Russellc
Addicted Member
I understand about what Wayne was doing. What I cant get a hold on is how,
from a response plot shown, one determines the forward axis.
russellc
Zilch's AK Design Collaborative - Econowave Speaker
- Thread starter Zilch
- Start date
I understand about what Wayne was doing. What I cant get a hold on is how,
from a response plot shown, one determines the forward axis.
russellc
Russellc
Addicted Member
Through the crossover region, both woofer and waveguide are playing the same frequencies. That's two sources with the same program, and they add to and interfere with each other to varying degrees depending upon the relative distance of the listener (or measurement microphone) from each of them. When the distance to each is the same, and both are in phase, they sum maximally, and conversely, when they are out of phase, they cancel maximally. We design such that the listener listens from where the two sources are in-phase and summing maximally.
The listener and the two sources in a single speaker, the woofer and waveguide, form an imaginary triangle in the vertical plane (on-axis horizontally) when they are mounted one above the other. When the distance of the listener to the acoustic center of each of them is the same, that triangle is isosceles, and the summation is maximal. Viewed from the side, it is easily seen that, as the listener moves up and down from the point midway between the drivers one side of that triangle (representing the distance of the source to the listener) becomes longer, while the other becomes shorter, the distances are no longer identical, and the summation is less perfect. In waveguide speakers, the vertical distance between the sources themselves (the base of the imaginary triangle) is large, and small movements of the listener (or microphone) up or down at typical listening distances create large differences in the relative distance to each of the sources.
What we call the "forward vertical axis" of the speak as line between the apex of that triangle and the midpoint between the acoustic centers of the two sources vertically, the midpoint of the base of the triangle, when the two sides of that triangle, representing the distances from each source to the listener, are equal, and the summation, maximal.
While it is possible to determine the exact location of the acoustic centers of the two sources via measurements, it may be easily seen that, if the upper one is displaced further back than the lower one, as is typically the case with a woofer and waveguide, the triangle "points" upward from the midpoint between them. With a typical woofer and tweeter, the situation is opposite, with the woofer being behind, and the triangle pointing downward. In both cases, we can manipulate the virtual locations of the acoustic centers via delay, either actual or phase delay, or both, to "aim" the triangle wherever we want, typically such that the apex of the imaginary triangle, the forward vertical axis, "points" the listening height of the listener.
For any given system under design and measurement, how might we easily determine the actual alignment and get it "right?" The determination part is easy, actually: if we invert the polarity of one driver or the other, instead of maximally summing at the apex of the triangle, the forward axis, the drivers maximally cancel there, and a deep notch appears in the amplitude response measurement of the system. By moving the microphone up and down and taking repeated measurements, we can locate exactly where that notch is deepest within 1/2 the diameter of the microphone aperture, typical, ~1/8". Knowing the height of the microphone relative to the midpoint between the drivers and the horizontal measuring distance, we can calculate the displacement angle of the forward axis relative to normal (perpendicular to the baffle) in pursuit of our design target.
Cool, huh?
THANK-YOU!:ntwrthy:
:ntwrthy:That is about the most concise writing describing every aspect of the puzzle I couldn't unwind. Outstanding explanation! Every question answered,
I hope this cleared a few things up for others as well. Thanks for going to the trouble to expound on this on a level I could understand clearly.
Much appreciated,
Russellc
Hi Pat,
Just a couple of notes. First, there is not much BSC in this design, as I intend to put these fairly close to the wall. Second, this worked great for my 123A-3, but when I tried my 2213h, there was a dip of several db near the crossover frequency. Not all woofers, even of the same type are created equal, so you may have adjust slightly on the woofer inductor and cap to suit your particular situation.
Markus
4343
Super Member
2.5 Way?
2.5 way has been done. MTM really only works well with small mid-woofers and with small tweeters. DC300=Not Small, Waveguide=Not Small...
2.5 way is really simple, you just need to calculate the highest frequency that the woofers are able to mutually couple and filter everything above that frequency out of the woofer farthest away from the waveguide (I stayed an octave down from there to be safe). An inductor feeding the woofer-that-needs-filtering works great if your woofers are in parallel, use a capacitor across the terminals of the woofer-that-needs-filtering if you need to run them in series. I was able to use 4 ohm car speakers in series on two pairs doing this. The trick is deciding what impedance the crossover will see at the crossover frequency. With two 8 ohm speakers in parallel, the load at low frequencies is 4 ohms, but when the frequency rises above the cutoff point of the inductor it will be 8 ohms. (This is a recipe for great bass, as most amps will put out more power into a 4 ohm load.):thmbsp: For this configuration, you would use the 8 ohm version of the EconoWave high pass.
With two 4 ohm woofers in series, just the opposite happens, with the load starting out at 8 ohms, and falling to 4 ohms before the crossover frequency. Thus I used the 4 ohm version of the EW high pass.
I've had pretty good luck just setting the bass levels with the tone controls so far.
Obviously, if you put two 4 ohm woofers in parallel, your amp will need to be able to handle a 2 ohm load, so I decided to go with series for that reason.
Dang, I gotta go turn the 2.5Wayves on for a while!
I have a budget of approximately $300 to build some Econo-Waveguides. These will be used as PA speakers (though only in very small rooms - the sort of things people use those rubbishy little Bose PA speakers for), and need to be able to produce some fairly impressive SPL. Can anyone recommend a woofer/xover to me? Or perhaps sell me some parts?
EDIT:
Has anyone considered either a 3.5-way or MTM configuration using two DC300s? It seems a shame to spend more on your xover than you are on your woofers.
2.5 way has been done. MTM really only works well with small mid-woofers and with small tweeters. DC300=Not Small, Waveguide=Not Small...
2.5 way is really simple, you just need to calculate the highest frequency that the woofers are able to mutually couple and filter everything above that frequency out of the woofer farthest away from the waveguide (I stayed an octave down from there to be safe). An inductor feeding the woofer-that-needs-filtering works great if your woofers are in parallel, use a capacitor across the terminals of the woofer-that-needs-filtering if you need to run them in series. I was able to use 4 ohm car speakers in series on two pairs doing this. The trick is deciding what impedance the crossover will see at the crossover frequency. With two 8 ohm speakers in parallel, the load at low frequencies is 4 ohms, but when the frequency rises above the cutoff point of the inductor it will be 8 ohms. (This is a recipe for great bass, as most amps will put out more power into a 4 ohm load.):thmbsp: For this configuration, you would use the 8 ohm version of the EconoWave high pass.
With two 4 ohm woofers in series, just the opposite happens, with the load starting out at 8 ohms, and falling to 4 ohms before the crossover frequency. Thus I used the 4 ohm version of the EW high pass.
I've had pretty good luck just setting the bass levels with the tone controls so far.
Obviously, if you put two 4 ohm woofers in parallel, your amp will need to be able to handle a 2 ohm load, so I decided to go with series for that reason.
Dang, I gotta go turn the 2.5Wayves on for a while!
Last edited:
Put in a box modeling program and see what the SPL limitation is at what frequency.
If you're going to add a second woofer, do it in "0.5" mode. I'd just build a second identical cab for that, and stack them.
You said PA in small venues better than crappy speakers. eWave "Standard" will get you that adequately. Now you want 6 dB more?
As shown, F3 is 43.94 Hz, and F6, 36.05 Hz, where it'll do 104.8 dB RMS with 31.2 watts, 1.64 mm excursion, and does not run out of that at its rated 80W until 23.1 Hz by my model. The knee of the rolloff occurs at tuning, 30 Hz.
Does not compute.... :tongue:
Last edited:
Spasticteapot
Nuclear Liberal
I'm afraid I'm stuck without any software for excursion calculation at the moment - WINE is a bit unhappy on my Macintosh. A 2.5-way with a simple electronic lowpass seems like an ideal solution for boosting the low octave a bit, and the additional cost is minimal.
You don't get 6dB. You get 3. 3 is a doubling of energy. You only get 6dB of voltage sensitivity when paralleling 2 drivers. That's voltage sensitivity, efficiency only goes up 3dB just like max output.
"Mutual Coupling" is a fancy way of saying "Double the cone area", it's just saying that within the frequency range where wavelengths are long relative to the driver spacing, they will reliably sum as a single coherent source with double the surface area.
Best said I've heard was paraphrased as "It's not something good that happens at L.F., it's the absence of the bad (lobing cancellations) thing that happens at H.F."
You get 3dB of efficiency from doubling the surface area (and cabinet volume to accommodate the same F3), no matter how you wire them. The voltage sensitivity will be equal to a single driver but a 16 ohm load drawing half the current (3dB less power input for a given SPL), if you series them. The voltage sensitivity will be 6dB higher than a single driver if paralleled, but 3dB of that is due to the doubled current input (3dB more power input for a 6dB higher SPL).
"Mutual Coupling" is unfortunately misleading. You get people talking about the compressibility of air and whatnot. I prefer just to think of it in terms of box size and drivers.
Model enough drivers and you see that for 3dB more in a similar alignment, you need 2x the box. Doesn't really matter if that's made up of 2 motors and cones, it's just a compound driver, until the wavelengths are short enough that lobing is an issue.
The "Efficiency Sensitivity" discussion takes place over and over and over and even the experienced misunderstand and/or misstate it. I think a lot of that is terminology related, which is why I like to pin it down a little bit- "Voltage Sensitivity", "Twice the current" "Half the current" etc. When you use the right words, it all "clicks".
Last edited:
See Fig. 9 and related text here:
http://www.harman.com/EN-US/OurComp...p/Documents/Scientific Publications/13000.pdf
http://www.harman.com/EN-US/OurComp...p/Documents/Scientific Publications/13000.pdf
dnewma04
The Healer
Seems to corroborate what badman is saying, unless I am reading something incorrectly. 3dB for 2 drivers, 6dB for 4 drivers, it would be 9dB for 8 drivers. If the nominal load drops from 8 to 4, you gain another 3dB assuming the amp doubles down and you lose 3dB if it rises from 8 to 16 if the amp produces half the power into 16 ohms.
Read the first paragraph under "Mutual Coupling" there again. You get 3dB more with the total power (1W) held constant (the voltage dropped from 2.83V to 2V). Hold the voltage constant at 2.83V (for a parallel connection) and you get 3 dB more from giving the pair 2W, instead.... :yes:
Last edited:
Assuming them to be driven by a voltage source (read: big tube or SS, low output impedance) amp, and the drivers in parallel in twice the box volume as a single driver and the range in which the .5 driver is operating below 1/4 W/L (or 1/2 depending on your definition) of the CTC spacing, 6dB.
You could also have them be 0, and even below 0, from a voltage sensitivity standpoint, if you wired them in series. Or raise the F3/Tuning if you don't double the box volume.