Dipole idea: mount midrange drivers back to back without baffle

[OllBoll]

Hello there all, I just thought of an idea:

What I want to solve is uniform front and rear dispersion in the approx range 400-2000 hz. I also want this with reasonably high efficiency since I will drive it with an F2J to get the current source love.

The problem however is that the front dispersion of drivers is vastly different from the rear dispersion. I've had threads about this before like this idea where a sealed dipole would be made. The problem here would if I understand correctly though that path length difference is too long so I get dipole peak in the middle of the response and a null at 2000 hz or lower, which I'd rather avoid.

But mostly what made me wonder was that if I understand correctly the cones are fairly acoustically transparent so even in that idea the back wave would travel through the cone of the second speaker so the box has even less purpose.

I've thought about using the TB75 dome but it has too much high order distortion at 400 hz for my taste and the directivity would be hard to control unless I build a waveguide like the ATC but then again path difference becomes too large and the peaks and nulls come.

So, here comes my idea:

I thought about the earlier sealed and that if both drivers interract anyway why not just remove the box altogether and just push highpass crossover higher? So in practice the drivers would be mounted without a baffle in free air back to back like in my drawing. Current driver choice is still the 6ND430 and two of them should very easily be able to handle 400-500 hz highpass without a baffle. The only problem as I see is that impulse response would be compromised at higher frequencies since the two drivers won't add nicely and instead mess upp the response slightly. The question though is if this would matter in the real world, or if it is a much smaller evil than the other solutions... which is why I ask you

If you don't have enough good reasons for why this would be stupid I think I'll buy two drivers, test and then see for myself



The alternative is of course to mount them vertically with one driver inverted, but then the dirspersion is only uniform horizontally and it takes up very much vertical space and I want to produce good sound quality in a small package

[puppet]

I don't think it hurts to try anything .. as long as it doesn't involve explosives.

[Rudolf]

Originally Posted by OllBoll

What I want to solve is uniform front and rear dispersion in the approx range 400-2000 hz. ...

The problem however is that the front dispersion of drivers is vastly different from the rear dispersion.

- Define what grade of quality you want from the front dispersion at 2000 Hz.
- Define how much is "VASTLY different from the rear dispersion" in your mind.

Improving the first will inevitably compromise the second and vice versa. There is only a small corridor where the rear response is close to the front response and the front dispersion will still act below the dipole peak only. A 15 cm wide driver is already too large to achieve that at 2 kHz. Think about a single 12 cm driver.

[OllBoll]

Originally Posted by Rudolf

- Define what grade of quality you want from the front dispersion at 2000 Hz.
- Define how much is "VASTLY different from the rear dispersion" in your mind.

Improving the first will inevitably compromise the second and vice versa. There is only a small corridor where the rear response is close to the front response and the front dispersion will still act below the dipole peak only. A 15 cm wide driver is already too large to achieve that at 2 kHz. Think about a single 12 cm driver.

I want symmetrical front-rear response, ideally I want it so it doesn't have a front and rear, but just two directions it plays sound.

If I look up measurements of say the 6ND430 the front looks nice: (link to the site with the measurements)



The rear however isn't as nice:



And if I've understood correctly if I use a 6ND430 (which If I understand correctly has a ~ 12 cm cone and then with surround and frame totaling 16 cm ) without a baffle then then shouldn't first null be pushed to over 4 khz and the peak to about 2 khz. Atleast that's what it looks like in Edge =)

But since the drivers loose some directivity at 2 khz then dipole peak shouldn't be an issue like it would with a dome. The next issue is to through EQ fix the response which hopefully can be done quite well but even if there are some issues it will atleast be front back symmetrical.

As before I'm not planning to use them as normal speakers and listen to the front response but rather use them as surround speakers to fill the room with sound hence why I want symmetrical front-back response.

[StigErik]

In my opinion, its better to do it like this:



These are 15 cm units, crossed over at 1700 Hz.

[OllBoll]

That is the alternative yeah.

But the main problem with that is that it's so tall I want small cute speakers so the question is mainly if the slight smearing of the impulse response by placing them back to back is a problem or if it doesn't matter in practice. I have a pair of speakers now that if 2 khz smearing was a problem should sound horrible but they still sound great which makes me think that it might not be a problem at all.

[StigErik]

Tall? Most houses have 2.4 meter or so ceiling height .....

[Rudolf]

OllBoll,
I don't see measurements of any "nude" drivers. Those are the only valid ones if talking about a "baffleless" application.

There are drivers which seem to do better than yours. How about this Monacor SPH-176:
SPH 176 vergl 2.gif

[OllBoll]

Originally Posted by StigErik

Tall? Most houses have 2.4 meter or so ceiling height .....

Yeah, but then I'd have to use up the floor and dedicate the space. My goal is to make the satelites at most 30 cm wide, 60 cm high and 30 cm long, while performing awesomely ofc =). Separate subwoofer for under 80 hz which is bigger but is easier to hide somewhere.

Originally Posted by Rudolf

OllBoll,
I don't see measurements of any "nude" drivers. Those are the only valid ones if talking about a "baffleless" application.

There are drivers which seem to do better than yours. How about this Monacor SPH-176:
Attachment 289801

True, so the measurements might not be accurate. Best is probably to buy a pair and measure myself. Or find another driver that has as good performance but I haven't found one yet. All others seems to have far less sensitivity or way more odd harmonics in the tests or some other fatal flaw like sub par performance @ 400 hz like the TB75.

[Rudolf]

Looking for high efficiency in a small dipole is equivalent to believing in miracles. How often did you win the Grand Lottery up to now? If you count less than three - forget about your approach.

Rudolf