RS-P WWMTM 3 Way underway ( pic heavy )

[Juhazi]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

I believe that this will be a value killer! And absolute sound quality killer too!

Double bass and mid will give minimal distortion, and lowish xo from mids to tweeter will give good vertical polars too. Smooth horizontals are guaranteed! Low position of woofers helps to minimize floor bounce. Bass to mid xo has lots of Fx possibilites (around baffle step), LR2 is most likely the best slope. Double drivers can be parallel (if 8R) or serial.

[jcandy]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

I believe that this will be a value killer! And absolute sound quality killer too!

Double bass and mid will give minimal distortion, and lowish xo from mids to tweeter will give good vertical polars too. Smooth horizontals are guaranteed! Low position of woofers helps to minimize floor bounce. Bass to mid xo has lots of Fx possibilites (around baffle step), LR2 is most likely the best slope. Double drivers can be parallel (if 8R) or serial.

Agree with everything, except perhaps the comment about low woofer positioning. While you are certainly on the side of common wisdom with respect to this issue, in reality its more of a mixed bag. It is true that the frequency of the floor-bounce null increases as the woofer height decreases. And it is also true that the low-frequency gain increases. These effects can be visualized simply by placing an image driver at height -h for every real driver at position h (in reality, the reflection coefficient is a function of the angle of incidence, so the accuracy of this trick is limited). Here, the floor is h=0. The wrinkle of course is the polar pattern at the woofer-mid crossover frequency. When the woofers are close to the floor, the strength of the image drivers approaches that of the real drivers, and you may wind up with a very erratic polar response at the listening position.

So, its a tradeoff, which can be perhaps more concisely stated by considering a simple W-M-T 3-way. By lowering W you bring it closer to its image and thereby enlarge the W-W lobe (good). But by lowering W you also move it farther from M thereby narrowing the W-M lobe (bad). The low-woofer design is also more difficult to model for various reasons. This is why I opted to maximize woofer height in the related design posted earlier in this thread.

[Jeff B.]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

Agree with everything, except perhaps the comment about low woofer positioning. While you are certainly on the side of common wisdom with respect to this issue, in reality its more of a mixed bag. It is true that the frequency of the floor-bounce null increases as the woofer height decreases. And it is also true that the low-frequency gain increases. These effects can be visualized simply by placing an image driver at height -h for every real driver at position h (in reality, the reflection coefficient is a function of the angle of incidence, so the accuracy of this trick is limited). Here, the floor is h=0. The wrinkle of course is the polar pattern at the woofer-mid crossover frequency. When the woofers are close to the floor, the strength of the image drivers approaches that of the real drivers, and you may wind up with a very erratic polar response at the listening position.

So, its a tradeoff, which can be perhaps more concisely stated by considering a simple W-M-T 3-way. By lowering W you bring it closer to its image and thereby enlarge the W-W lobe (good). But by lowering W you also move it farther from M thereby narrowing the W-M lobe (bad). The low-woofer design is also more difficult to model for various reasons. This is why I opted to maximize woofer height in the related design posted earlier in this thread.

I know you know this, Jeff, but I think it bears clarification. Lobing is a function of driver separation and wavelength. If the crossover between the woofer and the midrange is at 300 Hz then that corresponds to a wavelength of 45 inches. For almost any tower speaker, even one with the woofer near the floor, it is likely to be separated by much less distance than this. On the other hand, a 2000 Hz crossover will be dealing with a wavelength of about 6.7" and this will be pretty close to the separation between a typical midwoofer and a dome tweeter. We tend to usually feel pretty comfortable with the 2 kHz crossover though. I really don't think lobing concerns are a good reason to give up the sensitivity gain from placing the woofer near the floor boundary.

Jeff B.

[jcandy]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

Lobing is a function of driver separation and wavelength. If the crossover between the woofer and the midrange is at 300 Hz then that corresponds to a wavelength of 45 inches.

Right, but the lobe occurs when the path difference is half a wavelength. To be concrete, let's use the present design as a reference, and consider the interaction between (1) a woofer at the location of the lower woofer, and (2) a midrange located where the tweeter is (i.e., the average distance of the two midranges). The resulting lobe in shown in the image below. I assume LR2 at 350Hz. As you can see, at a listening distance of 1m, there is a wide null about 25cm above the tweeter axis.

For almost any tower speaker, even one with the woofer near the floor, it is likely to be separated by much less distance than this.

Well, sure, 45" is an extreme separation. But in the present example I have taken the distance to be 75cm (29"), which is more reasonable.

On the other hand, a 2000 Hz crossover will be dealing with a wavelength of about 6.7" and this will be pretty close to the separation between a typical midwoofer and a dome tweeter. We tend to usually feel pretty comfortable with the 2 kHz crossover though.

It goes without saying that I would not be entirely happy with 6.7" between midrange and tweeter crossed at 2kHz.

I really don't think lobing concerns are a good reason to give up the sensitivity gain from placing the woofer near the floor boundary.

Well, this is another can of worms. The gain depends on location and frequency. At very low frequency, ignoring room modes, moving a woofer closer to the ground (say from 24" to 12") results in insignificant gain. Do you agree?

[jcandy]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

Here is an example of the gain versus frequency assuming a perfect reflection from the floor and a point source for the woofer. The listening position is at a distance of 1m from the baffle. In the plot I scan the height of the woofer from floor to ear height (1m). Each curve corresponds to a different frequency. At low frequency there is actually a very small loss as the woofer approaches the floor.

[Jeff B.]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

Right, but the lobe occurs when the path difference is half a wavelength. To be concrete, let's use the present design as a reference, and consider the interaction between (1) a woofer at the location of the lower woofer, and (2) a midrange located where the tweeter is (i.e., the average distance of the two midranges). The resulting lobe in shown in the image below. I assume LR2 at 350Hz. As you can see, at a listening distance of 1m, there is a wide null about 25cm above the tweeter axis.

[ATTACH=CONFIG]49365[/ATTACH]


Well, sure, 45" is an extreme separation. But in the present example I have taken the distance to be 75cm (29"), which is more reasonable.


It goes without saying that I would not be entirely happy with 6.7" between midrange and tweeter crossed at 2kHz.


Well, this is another can of worms. The gain depends on location and frequency. At very low frequency, ignoring room modes, moving a woofer closer to the ground (say from 24" to 12") results in insignificant gain. Do you agree?

Yes, I do agree. But it's more important in the crossover range where it can help to reduce, or even eliminate the necessity of baffle step compensation. This is where the overall gain in sensitivity comes in. This is also part of what Roy Allison was trying to accomplish in his designs. Sure there are trade-offs. Everything in speaker design is a series of trade-offs one way or another. It becomes a matter of the priorities for that design.

[jcandy]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

Yes, I do agree. But it's more important in the crossover range where it can help to reduce, or even eliminate the necessity of baffle step compensation. This is where the overall gain in sensitivity comes in. This is also part of what Roy Allison was trying to accomplish in his designs. Sure there are trade-offs. Everything in speaker design is a series of trade-offs one way or another. It becomes a matter of the priorities for that design.

Ah, right. Now I see what you mean by sensitivity gain -- in connection with mitigating the baffle step. Sorry for the diversion.

[filmslayer]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

now that we got that straight , how does my woofer placement look ?

[Jeff B.]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

now that we got that straight , how does my woofer placement look ?

It looks fine to me.

[filmslayer]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

It looks fine to me.

PHEW ! thanks Jeff ... i gotta admit that my brain is about mush . i've been cramming everything i can about building speakers going on 9 months now and that last exchange pushed me close to the edge ... lol

[jcandy]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

PHEW ! thanks Jeff ... i gotta admit that my brain is about mush . i've been cramming everything i can about building speakers going on 9 months now and that last exchange pushed me close to the edge ... lol

Its worth thinking about. The point is that to minimize lobing due to the large distance between woofers and mids, you'll want to keep the woofer-mid crossover low. There is also the issue of the crossover and using the proximity of the woofers to the ground to mitigate the baffle step. Curious to see how that goes. Whatever happens, you are going to have some towers that can play loud loud loud.

[filmslayer]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

i completely understand and learned quite a bit J from that exchange ... and while we're at it can i get some tips on measuring the woofers ?

i plan on getting all the measuring done this weekend ...

[jcandy]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

i completely understand and learned quite a bit J from that exchange ... and while we're at it can i get some tips on measuring the woofers ?

i plan on getting all the measuring done this weekend ...

While the approach to designing a speaker with flat anechoic response is fairly standard, the approach to designing one that includes the woofer-floor interaction is less standard. Perhaps Jeff B. could explain what his preferred method is.

[Chris Roemer]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

If you measure indoors using something like ARTA, with about a 3.3 msec "window", you'll only get useful data down to about 300Hz.
You can then measure a woofer "nearfield" (like, at about 1/4" from the dustcap) and your curve will be OK, but you won't know how to relate that data (SPL-wise) to your windowed data taken at 1m. You COULD try a box sim program (like WinISD) to get your sensitivity in the ballpark, but THEN you also have to add in a measurement from your port (OR, just use an output curve from your box modeling software that includes the driver and vent). You really SHOULD plan to cross near 300Hz or so, but accounting for that using your own measurements is not very "cookbook".
You'll also have room interactions to consider.

[filmslayer]

Re: RS-P WWMTM 3 Way underway ( pic heavy )

thanks Chris , i believe Josh already did the in box modeling when he tweaked the files that you sent me . we were more concerned with in box response of the 125's and their inherent dip and while i was getting those measurements thought it might be fun to see what the woofers looked like .

i keep forgetting my screw-gun at work and haven't started measuring yet , i needed a couple days digestion and recharging anyway . i'm ready to get chewing again this weekend . i'm really starting to see how ambitious this project is for a noob like myself , but we'll get it !