[Danny Richie]

I will try to keep this a narrative of the history or progression of one thing that lead to another that comes around to the discussion at hand. I think the brief history will add some useful background. I will try not to make it commercial or cover any of my own products in an advertising manor.

Okay lets start with that BG tweeter that you are considering.

I was the first company outside of BG to get those tweeters when they first came out. One other company in Germany also received them about that same time too. So technically I am at least one of the first. So let's just say that I have been using them for a long time.

I used the standard Neo 3 in line sources that I designed for Epiphany Audio and used Neo 8's in a line source kit that I offered many years ago called the Alpha LS. I then came out with several kits using the Neo 3pdr that I still offer. As a single tweeter the pdr had better dispersion so it was the one to use.

I later was allowed to use an OEM version of the tweeter, that until that time, BG had kept only for their own use. It used a deep back cup that unloaded the diaphragm and allowed it to act as if it was in free air. This version took the knee out of the response in the lower region so instead of a flat response that finally dropped off steeply, it had a smooth response that started rolling off a little sooner but did so much more gradually and had a nice transition.

The deep cup version sounded better too. It was more transparent and at ease. The crossover didn't have to be as rough on it to force a smooth roll off either. It was a more staggered set of values that made better use of the natural curve and went with it rather than forcing it.

This better sound of the unloaded diaphragm was what pushed me to the next level and that was trying a full open backed version.

So I had BG make me some samples of a bit of a hybrid version. It keep all five rows of magnets like the standard Neo 3 but used a dense felt like the pdr version one each side of the diaphragm, but only on the front side only. It was like a pdr but with a little higher sensitivity.

When I tried it in several different test baffles for the prototype speaker that I was working on, it became real apparent that the way the back wave was handled was a big issue. If you just cut a big hole to mount it in it, then the response was disrupted, and the front did not match the back. So I had to have several baffles CNC cut with different types of small wave guides on the back side to control the rear wave. The best one was a through hole that was just the size of the diaphragm and had a 1.25" radius on it. My test baffle was 1.25" thick BTW.

That covers the surface of the tweeter testing. I can post pics of baffles if anyone is interested.

More on the rest of the design in a little bit.

[Danny Richie]

This new model that never came to be was to be the next step up from my OB-7 plus, OB-7, and OB-5 models that I sell as kits. Those use an MTM design in an open baffle and the deep back cup Neo 3 pdr. The bottom end is handled by conventional sealed or ported woofers depending on the models. The new design was to be a full on open baffle from top to bottom.

It is very similar to what Nyal has in his mind (I think).

I wanted to stick with my M-130's for the MTM section as those were just my favorite little woofers. For various reasons I like them much better than the Seas Excel  line (tried them all), Scan Speak, etc, but that's another story.

I also had been planing and developing some woofers with Brian Ding of Rythmik Audio using his advanced servo control technology, and applying that to 12" woofers that would be designed from the get go to be used in an open baffle configuration.

Here is an early box drawing: http://www.gr-research.com/pdf/Super%207%20drawings.pdf

And here are a few pics of the test box:





I also decided to add a couple of my M-130X woofers to the outer ends to help fill the range between the MTM and the servo subs. I wasn't sure at first how well the servo subs would handle playing up to 200Hz or so as the control was only suppose to be maintained up until the inductive reactance of the woofer started to cause the impedance to rise.

The outer woofers also shared that load in the range below or around 200Hz to increase power handling and lower distortion that could be caused by driving them hard.

It proved later to not really be necessary as the servo subs easily handled clean output to 300hz or so.

More in a little bit.

[Nyal Mellor]

Methinks this needs a rethink.
OK, so the tweeter is physically forward of the woofer and needs to be delayed, yes? Now, turn the baffle around. The tweeter is now behind the woofer and it is the woofer which needs delay. It would take quite the circuit to solve that pickle. Of course physical alignment solves the problem without the paradox.

After reconsidering this for all of about 5s, you are are of course right.

[Danny Richie]

Another problem that I had with the design was the look.

Testing proved that a wider baffle up top really hurt the imaging as it really increased the surface reflections around the drivers. So I wanted to keep it narrow. But then I had these big woofers at the bottom. I flared the sides out a little bit as seen in the drawings, but it still did not look right.

So I decided to lay the woofers over side to side instead of up and down like the W frame that I was testing in. This made the speaker more narrow, but a little taller.



Here are the new drawings: http://www.gr-research.com/pdf/Super7cabinet.pdf

I designed a little scooped area right under the lower woofer to force the drivers a little lower and help keep the tweeter at ear level.

The prototype was incredible. I must say that it was easily one of the best sounding speakers that I ever heard.

There were some issues though.

For one, I design products for a lot of different companies in the retail market. So as not to compete against them I try to keep my own products out of that market and sell them as kits. This speaker was already to the point to where the average guy can't build it. The woofer section was to be a stacked lamination of Baltic Birch, and the structure that held the woofers sideways was complex. All of this had to be CNC cut. So did the upper baffle because it was critical that the tweeter have a CNC cut section.

And if I wanted to order a production run of those tweeters from BG (a custom version) then I'd have to pony up for a 500 piece order.

The whole thing then just got put on hold.

The servo subs in an open baffle were a home run though, and I still used them in a new series of speakers that even exceeded this prototype. So it wasn't a total waste of time. One thing always leads to another.

I also still have some new ideas about making a speaker like this happen, and have a plan for a full open baffle line using the Neo's, and some new custom 6.5" woofers designed for free air operation. I even have some 8" servo subs in development, also designed for free air operation. We will see where that goes. Maybe it will be a new product line for next year.

I hope this information helped or was at least an interesting read.

[Saurav]

Thanks for the writeup, that was pretty helpful. I'm interested in what other baffle configurations you tried with the tweeter. Also, the back of the cutout - is that similar to the photos you'd sent me when we were discussing this, or did you try some other things as well?

[Danny Richie]

Thanks for the writeup, that was pretty helpful. I'm interested in what other baffle configurations you tried with the tweeter. Also, the back of the cutout - is that similar to the photos you'd sent me when we were discussing this, or did you try some other things as well?

I don't recall for sure what I sent you. I may have showed you front and back sides of a couple so that you could see the differences in what I tried.

[Nyal Mellor]

Hi Danny

Thanks for the information you posted! Always interesting to read about someone else design process, outcomes and challenges.

I'm not trying to get dipole radiation down below 100Hz or so. I'm going to use multiple monopole subs (at least two, maybe three).

The challenge seems to be 'how to maintain good polar response from a midrange driver run between 100Hz and the crossover to the tweeter'.

Using something like the BG Neo 2 or the majority of other tweeters would seem to lead you down the path of a three way or a two way with multiple smaller midrange drivers due to the still relatively high XO point that you need. Both seem to be compromises but maybe they are required. Is it possible to identify a design that would not compromise max spl (due to displacement limitation) whilst allow a two way, two driver dipole above 100Hz? By my rough in my head calculations we would seem to need a single driver capable of 700Hz to 20kHz, crossed over to something like Linkwitz uses, an 8" driver?

[Danny Richie]

I'm not trying to get dipole radiation down below 100Hz or so. I'm going to use multiple monopole subs (at least two, maybe three).

Okay multiple subs can work well. I have a lot of customers doing that with our servo subs. One of them just posted right here about it: http://www.audiocircle.com/index.php?topic=84617.0

100Hz is a pretty high crossover point for a separate sub though. That can cause you some problems.

Check a wavelength chart: http://www.soundoctor.com/freq.htm

100Hz is about 11 feet long. So you really want to be within half of that distance to the driver that you are crossing to in order to avoid cancellation issues in the primary output. It will be bad enough that you will have cancellation going on in the room reflections. So you have to have the sub within 5.5 feet of the speaker at that wavelength.

The challenge seems to be 'how to maintain good polar response from a midrange driver run between 100Hz and the crossover to the tweeter'.

The Neo 3 and several other tweeters out there can be crossed fairly low (in the 1,800Hz range or less). You can do this and still get great polar response (vertical off axis response).

Read this thread and look at the examples and the measurements to see what is possible: http://www.audiocircle.com/index.php?topic=83841.msg815287#msg815287

It is all about the crossover point and distance from acoustic centers.

Using something like the BG Neo 2 or the majority of other tweeters would seem to lead you down the path of a three way or a two way with multiple smaller midrange drivers due to the still relatively high XO point that you need. Both seem to be compromises but maybe they are required. Is it possible to identify a design that would not compromise max spl (due to displacement limitation) whilst allow a two way, two driver dipole above 100Hz? By my rough in my head calculations we would seem to need a single driver capable of 700Hz to 20kHz, crossed over to something like Linkwitz uses, an 8" driver?

The upper crossover point is not an issue. The lower one that you are shooting for is a big issue.

To get a single driver (even an 8 inch) to play down to 100Hz in an open baffle is not a happening thing. If a driver were designed to hit that goal then it would give up a lot of sensitivity and SPL capability.

You know that it is going to all be about surface area and X-max to hit any kind of SPL levels down there.

I tried something using 5 of my M-165X woofers (XBL^ motors) and a single Neo tweeter. The tweeter only crossed to the top woofer though. You run them all in series then by-pass all of the lower woofers with a cap value in the 56uF to 75uF range. You then just get the top woofer playing to the tweeter but all of them sharing the load in the lower ranges.

Doing this I got it to play down pretty low in free air, but my test baffle had some diagonal shaped wings on the back side to help stabilize the front baffle. By the time they reached the top woofer they were gone. They started at the top of the second to the top woofer and angled to the floor where they were about a foot deep.

A foot deep was way too much and it set up a real resonance cavity in there. It would have to be done with no wings.

The sensitivity is still the same as one woofer after compensating for baffle step loss. In this case about 85db with this woofer.

Something like this could be done and get down to below 100Hz with good levels. With a high Q woofer designed for this application, maybe down to 60 to 70Hz or so. You can't do it with one though. No way.

The real cut off point is below 200Hz. That is no problem if when the lower woofers are right there under the mid-bass driver. Then power the lower woofers with their own amp.  Problem solved.

[Nyal Mellor]

Danny

Not an issue to place subs where you want them by time delaying the mains, phase cancellation is an issue if you aren't though, I understand that.

I understand your point about displacement limitations from the 100Hz XO point, though that is what I have in the Orions at the moment with the 8" driver and there is no displacement issue. The problem though with that size of driver though is maintaining smooth polar response until the handoff to the tweeter, which AFAIK is 1.4k in the Orion. So that is why you suggest multiple smaller drivers, in a 2.5 way type configuration. Maybe that is the best answer. The other approach would seem to be using something like a Beyma TPL150 and crossing over lower. Still if we use a 6" driver we would sacrifice some SPL unless we experiment with higher XO to the mains from the subs, maybe 120Hz. I have read though that we start to localize bass at these frequencies so that may not be the best idea....

If, as in your suggestion, we use a midbass down to 200hz, then we would use another driver to cover the rest of the range, and putting a bass driver there is not the best design option in my opinion. Always a compromise between soundstage, bass response and don't forget speaker boundary interference. Even with a dipole which excites room modes there is still an issue with coplacement of bass and mid/treble systems. Much better IMO to use subwoofers placed in corners for the range under 100Hz or so. Remember that I am talking about implementation in a system with time delay and EQ capability. Obviously most audiophile systems don't have that but mine does and I'm used to this kind of thing from the high end home theater side.

[Danny Richie]

Danny

Not an issue to place subs where you want them by time delaying the mains, phase cancellation is an issue if you aren't though, I understand that.

Actually, when you time delay them to align them you are only doing so for one point in space. So let's say you have a couple of subs right between the speakers and 4 feet back. You can shift the phase to nail it at the listening position, but still have a cancellation to the left and right because you have delays in those directions. It is all about distance and the wavelength.

Now it the crossover point were half of that, say 50Hz. Now you are talking about a 22 foot long wavelength. So if that side wall area had a 180 degree out of phase rotation at 100Hz, it will only be a 90 degree phase rotation at 50Hz.

If the subs are part of the speaker, like the Orion, then the phase is the same regardless of direction because the distance doesn't change.

Distance is still important.

understand your point about displacement limitations from the 100Hz XO point, though that is what I have in the Orions at the moment with the 8" driver and there is no displacement issue.

That's true but it is very steeply high passed with your electronic crossover, and there is still an SPL penalty. Sensitivity is only 85db with one watt.

The problem though with that size of driver though is maintaining smooth polar response until the handoff to the tweeter, which AFAIK is 1.4k in the Orion.

What the Orion has fighting it is a greater acoustic center spacing. Still with that low of a crossover point, polar response will be pretty good. Did you read that thread I sent a link to?

So that is why you suggest multiple smaller drivers, in a 2.5 way type configuration. Maybe that is the best answer.

Well, the smaller drivers do allow for a closer center to center spacing. If you notice with my designs I even overlap my woofers onto the tweeters to get them even closer. And using more drivers does allow you to let them play lower.

The other approach would seem to be using something like a Beyma TPL150 and crossing over lower. Still if we use a 6" driver we would sacrifice some SPL unless we experiment with higher XO to the mains from the subs, maybe 120Hz. I have read though that we start to localize bass at these frequencies so that may not be the best idea....

When you go to a bigger driver then you really have to start considering the coaxial designs. Then you just have to worry about the falling off axis of the woofer especially if they get large. With the two coaxial's I am using in my V-1 and Super-V the crossover points are 1kHz. So no problem.

You only have to worry about localization of the bass frequencies if the subs are separate from the speakers. And it really starts around 100Hz or so.

[Nyal Mellor]

Danny, I understand the physics behind phase cancellation, that is one of the things that makes me think that separate subs below 100Hz or so is a good thing, that way you avoid the frequency response suckout created by the bass waves reflecting off the back wall interacting with the direct sound from the speakers which causes destructive interference at the listening position.

Why would I worry anyway about cancellation at other points apart from the prime listening position?

[Nyal Mellor]

Re sensitivity, don't really care, that's what big amps are for 

[Danny Richie]

Why would I worry anyway about cancellation at other points apart from the prime listening position?

Because we don't listen in an anechoic chamber. We listen in a room and we hear the in room response, not just the on axis response.

Using multiples doesn't take away from the fact that there are cancellations going on. But adds other peaks and dips to the whole picture in an attempt to even them out.

Re sensitivity, don't really care, that's what big amps are for 

Yeah I know. I also know exactly what it cost to build a top notch level amp and have big power. I can get the same quality in a lower powered amp for a fraction.

[Nyal Mellor]

Danny

I don't understand your last point. If you place a measurement mic at the listening position and that shows no cancellation because you have time delayed your subs, then that is what you hear. Of course that is a combination of the direct sound and reflected sound, but I don't see what that has to do with worrying about cancellation at other points apart from the listening position. ANY time you have two sound sources at different distances you will get cancellation. Whether they are bass units under the mid/tweeter or separate subs make no difference.

If you measure your dipole speakers in room you will see a suckout probably centered on 80hz if your speakers are 4ft from the back wall. If the difference between the direct and indirect path lengths (where path length is the distance the sound has to travel) is equal to half a wavelength then the two sound waves will combine destructively and a notch in the frequency response will occur. The frequency that interference occurs at can be calculated through application of the wavelength formula as follows: cancellation frequency = speed of sound / (2 * path length difference). That is what I am trying to avoid by separating out the low bass units and putting them against boundaries. I don't believe I have created any new problems that cannot be fixed through DSP based time delay. But maybe I am wrong, I would like to learn why.

In terms of compromises, I would rather compromise on efficiency in terms of SPL per watt than compromise on speaker design, at least in a home setting where achieving very loud SPL levels is not difficult, even with insensitive speakers. There are many very powerful amps available cheaply these days and anyway you are much more likely to run out of displacement than you are amplifier power, especially in an active system where each amp is only feeding one driver.

I am looking to improve on the Orion, which is in itself a very high performance speaker, not cost engineer a speaker to a pricepoint.

[Danny Richie]

I don't understand your last point. If you place a measurement mic at the listening position and that shows no cancellation because you have time delayed your subs, then that is what you hear. Of course that is a combination of the direct sound and reflected sound, but I don't see what that has to do with worrying about cancellation at other points apart from the listening position. ANY time you have two sound sources at different distances you will get cancellation. Whether they are bass units under the mid/tweeter or separate subs make no difference.

It makes a difference.

Let's take a single speaker in the room. We want a good polar response, good horizontal dispersion, and even response across the board. Why, because we hear the interaction with the room and we want side wall and ceiling reflections to be in direct relation to the on axis response and not a peaked or dipped reflection.

All of the sound from 20Hz (full range speaker) and up comes from one point in the room (same example). So the reflections in the room do cause a peak or a dip somewhere and at the listening position we can treat it with room treatment etc, to minimize it all.

All this is a given, you knew it all already...

Now we split from 100Hz or so away and position it from somewhere else in the room. Now we have the same thing to contend with before but from two sources. So on top of a peak or a dip that we get from the room response, we also get a secondary peak and/or dip from an out of phase cancellation from some of those reflection points. And remember 100hz and up starts getting directional so you hear it.

Just because the amplitude measures fairly flat at the listening position doesn't mean it will sound right.

Years ago there was a system showed with Tech audio gear (something like that), maybe it was Tec Audio. Anyway they separated the 200Hz range and below and sent to to subs that were corner loaded. And they relieved the mains from 200Hz on down.

They took in room measurements and made digital room correction and all of that so it was flat as can be from the listening position. The corner loaded subs may have evenly loaded the room, but the music was very disjointed. You could clearly hear the bottom end firing out at you from the corners of the room. 

[Danny Richie]

If you measure your dipole speakers in room you will see a suckout probably centered on 80hz if your speakers are 4ft from the back wall. If the difference between the direct and indirect path lengths (where path length is the distance the sound has to travel) is equal to half a wavelength then the two sound waves will combine destructively and a notch in the frequency response will occur. The frequency that interference occurs at can be calculated through application of the wavelength formula as follows: cancellation frequency = speed of sound / (2 * path length difference). That is what I am trying to avoid by separating out the low bass units and putting them against boundaries. I don't believe I have created any new problems that cannot be fixed through DSP based time delay. But maybe I am wrong, I would like to learn why.

My open baffle speakers are further out into the room and mine occurs lower in range.

I then add gain with the one band EQ and fill it back to flat again.

See amp: http://www.gr-research.com/index.asp?PageAction=VIEWPROD&ProdID=142

Those amps are great.

You can do what you are talking about and it will work well also, but you are going to have to lower that crossover point for it to blend well. Try to get those main speakers to play down into the 50Hz range and you have it licked. The lower crossover point will also minimize cancellation effect between sources.

[Danny Richie]

In terms of compromises, I would rather compromise on efficiency in terms of SPL per watt than compromise on speaker design, at least in a home setting where achieving very loud SPL levels is not difficult, even with insensitive speakers. There are many very powerful amps available cheaply these days and anyway you are much more likely to run out of displacement than you are amplifier power, especially in an active system where each amp is only feeding one driver.

I am personally not willing to compromise anything. So there is no way that I would run a powerful but cheap amp. And there is nothing like having a lot of head room.

If I were to go buy the amps that are in my listening room, brand new, and buy enough of them to power a set of Orion's, and I used one amp to cover both of the lower woofers on each speaker (six channels in all) then I would be spending $150,000 on amplification. No thanks. Good clean power just does not come cheap when you want plenty of drive.

And as per your very last statement, that is not a problem.

[diyfan]

I tried something using 5 of my M-165X woofers (XBL^ motors) and a single Neo tweeter. The tweeter only crossed to the top woofer though. You run them all in series then by-pass all of the lower woofers with a cap value in the 56uF to 75uF range. You then just get the top woofer playing to the tweeter but all of them sharing the load in the lower ranges.

Danny,

Just wondering if you could clarify for me how this is wired?  Does the woofer to tweeter x-over go across all five of the series wired woofers and the cap then parallels the 4 lower woofers? 

Any chance you could post a diagram  or sketch and maybe elaborate a bit on how it works?

Thanks
Sean

[Danny Richie]

Danny,

Just wondering if you could clarify for me how this is wired?  Does the woofer to tweeter x-over go across all five of the series wired woofers and the cap then parallels the 4 lower woofers? 

Any chance you could post a diagram  or sketch and maybe elaborate a bit on how it works?

Thanks
Sean

That's a good question. See what I posted in this thread:

http://www.audiocircle.com/index.php?topic=83025.0

[JohnR]

Is it possible to identify a design that would not compromise max spl (due to displacement limitation) whilst allow a two way, two driver dipole above 100Hz?

Why do you want it to be a 2-way? (Sorry if I missed it.)