Small 3-Way Tower, Front and Rear Woofers?

[superspeeder]

I was reading through another post here regarding a very $exy thin tower build by devinkato:

http://techtalk.parts-express.com/showthread.php?234584-3-Way-Active-High-WAF-Hi-Vi-Fas

There is a really good discussion there regarding driver diameter and directivity. Not wanting to hijack the thread I decided to start a new one on a semi-related topic.

A project idea that's been on my mind for a long time is a dual small woofer 3-way design, but mounting the extra woofer on the rear of the cabinet instead of the front.

In my mind this offers a few bonuses, the biggest of which would be low frequency cabinet vibrations cancelling out (the cones of the front and rear mounted woofers moving in opposite directions).

Directivity is my main concern with this idea, so my question here for the more experienced is:

How high would I expect to play a pair of ~4" woofers if they were mounted front and back on a 6x6 tower, and still be able to blend them smoothly with a small-ish midrange?

Having some idea what this frequency is would give me a much better idea of what to expect out of a small midrange. My naive mind thinks this frequency would not be much higher than 150-200 Hz, but that's just a gut feeling not based on any math.

The other option would be to treat the extra woofer at the back as a ".5", but if already dealing with a 3-way system it starts looking awful complex. Input appreciated!

[superspeeder]

Re: Small 3-Way Tower, Front and Rear Woofers?

Over 100 views... Bueller?

[ryanbouma]

Re: Small 3-Way Tower, Front and Rear Woofers?

DARN!!! Just lost all my post. Here we go again.

I've given this quite a bit of thought, and have decided it's a bad idea. It's a bipole. The rear woofer acts like the x.5 but without any electrical influence.

Take a look at this polar I made of a 2-way using a 6" woofer on a 9" wide tower (iirc):



You can see at 150hz (blue) the spl remains constant all the way to 135d. At 500hz (green) it's attenuated a little. And at 700hz (yellow) it's fully 2pi and radiating pretty much only 0 to 90d and not behind the baffle. I'm sure you're familiar with what's happening here. Baffle step.

Well, what I didn't show, because it didn't concern me, but would have helped this explanation, is that at 180d you start to see another lobe form. Like this: http://www.qscaudio.com/support/educ...1-5_polar3.png

Now the issue to me, with what you're proposing here, is there is going to be a funny lobe like that radiating forward from the rear woofer. This is undesirable, and I think it's obvious why. Back to my polar, at 150hz, you'll get a fairly uniform spl wrap. But then again at 500hz, you'll get this lobe. Then you'll get only 2pi radiation above that.

Because of this, and also comb filtering between the two drivers as frequency increases, I would say your cut off point would have to be fairly low, and enclosure size dependant. But as you already mentioned, likely around 150 to 200hz. Maybe 300hz for that size of enclosure.

[ReissM]

Re: Small 3-Way Tower, Front and Rear Woofers?

...A project idea that's been on my mind for a long time is a dual small woofer 3-way design, but mounting the extra woofer on the rear of the cabinet instead of the front. In my mind this offers a few bonuses, the biggest of which would be low frequency cabinet vibrations cancelling out (the cones of the front and rear mounted woofers moving in opposite directions).

With a heavily braced cabinet, these low freq cabinet vibrations won't be an issue for you. (You're talking about using a 4" woofer, right?) Go with 3/4" panels or thicker, and brace them well... then it won't make much of a difference if you mount the two woofers to opposite sides or not.

...Directivity is my main concern with this idea, so my question here for the more experienced is:

How high would I expect to play a pair of ~4" woofers if they were mounted front and back on a 6x6 tower, and still be able to blend them smoothly with a small-ish midrange?

Everybody has their opinions on this type of thing. You're asking for recommendations on the freq. your 4" woofer should hand off to the midrange driver, right? My personal opinion is to keep the crossover frequencies as low as possible. Remember that at the crossover frequency, both drivers are reproducing sound. That means you'll have path length differences from each driver to your ears. So in order to minimize the problems created by path-length differences, you need to keep the wavelengths "long" relative to the path length distance difference.

Example:
Let's say that you measure the path length that each driver takes on its way to your ears (or a fixed microphone) and you arrive at the conclusion that the sound from the woofer travels 3.5 inches further than the sound from the midrange driver. At some frequency the woofer and midrange will be acoustically out of phase due to their geometric configuration, and will acoustically cancel each other. So what frequency has a half-wavelength that is 3.5"? About 1940 Hz. Ok, what does that mean? It means you'll have a huge frequency response dip at 1940 Hz (but only at that particular microphone location).

But here's the real problem... As you can imagine, any location that is not equidistant from both drivers will experience this path-length cancelation problem. The solution is to simply pick a crossover frequency that is much lower than the calculated half wavelength. (1940Hz in this example) For this example, how about using 500Hz as the crossover point? 500Hz has a wavelength of about 27 inches. That makes the 3.5 inch path-length difference no big deal.


My suggestions for your particular case...
For woofer-to-midrange: Try about 200Hz.
For midrange-to-tweeter: Try 1700 or 1800Hz (if your tweeter can handle the power, and has a low Fs down around 900 Hz)


Just out of curiosity, years ago I created an Excel spreadsheet to do all the math for me based on the geometry of the two drivers and the listening position. Then I went ahead and predicted the frequencies that would be associated with the "path-length differences" for several off-axis locations. Later, when I took some acoustic measurements, the math was only off by a few Hz. (e.g. predicted null at 1920Hz, actual null at 1927Hz) By simply picking a lower crossover frequency, (longer wavelengths) ... this problem can be significantly reduced.

Here's a snapshot of the recently updated spreadsheet that shows the geometry and some typical numbers. Note that problems also occur at 1.5 * the wavelength in addition to 0.5 * wavelength.

[ReissM]

Re: Small 3-Way Tower, Front and Rear Woofers?

The diagram in my previous post can be rotated such that the two drivers are horizontally configured instead of vertically mounted. The math doesn't change. In this configuration, it reminds me of the typical center channel configuration with two horizontally mounted woofers.

I enter some basic dimensions and the spreadsheet crunches the numbers to give me the off axis angles and the path-length problem frequencies. I use that information to pick my center channel crossover frequency.

e.g.- if you know the two woofers will create cancelation 30 degrees off axis, at 3000 Hz, due to their physical separation... well... don't let those two woofers play up that high. Cross them over lower... maybe at 1700Hz. Curt Campbell uses this strategy in his Aviatrix center channel design to achieve excellent off-axis response even though the two woofers are horizontally positioned. (I don't recall Curt's crossover frequency, but I know it's fairly low for this exact reason.)

The tradeoff is that you'll need to buy a tweeter than can handle the extra power. Also, keep in mind the general rule of thumb is to crossover a tweeter at double the tweeter's Fs, or above. (we all break the rules once in a while, but do your best to find a low Fs tweeter if you're designing a horizontally configured center channel.)

In addition to choosing low XO freqs, you're going to want to mount the two woofers as close as physically possible while still squeezing the tweeter in there somewhere. The closer you can mount the two woofers, the fewer the problems you'll have with path-length cancellation issues.

These concepts can be applied to many different speaker designs, not just towers or center channel speakers.

[superspeeder]

Re: Small 3-Way Tower, Front and Rear Woofers?

Wow, some great information here!

WRT this strange lobe showing up at 180, do you think this would still happen using another driver on the rear? I'm under the impression that the output "wave" of two drivers operating at the same frequency will merge together, not act independently?

If this is the case, the merging of outputs would happen at the sides of the enclosures, at 90 and 270. Am I correct in thinking this way?

WRT the Excel sheet, that's good stuff! It clearly shows some very interesting phase interactions.

WRT the driver separation vs freq, a 500 Hz signal has a wave of 27", but would a 3.5" driver delta introduce some sort of group delay, distortion or something similar?

Sorry I didn't take the time to break this down into a multi quote response, I'm doing this from my cell phone...