OK, lets design and build the best speaker possible (“No compromises?”)

[Paul H]

What a drag, Paul! How long till you get them setup?

A while - let's just say there might be another "I'm building a home theater and looking for suggestions" thread soon

[Fryguy]

I think a main point to bring up in this is "how do you approach perfection?"

Some people like to pursue exotic designs like folded horns and transmission lines matched with crazy looking baffles, while others are going to use straight baffle sealed enclosures with tons of passive crossover circuitry, while others are going to go fully active and tri-amp with customized electronic crossovers doing everything.

Perfection is a funny thing in audio, as it's unattainable, and yet everyone tries to do it, and in vastly different ways.

[mazurek]

I have a project in the works for ultimate active three way box loudspeaker. Will be printing crossover boards at end of week, signal psu is done, next step when I have more money is to rebuild the amplifiers and package them all with the speakers.

If you are into it, this is a good learning project, it involves all levels, cabinet work, crossovers, psus, metalwork, etc, with each part being individually testable. Its not quite ready for release though, and I figure I'll be documenting it in about 2-3 weeks. If you want any details/plans for this type of project email me. (lee@mazurek.net -its full of spam anyhow so public forum mention doesn't matter).

Mazeroth saw an early version, it sort of sucks due to crossover issues, but it is super-clean, and an excellent platform for building.

[SteveCallas]

My normal target through most of the range is effective sensitivity of 82-83 dB

Assuming you sit ~10' back from your speakers, and you want the dynamic capability of a clean 105db, aren't you going to need to be running ~1000 watts?

[JonW]

Thomas-
I like your Garrison Keillor quote.

Taco-
I’m not familiar with “ripoles.” What are they?

Mazureck-
Will you be posting a writeup of your project?

Perfection is a funny thing in audio, as it's unattainable, and yet everyone tries to do it, and in vastly different ways.

But isn’t that what makes it fun? If there was only one way to make a speaker, there would only be one main design out there. The only reason to deviate from that would be cost. No fun in that.

I've got about 480 hours of accumulated vacation time.

Jon-
“I can stop working any time I want…” I’ll believe you’ve taken the time off only after it actually happens.

[JonW]

A couple related questions for everyone:

What is the best DIY speaker you’ve ever heard?
What is the best commercial speaker you’ve ever heard?

If you can, provide details as to the style of the speaker (3-way, box or dipole, etc.), drivers, and such. Maybe approximate costs as well. It might help provide an idea of what people like a lot.

[seattle_ice]

The best speaker I have ever heard in person are my buddy's Dynaudio Confidence C4's. Of course, he has spent countless dollars and hours getting the room set up just right and the elctronics gear suitable for high end audio.

I am thinking about a set of TMWW with maybe the Accuton C23-6/C90-T6 and a pair of 10" woofers. Aurasound? Not sure.

[TacoD]

Thomas-
I like your Garrison Keillor quote.

Taco-
I’m not familiar with “ripoles.” What are they?

A ripole is a special variant of the dipole system. With a small chamber at both sides of the woofer. Several configurations exist, I used the W-profile with great succes. For that I used two 99 dB woofers which resulted in ~ 93 dB system sensitivity.



For more information see this link.

Axel Ridtahler is not very open about the exact configuration, but my findings are that you have to keep the back chamber as small as possible. As for driver choices, an high Q helps (like every dipole), but I got good results with low Q drivers (needs extra eq to get a smooth response, resonance peak of the system). One of the big advantages is that the Fr is lowered.



A really nice picture of such a system with two W-ripoles per side:

[SteveCallas]

Taco, that's very interesting - a few questions if you don't mind.

What does the ripole accomplish, or what is the theory behind using such a configuration?

Are the two woofers wired in phase in the W Linkwitz variety?

With the woofer section of a three way speaker typically covering up to ~200-300hzhz, and no direct "line of fire" from the face of the woofers to the listener, are their problems with locatability or directionality with the sound coming out of the chamber?

[seattle_ice]

Are the two woofers wired in phase in the W Linkwitz variety?

You may be asking about the Linkwitz, but just to clarify, looks like the one in the pic is the Ridtahler version with the woofers facing each other. I could be mistaken, though. And I would like to know how they are wired also.

[Davey]

There is only one possible wiring scheme that would work. If reversed there would be very little output.

In the Ridtahler configuration both cones would move away from their magnets at the same time. This is a valid configuration, but unfortunately it doesn't have the advantage of minimizing even-order distortion that the standard Linkwitz W-style does. Of course, maybe maximizing some distortion was the intent.

Also, the Ridtahler "configuration" appears to be identical to a stacked (and reversed) woofer configuration that Linkwitz advocated many years ago. Although he didn't advocate shrinking the cavity chamber to such a small dimension.

http://www.linkwitzlab.com/woofer3.htm

Cheers,

Davey.

[Dennis H]

Yeah, I'm not buying the whole "ripole" thing. Sounds like an advertising buzzword to me, where the designer uses his own initial in the name, and I think we can do without it. I think for the DIY community, we can call them what they are, either a W frame or an S frame (N frame?). The advantages and disadvantages of making the internal volumes larger or smaller can be discussed as a separate issue.

[AJINFLA]

LOL, I just said pretty much the same thing on DIYAudio. You can call me a monday morning QB or flat out liar if you will, but I had sketched out the BMC configuration long before I saw their sub, simply as a way of compacting a W down to minimal size. Including the magnets sticking out the sides using my cheap but deep TB woofers! I was unaware of the air mass loading this would create, since I never did build it - or a thousand other things I'd doodled up.
The loading comes at the price of an even lower cavity resonance than the H.
It's a W or a dipole if I ever saw one LOL.

cheers,

AJ

p.s. if I ever build one of my other configurations, maybe I'll call it a Jipole :B
Although Apole might be more apt.

[jdybnis]

It's not so much the design itself as much as the analysis of the design that he's taking credit for. The secretive part (which I think is a silly thing to do) is the model he uses to simulate and derive a correction filter. To give him credit the word "ripole" doesn't appear in that pdf documentation, he calls it a quasi-dipole.

I'm guessing that the big benefit, aside from the compact form factor is the mass loading lowering the fs. I also think the cavity resonance is higher than in a w-frame. The cavity being smaller and all. I don't quite buy the measurements in the pdf. I don't see how he's getting a consistent cardioid shape across the who frequency range.

[Dennis H]

The secretive part (which I think is a silly thing to do) is the model he uses to simulate and derive a correction filter.

Probably because he does it just like SL and anyone else designing a dipole. Build it, measure it and design the filter from the measurements.

[SteveCallas]

So the main benefit is lowering the Fs? If looking to use drivers for the woofer section of a 3-way, wouldn't it just be easier to use a driver with a naturally lower Fs? I think the Dayton HiFi 15" would work nicely for a woofer section.

[TacoD]

Taco, that's very interesting - a few questions if you don't mind.

What does the ripole accomplish, or what is the theory behind using such a configuration?

Are the two woofers wired in phase in the W Linkwitz variety?

With the woofer section of a three way speaker typically covering up to ~200-300hzhz, and no direct "line of fire" from the face of the woofers to the listener, are their problems with locatability or directionality with the sound coming out of the chamber?

- A ripole is less sensitive for room placement, and has higher output for the lower octaves.

- woofers are wired in phase.

- you cannot locate the sound of the ripole in my setup, I use it only at <200 Hz. It sounds very clean with strong bass. It works in my room better than the same woofers on a straight baffle.

Here is my measurement of the ripole (W-config). The two woofers I used have a Fs of 38 Hz.




[off-topic, do not read if you are easy offended :dancenana: ]
I do not say it's the best configuration, I only tell what a ripole is. For me it works very well and I like the mass loading, it works better for my needs. I hope this board will keep their open minded people, if I want flame wars I'll read diyaudio.
[/off-topic]

[SteveCallas]

I didn't mean to come off as resistant to this type of design, it's just the first time I've heard of it. Here's a few more theaoretical questions.

- A ripole is less sensitive for room placement

When compared to what?

- you cannot locate the sound of the ripole in my setup, I use it only at <200 Hz

Doesn't bass only stop being locatable for average hearing <100hz?

[TacoD]

I meant the blend with the rest of the system, therefore in my 3way configuration I reached a balance in which I cannot separate the ripole from the midbas (I used a double ripole on each side)

[SteveCallas]

Ok, let me try again, and realize that I'm just being curious. If the <200hz frequencies are exiting through side slots of the speaker as opposed to the front of a speaker under most normal configurations, does it affect the way the ear locates or places those frequencies?

By any chance, are the speakers positioned so that the side slots face the listener as opposed to the front or back of the speaker?

[TacoD]

I have positioned the ripole as depicted on the photo from my initial post. The output is shaped like an '8' (not symetrical, front lobe is bigger) And is coming from the "vents" at the back and front of the system.

For conventional loudspeakersystems low frequencies are also radiating in the 360 plane, depending of baffle width. So there is for the ear not a big difference with locating those frequencies. In the case of a Ripole where output minimum is at the left and right of the system, (imagine the 8 shape) there the output is almost nihil.

[Davey]

Taco,

Thanks for sharing your info. Do you have a nearfield (front and back) SPL plot also? (I think that would be very illustrative.)

FYI, here's an impedance measurement and nearfield SPL (no EQ or xover) plot from one of my Phoenix stacked woofers with 1252DVC drivers installed.





Note that the cavity resonance peak is approximately the same frequency as yours but less in magnitude. It's also of note that the "mass loading" of the drivers has lowered the Fs from approximately 17Hz in free-air to 13Hz in my speakers. (The -3db point in the SPL measurement is approximately 17Hz.) These drivers have Qts of 0.5.

It seems to me that the "ripole" approach might be valid for drivers (as you've used) which have somewhat higher Fs and roll-off points, but a better approach would be to select more suitable (lower Fs) drivers to begin with. Such drivers would probably have heavier cones and already be "mass loaded" before you start.

Little did I know that I already had ripoles in my listening room for the last six years.

My two cents.

Cheers,

Davey.

[TacoD]

I only have a nearfield measurement of my test ripole (bigger chamber behind the woofer, so resonance doesn't match with my graph posted earlier). My other measurements are with x-over filters (changing the filter on the fly and then measure again repeat this to get a nice freq response).



I haven't made polar responces but on axis the output on both sides is about the same. In my small room it is very difficult to measure the low frequencies. For my application I used PA woofers so I get very high efficiency. With long throw woofers you won't get that.

[jdybnis]

Davey,

It looks like the first resonance on your graph is at about 150hz and on TacoD's at about 250hz. Not surprising since the cavity is smaller on the ripole.

Doesn't bass only stop being locatable for average hearing <100hz?

AFAIK, a 400hz tone is not localizable. The wavelength at 400hz is about 4x the distance between your ears. Above that frequency you can localize pure tones. That means though by 400hz you want a sub's output to be at inaudible levels.

Some time ago I ran the numbers and came up with ~200hz, as the frequency of a 10" diameter source becomes truly (360 degrees) omnidirectional.

[Davey]

Hi Josh,

No, that peak at 150hz is something else related to the Madisound driver itself. It shows up even when I do an impedance sweep with the driver in free-air. Note that it corresponds to a slight dip in the SPL response vice the large, primary peak at 230Hz.

I'm not sure how much smaller the cavity actually is on the ripoles. In my enclosures the magnet back is only about 0.5" away from the side panels. Any smaller and I would have to bore holes in the sides ala ripole.

Cheers,

Davey.

[jdybnis]

Davey,

You've pretty much swayed me to your opinion. It's not significantly different than the Linkwitz w-frame. And like I said before I don't trust the measurements that show a cardioid radiation pattern. There is no apparent way that the rear wave could be lower level than the front wave.

[Davey]

I'm not sure I was trying to sway anyones opinion. I think this is mostly semantics since both concepts are essentially the same thing. And they're both a long way from a cardioid radiation pattern. The "Blow Motion Concept" is a pretty gimmicky/funny label though.


On a similar note I think this is of interest: (maybe you and others have already seen this.)

http://web.telia.com/~u82608519/WF8.ppt

Stefan did a lot of work using the latest version of LspCAD 6.0 to simulate the response of a "W"-frame setup but ultimately found the actual response was quite a bit different than the simulation.

It certainly shows how difficult and complicated simulating something like this can get. He's the first one I've seen go to this length.

Cheers,

Davey.

[JonMarsh]

Thanks for the interesting link, Davey.

~Jon

[JonMarsh]

Assuming you sit ~10' back from your speakers, and you want the dynamic capability of a clean 105db, aren't you going to need to be running ~1000 watts?

Any nominal sensitivity MT built with "conventional" drivers having 90 dB sensitivity and full BSC, after 1 dB of insertion loss will only be about 83 dB sensitive at one meter. Some will be even lower, like the Joeseph Audio RM33si. Maggies are about 82-83 dB.

BTW, that works out to 128 watts peak for 104 dB at one meter with everything padded down to match the bass drivers I was last working with. With the newer woofers, will probably pick up 2 dB, maybe three.

So, accounting for fall off in my listening room, to have similar levels at the listening position might indicate another 3 dB of power- 250, let's see, for peak levels. Aragon 8008's clip at about 290, and my Palladiums (which this project was designed for) clip at about 580 watts.

I think that's plenty loud for me...

But for when it isn't, I'm arranging the crossover modularly so that an additional woofer panel can be crossed in, and the midrange sensitivity raised, ala Gradient, Beethoven Grande, Apoggee's, etc. Then we should have another 6 dB of output available. Me, I don't think I need 110 dB in my listening room, but I've done silly things before, so who knows?

It's also a funny situation in the midrange, as that driver by itself is 92-93 dB/watt, so the actual power it needs for 104 dB is only 16 watts.

[seattle_ice]

But for when it isn't, I'm arranging the crossover modularly so that an additional woofer panel can be crossed in, and the midrange sensitivity raised, ala Gradient, Beethoven Grande, Apoggee's, etc. Then we should have another 6 dB of output available. Me, I don't think I need 110 dB in my listening room, but I've done silly things before, so who knows?

Wouldn't that be a passive version of Thomas's Ultra Arvo's? Isn't that essentially what those were?

I have been following this thread (including the tangential ripole discussion) with some interest.

I just finished building a 4700 sq. ft. house, and am looking at building a dedicated theater in the 600 sq. ft. space below the garage floor. So I try to follow all the different ideas floating around the forums.

I really love the Arvo's, if not simply because they are not the typical 'I bought a stereo and some speakers, woohoo.'. Not too sure how well they would work in a theater/music studio setting though.

My wife now thinks I have gone crazy, since I have built three subs and four sets of speakers in the last two months. But I love to build things, and love to learn. But man, can this stuff get overwhelming in a hurry.

I also have to say that I am truly impressed with not only the work, but the time that the knowledgeable people put into the designs and helping people on this and other forums. Time is a very valuable commodity, so thanks.