or Connect
AVS › AVS Forum › Audio › DIY Speakers and Subs › A 3 way 99db multi configurable SEOS design
New Posts  All Forums:Forum Nav:

A 3 way 99db multi configurable SEOS design - Page 2

post #31 of 670
Quote:
Originally Posted by tuxedocivic View Post

Thanks!

No abondoning going on. The polars were very tricky to get right. A LOT of thought went into this design. But it paid off because the polars are very good for a 12" wide speaker. I am going to offer a single WOOFER design. So it'll be TMW. Saves $80 and 14" height.
Thanks. That's a ground plane measurement, so you can see it hits an F3 of about 80 or 90hz. This surprised me when I measured because it sounds better than that in room. Then I listened in my shop and they were very bass shy. So I decided a shorter port will bring up the tuning and give about a 70hz F3. Right now the tuning is around 45hz. To low for this driver.
Ya, the T/S parameter sensitivity is about 89db each. But the actual output is more like 91db in the pass band I'm using. So it is more like 97db/2.83V/m not the 98db I said. Still, where the 99db comes from is the bandpass gain. Inquisitive question biggrin.gif
Tux are the speakers still "bass shy" did you settle in on a final tuning?
post #32 of 670
Thread Starter 
I haven't shortened the ports yet. There's really no need to if using subs, but I'll still tune them to about 55hz in the end I think.
post #33 of 670
Thread Starter 
Ok, now for the hard bit to explain. I definitely need to explain this a bit. Here's the side by side woofer portion for the center style.



This chart shows the general woofer to mid XO region.

Now, first off note that there's about a 40 degree listening window there. So that's pretty good. I think that works for most people. Second note that with most center channels, side wall reflections are far away. Also note, you never see a polar like this for a center channel MTM speaker. Because frankly, they're very very very bad, all the time. By comparison, this is really good and part of why I'm presenting the data this way. It's the easiest way to see nulls. Ok there's a few things going for us.

What's bad about it, well, there's some pretty deep nulls around 45 degrees. I've proposed to get them out closer to 60 degrees by moving the woofers closer to the mids. Originally I planned about a 500 or 600hz cross over. In the end there was a lot of overlap in the XO and it was around 750hz. This was done for power handling, bandpass gain, and lower/cheaper XO parts count. So in my plan, I spaced the woofers apart a little more than needed to make constructability easier and hoping to match directivity better. With the finished product, the woofers will be as tight as possible, and then hopefully I can remeasure and see the improvement.
post #34 of 670
Awesome work Tux. I'm pretty excited about this design.

I haven't looked at many (maybe any) plots of this type for a sideways MTM, but is the early slope off axis to be expected? It looks like the -6dB point is around 15 degrees. I sort of expected it to stay flatter out to say 35 and then just take a nose dive. Is that effect limited to the crossover region that you've plotted here? Is it a driver interaction thing that will be pushed out some when you bring the spacing in tighter?
post #35 of 670
Thread Starter 
You have it exactly right pitviper. Bringing the drivers a little closer will improve things. I feel like the current configuration is closer to -6db at 20 degrees, but ya, around there. The nulls are expected, because the woofers cancel each other out once out of phase. The way to fix this is to get them closer, or lower the cross over. I could have lowered the cross over, but as we know, everything is a compromise, and this is one I made.

It is limited to the XO region. Above 1200hz you can look to my other polar to see the kind of polar response you can expect.

The reason you don't see this kind of plot with a typical MTM center is because it would look really bad.
post #36 of 670
Quote:
Originally Posted by tuxedocivic View Post


The reason you don't see this kind of plot with a typical MTM center is because it would look really bad.

So says you!
biggrin.gif
post #37 of 670
Thread Starter 
I wanted to find some examples of MTM centers. I found this from zaph: http://www.zaphaudio.com/BAMTM-polar-1000-2000-100step-red-blue.gif

That's a 1450hz cross over with a dome tweeter. Very little power handling to achieve the wide nulls he did. And they're still not as wide as my speaker. Most all other MTM centers will be far far worse. Zaph does endorse that speaker for center channel use also.

I've looked at the models, and if Erich can squeez those woofers together even another 2 inches, that should open the -6db point to 30 degrees.
post #38 of 670
I will be able to get the woofers closer once the flat packs are cut. Samples for the plastic waveguide are suppose to be here next week so I can then start doing the CNC design work.
post #39 of 670
Quote:
Originally Posted by tuxedocivic View Post

You have it exactly right pitviper. Bringing the drivers a little closer will improve things. I feel like the current configuration is closer to -6db at 20 degrees, but ya, around there. The nulls are expected, because the woofers cancel each other out once out of phase. The way to fix this is to get them closer, or lower the cross over. I could have lowered the cross over, but as we know, everything is a compromise, and this is one I made.

It is limited to the XO region. Above 1200hz you can look to my other polar to see the kind of polar response you can expect.

The reason you don't see this kind of plot with a typical MTM center is because it would look really bad.


Tux, can you explain how bringing the drivers closer together will improve things? Should the drivers always be as close as possible with most typical builds? When is this not to your advantage?
Another concept that I am having trouble grasping is how phase works. How do the woofers become out of phase at some point? How and why does that happen? When two or more woofers are used in
a speaker, will the woofers always cancel each other out when out of phase? How do the woofers cancel each other out?

I appreciate you taking the time to answer my questions!
post #40 of 670
Thread Starter 
Hey Marty. Pretty big topic to cover in a forum post. My physics 101 prof spent probably 3 or 4 lectures on it. Here's a summary: https://en.wikipedia.org/wiki/Interference_(wave_propagation)

The link is more about light, but the same physics apply. Basically, it's a function of wavelength (frequency dependant), phase, and distance between sources (two woofers, two mids). Because it's a function of distance, if we can change distance, we can change the cancellation. By shortening the distance, we widen the cancellation nulls.

Usually closer is better. There are exceptions where wider is better. I thought this might have been one of those times, but in the end I crossed a little higher than expected for other benefits.
post #41 of 670
Thread Starter 
Quote:
Originally Posted by Erich H View Post

I will be able to get the woofers closer once the flat packs are cut. Samples for the plastic waveguide are suppose to be here next week so I can then start doing the CNC design work.

Hopefully this weekend I can build a new baffle and test out a closer spacing.
post #42 of 670
Quote:
Originally Posted by tuxedocivic View Post

Hey Marty. Pretty big topic to cover in a forum post. My physics 101 prof spent probably 3 or 4 lectures on it. Here's a summary: https://en.wikipedia.org/wiki/Interference_(wave_propagation)

The link is more about light, but the same physics apply. Basically, it's a function of wavelength (frequency dependant), phase, and distance between sources (two woofers, two mids). Because it's a function of distance, if we can change distance, we can change the cancellation. By shortening the distance, we widen the cancellation nulls.

Usually closer is better. There are exceptions where wider is better. I thought this might have been one of those times, but in the end I crossed a little higher than expected for other benefits.

Good info and details for DIYers here. Most brush off the reality of executing a passive crossover at a lower frequency. Actual implementation can be tricky when trying to balance multiple priorities which is why I've suggested many times for DIYers to do a passive Mid-High and actively cross to woofers. Thank you also for making a good demonstration of why I have the crossover on my big Catalyst speakers down in the 200-250Hz range.

It should be noted that the crossover and overlap of the chosen type also affects the resulting pattern, not just the spacing. Phasing and overlap add another source in the middle which can help... or sometime even hurt the cause.
post #43 of 670
Thread Starter 
Quote:
Originally Posted by Mark Seaton View Post

which is why I've suggested many times for DIYers to do a passive Mid-High and actively cross to woofers.

For sure. Which reminds me, I have an active/passive version of this as well which crosses the woofers a little lower and harder.
post #44 of 670
Quote:
Originally Posted by tuxedocivic View Post

Hey Marty. Pretty big topic to cover in a forum post. My physics 101 prof spent probably 3 or 4 lectures on it. Here's a summary: https://en.wikipedia.org/wiki/Interference_(wave_propagation)

The link is more about light, but the same physics apply. Basically, it's a function of wavelength (frequency dependant), phase, and distance between sources (two woofers, two mids). Because it's a function of distance, if we can change distance, we can change the cancellation. By shortening the distance, we widen the cancellation nulls.

Usually closer is better. There are exceptions where wider is better. I thought this might have been one of those times, but in the end I crossed a little higher than expected for other benefits.

Thanks for the reply, Tux.I have a couple of more basic questions. No need to get too deep in your explanation..

How do you know when and what frequencies the woofers will be in phase versus out of phase?

The thing that messes me up is the fact that I just do not understand phase. (ie what is phase? How does it work? Can it be predicted? ect..)

I would imagine that going with an active crossover would give you the freedom to adjust the crossover in order to correct the phase issues that are cancelling out the bass frequencies between the two woofers, which imho would be much easier than constructing a new baffle in order to bring the woofers closer together.
post #45 of 670
Thread Starter 
Actually active wouldn't fix the problem. The problem is "hard wired" into the speaker like the directivity of a horn. What active can do is lower the cross over and make it steeper which can mitigate and even eliminate the problem. Like what mark said he did in the Catalysts.

This of it like this. Standing directly in front of the speaker, both woofers are exactly the same distance from you. Now move to the left. Now the left woofer is closer and the right woofer is further. Ok now imagine the woofers are shooting out waves, cause they are. The wave coming from the right woofer will be lagging the wave from the left woofer the distance that they're apart from you. If you look at that Wikipedia link, you'll see that waves can construct or destruct. They're destructive at 45 degrees.

Now if we move them closer together, then they're not as far apart when you move off axis. So they'll destruct only when you move further off axis.
post #46 of 670
Quote:
Originally Posted by tuxedocivic View Post

Actually active wouldn't fix the problem. The problem is "hard wired" into the speaker like the directivity of a horn. What active can do is lower the cross over and make it steeper which can mitigate and even eliminate the problem. Like what mark said he did in the Catalysts.

This of it like this. Standing directly in front of the speaker, both woofers are exactly the same distance from you. Now move to the left. Now the left woofer is closer and the right woofer is further. Ok now imagine the woofers are shooting out waves, cause they are. The wave coming from the right woofer will be lagging the wave from the left woofer the distance that they're apart from you. If you look at that Wikipedia link, you'll see that waves can construct or destruct. They're destructive at 45 degrees.

Now if we move them closer together, then they're not as far apart when you move off axis. So they'll destruct only when you move further off axis.

That makes sense. So by bringing the woofers closer together you are improving the off axis response. Does bringing them closer together also improve the phase?
Also, how does the cancellation between the woofers work? This has always baffled me. Is this a problem when doing a design that only employs one woofer?

After google-ing for about a while, I feel like I understand the rest off this, just do not understand how cancellation works.
post #47 of 670
Quote:
Originally Posted by tuxedocivic View Post

Actually active wouldn't fix the problem. The problem is "hard wired" into the speaker like the directivity of a horn. What active can do is lower the cross over and make it steeper which can mitigate and even eliminate the problem. Like what mark said he did in the Catalysts.

This of it like this. Standing directly in front of the speaker, both woofers are exactly the same distance from you. Now move to the left. Now the left woofer is closer and the right woofer is further. Ok now imagine the woofers are shooting out waves, cause they are. The wave coming from the right woofer will be lagging the wave from the left woofer the distance that they're apart from you. If you look at that Wikipedia link, you'll see that waves can construct or destruct. They're destructive at 45 degrees.

Now if we move them closer together, then they're not as far apart when you move off axis. So they'll destruct only when you move further off axis.

Quote:
Originally Posted by Martycool007 View Post

Quote:
Originally Posted by tuxedocivic View Post

Actually active wouldn't fix the problem. The problem is "hard wired" into the speaker like the directivity of a horn. What active can do is lower the cross over and make it steeper which can mitigate and even eliminate the problem. Like what mark said he did in the Catalysts.

This of it like this. Standing directly in front of the speaker, both woofers are exactly the same distance from you. Now move to the left. Now the left woofer is closer and the right woofer is further. Ok now imagine the woofers are shooting out waves, cause they are. The wave coming from the right woofer will be lagging the wave from the left woofer the distance that they're apart from you. If you look at that Wikipedia link, you'll see that waves can construct or destruct. They're destructive at 45 degrees.

Now if we move them closer together, then they're not as far apart when you move off axis. So they'll destruct only when you move further off axis.

That makes sense. So by bringing the woofers closer together you are improving the off axis response. Does bringing them closer together also improve the phase?
Also, how does the cancellation between the woofers work? This has always baffled me. Is this a problem when doing a design that only employs one woofer?

After google-ing for about a while, I feel like I understand the rest off this, just do not understand how cancellation works.

I'm a fan of looking to the more clear conditions where then it's a matter of understanding the transition and space between.

Stepping back to phase, the simplest explanation is that phase is the difference in timing between two signals as a portion of the wavelength. I think you are getting confused in different applications and context. The case of the pair of woofers is a function of physical spacing which exists when they get the identical signal. A high or low pass creates delay that varies with frequency. We most often display this delay as phase, which is delay/period. We do so to give appropriate context and to indicate how 2 signals will sum together.

IMO the first important understanding here is what happens when you have two spaced drivers. More specifically, what happens at the frequency where they are spaced 1/2 wavelength apart. As tux' mentioned, directly in front/center of the 2 speakers the sound arrives from each source having traveled the same distance. Now rotate to 90 degrees off to the side. From the side you have the maximum difference in arrival time. The sound from the close driver arrives 1/2 wavelength before the far driver, and the result is significant cancellation. The result is a transition from full combination in front to cancellation at the sides. Real distances and size of drivers keep the cancellation from being perfect, but it is significant.
post #48 of 670
Phase is the relative time at which different frequencies arrive at a point. You can start at 0 phase with any frequency. Then the change in phase at different frequencies will tell you the relative time at which they arrive at the point compared to the frequency at 0 phase.
post #49 of 670
Quote:
Originally Posted by Mark Seaton View Post


I'm a fan of looking to the more clear conditions where then it's a matter of understanding the transition and space between.

Stepping back to phase, the simplest explanation is that phase is the difference in timing between two signals as a portion of the wavelength. I think you are getting confused in different applications and context. The case of the pair of woofers is a function of physical spacing which exists when they get the identical signal. A high or low pass creates delay that varies with frequency. We most often display this delay as phase, which is delay/period. We do so to give appropriate context and to indicate how 2 signals will sum together.

IMO the first important understanding here is what happens when you have two spaced drivers. More specifically, what happens at the frequency where they are spaced 1/2 wavelength apart. As tux' mentioned, directly in front/center of the 2 speakers the sound arrives from each source having traveled the same distance. Now rotate to 90 degrees off to the side. From the side you have the maximum difference in arrival time. The sound from the close driver arrives 1/2 wavelength before the far driver, and the result is significant cancellation. The result is a transition from full combination in front to cancellation at the sides. Real distances and size of drivers keep the cancellation from being perfect, but it is significant.


So how is phase measured? Can it be measured with REW?
How can phase be adjusted in a passive or active crossover? How do you know how two, of any given drivers, will sum when both on-axis and off-axis?

I do, of course, realize that I am asking questions that are very complicated to answer in a short thread,, but, if you don't feel like typing then the cliff-notes answers will suffice!
post #50 of 670
Thread Starter 
Lets drop the phase questions and open a thread on the topic if you'd like. It could take pages of forum posts to explain the topic. Even the cliffs notes version is fairly involved. If you open a thread il scan some pages from some old text books for you.
post #51 of 670
Quote:
Originally Posted by Martycool007 View Post

So how is phase measured? Can it be measured with REW?
How can phase be adjusted in a passive or active crossover? How do you know how two, of any given drivers, will sum when both on-axis and off-axis?

I do, of course, realize that I am asking questions that are very complicated to answer in a short thread,, but, if you don't feel like typing then the cliff-notes answers will suffice!

To very simply answer your questions: A capacitor voltage leads current by 90 degrees, and an inductor voltage lags current by 90 degrees. These values change based on the combination of inductors and capacitors. Resistors have 0 degree phase. Look up phasor diagrams.

Here's something to look at to give a little bit of an idea what happens with distance between wave sources http://www.mwit.ac.th/~physicslab/applet_04/physics_classroom/mmedia/waves/ipd.gif
post #52 of 670
Quote:
Originally Posted by tuxedocivic View Post

Lets drop the phase questions and open a thread on the topic if you'd like. It could take pages of forum posts to explain the topic. Even the cliffs notes version is fairly involved. If you open a thread il scan some pages from some old text books for you.

Agreed. Marty needs to do a bit of reading on his own and start a separate thread as we're now well past the applicable questions related to this build.
post #53 of 670
Ok so after google-ing for several minutes and re-reading every ones post, I know have a pretty good grasp on phase! Thanks to Tuxedocivic, Mark Seaton, Rahula7 & JWagstaff!

Now please accept my apology for derailing the thread.


smile.gif
post #54 of 670
Any more news on this fantastic project? I'm tv stand shopping and want to make sure this will fit as a center. Right now I'm going with 40 inches wide by 12 inches tall for dimensions.
post #55 of 670
Thread Starter 
Those dimensions should work.

We have to wait on the plastic SEOS 10 to get here. Good to hear someone is thinking about building it. It's the waiting game of course. Erich probably has a better handle on the timeline. In the meantime I'll be working on a single woofer version, as well as a two-way version.
post #56 of 670
I'm so choked that i'm 4 weeks late to this party. Cool-stuff!
post #57 of 670
Any update on the timeline for this to be able to buy ? Any way to approximate cost 400-600?
post #58 of 670
Thread Starter 
It won't be ready until the SEOS 10 is ready. Without a cab the parts ring up to about $300.
post #59 of 670
How would the sound quality for bass and midrange of this unit compare to the tempest or.zephyr ? For me the vocals and.voices are really important for center channel and want more bass from 60hz to 100hz then I.am.getting from my current lcr. Your from bc tux ?
post #60 of 670
Thread Starter 
I haven't heard the zephyr or tempest. This speaker does do excellent midrange though. There are a number of technical reasons for this and I heard it with my own ears, its very good. It is not very good in the 60 to 80 range though. I can do a 60hz XO to subs, but not as well as others. This speaker is about high effeciency in a narrow package. That pushes bass down the priority list. Most of my speakers have a goal in mind and bass wasn't it on this one.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: DIY Speakers and Subs
AVS › AVS Forum › Audio › DIY Speakers and Subs › A 3 way 99db multi configurable SEOS design