and 500mm wide, (150mm deep) and each speaker has three columns of drivers, a centre line of AMT tweeters flanked left and right by a column of 6 inch bass / midrange drivers.
The bass mids cover 200Hz to 2,000Hz and the AMT's 2KHz up.
Each speaker sits above an 18 inch sealed box sub covering 20Hz to 100Hz.
Everything controlled by DSP and active amplifiers.
Thanks
Alex.
James, I have a question re multiple sources/centre to centre interference, hope you can help?
Will the following bass/mid driver arrays have any obvious sonic issues I need to worry about?
I have built (and love the sound of !) various on wall line arrays using full range drivers and never had a problem with driver interference when listening at 3 meters away or more.
I would like to build an on wall array where each speaker is 1550mm tall and 500mm wide, (150mm deep) and each speaker has three columns of drivers, a centre line of AMT tweeters flanked left and right by a column of 6 inch bass / midrange drivers.
The bass mids cover 200Hz to 2,000Hz and the AMT's 2KHz up.
Each speaker sits above an 18 inch sealed box sub covering 20Hz to 100Hz.
Everything controlled by DSP and active amplifiers.
Thanks
Alex.
PS, each speaker has massive Sd:
Sub - Single 18 inch sub with 1,165 cm square radiating area.
bass/mid - 18 (two columns of 9 drivers) times 150 Sd = 2,700 cm square radiating area.
Tweeters - 9 times 100 Sd = 900 cm cm square radiating area.
Total per speaker = 4,765 cm square radiating area.
A typical "big" three-way with a 12-inch bass (550 Sd) 4 or 5-inch mid-range (50 to 100 Sd) and a dome tweeter (under 5 Sd) for a total of around 600 to 650 cm square radiating area!
My aim is to have the absolute minimal driver cone/ribbon travel at any given frequency and SPL.
Sub - Single 18 inch sub with 1,165 cm square radiating area.
bass/mid - 18 (two columns of 9 drivers) times 150 Sd = 2,700 cm square radiating area.
Tweeters - 9 times 100 Sd = 900 cm cm square radiating area.
Total per speaker = 4,765 cm square radiating area.
A typical "big" three-way with a 12-inch bass (550 Sd) 4 or 5-inch mid-range (50 to 100 Sd) and a dome tweeter (under 5 Sd) for a total of around 600 to 650 cm square radiating area!
My aim is to have the absolute minimal driver cone/ribbon travel at any given frequency and SPL.
Last edited:
When you are listening from a position where your ears are equidistant from the two sources (e.g. midrange driver and tweeter), the distance between the sources isn't much of an issue (unless it is so large that it becomes obvious that different frequencies are coming from different directions). The inter-driver distance becomes an issue in two ways:
a) if you move in such a way that your ears are no longer equidistant from the two sources and this is sufficient to create a significant cancellation at frequencies around the crossover point,
b) the off-axis sound (with the frequency response problem described above) is reflected towards the listener by the room. The theory is that the sound of the reflection will not match the direct sound and that this will sound unnatural.
(Off topic, what is your connection with hydrogen and Cambridge? I recently started a synthetic biotech company with a research team in Cambridge.)
I feel like I agree with this theory, but I feel like its just one piece of the puzzle. The information regarding the critical zone, makes it seem like it can be an problem but it can also be a non issue depending on how the speakers are used. My perspective is from one that considers the sweet spot to be the only place that matters.
Thanks for posting, Dave. I didn't realize the recommendation is 1/4 wavelength.
I can't remember where I (mis)read the 1-wavelength recommendation. Forgive me, since I'm not very technical, but I was thinking about the typical 1" dome tweeter and 6.5" woofer 2-way bookshelf speakers. Most of these design crossover at 2.0Kz - 2.5Kz. The 1/4 wavelength at 2.25Khz is approximately 38mm, or 1.5" ---which would mean its impossible to create 1" tweeter + 6.5" woofer design that does not exhibit lobing.
Maybe the simple explanation is that 1/4 wavelength or less is ideal (as you stated above), and up to 1 wavelength is acceptable, but anything greater than that would result in sever lobing. .
per Danley. This criteria makes for such a low XO that only WAWs (or Coax, but they would need appropriate time delay somewhere) can achieve it, so you see various other compromized criteria are used.
dave
You didn't misread as the 1 wavelength or less is common in literature. A quote from the Loudspeaker Design Cookbook 6th Ed by Vance Dickason:
... separate the two drivers by a distance equal to, or less than, the wavelength of the highest frequency the driver will produce. For practical purposes, that distance is equal to one wavelength at the crossover frequency.
The ¼ wavelength distance pops up in discussions, especially on the WAW and FAST speakers which helps to promote that style of loudspeaker.
I haven't seen it in the literature I've read in the past but I've only read a small amount on the subject.
1/4 wavelength separation ensures that even at 90˚ off axis (the worst case example), the two drivers will still sum to an amplitude at -3dB relative to the on-axis response, which is still acceptable.
By the time you get to 1/2 wavelength driver separation, there would be a null created 90˚ off axis (assuming both drivers have the same off-axis response), with varying degrees of attenuation between 0˚ and 90˚ from the axis. This isn't likely to be a serious or even noticeable issue for most situations but it is advisable to avoid such design flaws if possible.
Using a 1 wavelength separation or more is asking for trouble as the nulls and attenuation quickly become an issue within 45˚ of the axis. This can degrade the direct sound quality as well as the reflected sound. See diagram for a visualization of nulls and amplitude variation:
Attachments
It's kind of artificially created problem and largely dependent of a room acoustics. CTC is important because of vertical directivity, but going as far as demanding 1/4 wavelength is pointless. When you look at the fact that you listen to the music in the spot that is vertically spread within +/-0,5m (even this is more than a real life situations) at the 3m distance then you get a window of +/- 10deg, and this is roughly how big vertical derectivity you need.
Seems to me that this effect really only helps if you are at a position where sound is coming from only one of the two sources, instead of both.
It is easier to talk about what happens, not as easy to describe what it might sound like. jamesblonde is right in saying that 1/4wl produces a lobe, with a beamwidth of 180degrees. This therefore affects power.
It should be taken case by case because it is based on a compromise to begin with. To put it in simpler terms, say you have a 1/2wl separation which produces a null at 90degrees off axis, but neither of the drivers is directing sound to 90degrees near the cross.
Well said James. Your picture reminds me of this one from Taylor:
https://p10hifi.net/TLS/downloads/taylor-line-array.pdf
dave