How compromised is this dipole tweeter?

[otto88]

I’ve just built the frames for a 3 way active dipole, with “nude” mids and tweeters. Now a tweeter “dilemma”:

I want a dipole tweeter that is high efficiency to allow high peaks with a moderate power tube amp, by which I mean drivers > 95 dB. I’m prepared to sacrifice some dipole “true behaviour”. How compromised is this:

I have a pair of 97 dB tweeters (Supravox, that I got 2nd hand); and two pairs of Scan Speak D2608 Discovery tweeters. The D2608 is c 92 dB.
My idea is to have the Supravox front firing, in the centre; and rear firing: one D2608 a side, wired in parallel (and attenuated to match the Supravox). The crossover will be c 2000 Hz. Mounted “nude” in a frame.

By one D2608 a side, I mean the 3 drivers form a row; with the two D2608s from above probably set a little forward, so the acoustic enters are in a line. The left – right distance between the centres of the two D2608s would be about 220 mm.

To the rear there would be comb filtering, above frequency ...?
But how much, and how would that rear comb filtering affect the sound?

[otto88]

Thinking further, the comb filtering could be reduced by mounting the rear firing tweeters further apart.
The way I've done the frame, they could by about twice that width apart, c 450 mm.
And they do not need to fire directly back, eg they could be angled at 10 - 30 degrees to either side.
Both these mountings having less comb filtering, but being less dipole in overall system response.

[Rudolf]

Originally Posted by otto88

The left – right distance between the centres of the two D2608s would be about 220 mm.

Left-right

I would mount them above and below the front tweeter. Can you remove the frontplates to get them nearer together?

[otto88]

I could do them: above and below - so that is better. .

Not sure about *remove the frontplates, but possibly could trim them(?), if I could avoided damaging them.

[otto88]

How would it be, if the Left – Right distance between the centers of the two rear firing tweeters were about 120 mm?

[cuibono]

I'm having trouble visualizing what your talking about. A drawing would help.

I've tried nude, back-to-back tweeters, and found them to sound off. The comb filtering/dipole peaking is obvious.

If your set on using two tweeters per side, you might have better luck if you separate them more - like Linkwitz does. His baffle is about 33cm wide. The tweeters probably will have a much smoother response this way.

[Rudolf]

Originally Posted by otto88

How would it be, if the Left – Right distance between the centers of the two rear firing tweeters were about 120 mm?

Top row is the horizontal pattern, bottom is the vertical pattern. Left is to the left side, right to the right side. "vorn" is 0°, "hinten" is 180°

If you mount the drivers vertically instead of horizontally, you have to switch top with bottom.

[otto88]

Much appreciated Rudolf.

Did you use Boxsim? By “the vertical pattern”, you mean with the rear tweeter above and below the front tweeter – yes, much better in that direction.

In the other direction, somewhere between 30 and 60 degrees, the behavior goes bad. Can I ask please, how does it sim at 45 degrees?

I recall a thread about cutting down the faceplates on conventional tweeters like these (I think cheap Daytons were used), as small as possible, to push the good polar behavior higher in frequency. On the directivity of dipole tweeters

I’m a bit reluctant to cut into the expansive Supravoxes. And even if I do, as the magnets are larger (also on the SS tweeters), I will still only be able to get good dipole behavior only so far.

So maybe I could cross these tweeters, to a Neo3 used as an active supplementary tweeter. I know the Neo3 is an ideal dipole tweeter, but rejected it because of SPL capability is just a little low for me.
However its power handling maximum of “20 watts RMS/ 50 watts max”, limits SPL ~ when it’s crossed around the “normal level” of 2 kHz. 91 dB + 20 watts gives a maximum of about 103 dB at full excursion.

I recall if frequency is doubled, that excursion reduced by factor of 4.
So if the Neo3 were crossed around say 5 kHz, 4th order, maybe I could safely put 60 watts into it (without quite full excursion, I’d guess) and achieve c 107 dB. IIRC the Nao Note is crossed (passively) around 5 kHz.

I would still need to trim the Supravoxes and Scan Speaks. But in conjunction with the Neo3, could get a much higher range of good dipole behavior.

Cuibono, I appreciate it could be difficult to visualize, I will try to arrange a drawing, but have no CAD experience, nor scanner

It’s actually 3 tweeters per side: one (97 dB) front firing; two (91 dB) rear firing, probably side by side.

What do you think of combining

• a wide Orion style baffle (330 mm), with
• no trimming of drivers; and possibly also
• the Neo3 crossed around 5 kHz

[otto88]

To reduce comb filtering from the two rear firing tweeters, perhaps i should, instead of having these two fire directly back in parallel, have their aimed at say 20-30 degrees apart. Then there would be much less interaction between them.

Possibly mount them each on rotating gimbals, to fine tune their angle of radiation.

[otto88]

Any comments welcome . .