Square Pegs

[Patrick Bateman]

Crunched some numbers with my calculator, and I think I came up with a schematic that will curve the wavefront of a flat driver, like the BGs:





Here's the original schematic. See how the additional length on the edges creates a delay? Basically we want the center of the diaphragm to 'lead' the edge by about four centimers.

[Patrick Bateman]

Originally Posted by avincenty

I have a bunch of sawafuji ribbon tweeters and decided to make a 4 paraline array to test how they work. I made an entry deflector that only deflects the radial holes from the tweeter, plugging the center hole. By using the paraline I will need a lot less tweeters (if I wanted to build a tall array). The one surprise is that the tweeter goes quite low. The vertical directivity is as predicted, big difference from the response in front of the array versus 8 inches on top of it. SPL levels up/down the array is constant so the paralines integrate well vertically. I attached some measurements although it is not a great setup as I want to stay air conditioned.

Building the paraline was not too difficult. I made a router template with indexing holes to repeat the cuts. I did a layer at a time. Next time I think it is easier and more precise to do the three eye layers at the same time.

In the end I cheated a little. I had made the template to use a 1/4 inch spiral bit which I did use, but I did not have enough 1/4 inch material to do it. I used 3/8 material instead. Also the front layer with the exit slot is not 3/8, it is 1/8 inch.

I want to try get some measurements with the array wedged into the corner and use the 90 degree wall as a horn.

Here's another driver that might make a promising mid range in a Paraline:

HiWave BMR12 Compact 2" Full-Range Square Speaker 12W 8 Ohm 299-208



Here's why this is a cool driver:

1) it's under six bucks
2) Because the diaphragm is flat, you don't need a compression chamber. Basically the reason we use the compression chamber on the mids is so that the compression driver doesn't reflect off the midrange drivers. But the compression chamber creates a plethora of problems, particularly by limiting the top end of our mids. The BMR driver solves that. On the downside, distortion is higher without the compression chamber. I can live with that.
3) Due to the very small size of the BMR driver, you can do some cool stuff with it. If I were to use this, I would orient the BMR drivers into a ring around the compression driver. So basically you have a pulsing 'ring' of sound. This keeps all the pathlengths the same, which simplifies the crossover, improves off-axis response, improves intelligibility, makes the response more uniform, improves imaging.


Rings are your friend.

[Nissep]

Talking about Hiwave got me thinking.
Why not go all the way and use the wall in the paraline as a speaker?
No surrounds on them to disturb the sound.
http://www.hi-wave.com/products/audio-exciters.php
HiWave HIAX25C10-8/HS 25mm Classic Efficient Exciter 8 Ohm 297-2105


I played around with vibration speakers before and depending on the surface you can actually get som decent midrange from them.

But, I wonder what kind of material and how thick one should use to get the most out of these exciters.

And where should we place them?

[Overkill Audio]

These pics were taken before I screwed in the drivers ( 128 screws!) and put on the black felt covers.
Anyway they are pretty amazing...!
I agree with Patrick, off axis and on axis, everywhere in the room is even balanced sound.
A great benefit is that when you Eq on axis for driver / room artifacts, it sounds just the same off axis...
The Sd is 960 ( more than a 15 inch driver!) per channel so even when you add in low end boost and power them with big amps the driver excursion is still under 2mm ( plus / minus 1mm)
I wire them as 4 groups of 1 Ohm in series so end up with 4 Ohm load but a sensitivity of 100dB per watt @ 1 meter!

I am going to start a build thread next month when I build another few pairs and I can take more pics during assembly.

Cheers
Derek.

[weltersys]

Originally Posted by Nissep

Talking about Hiwave got me thinking.
Why not go all the way and use the wall in the paraline as a speaker?

I played around with vibration speakers before and depending on the surface you can actually get som decent midrange from them.

But, I wonder what kind of material and how thick one should use to get the most out of these exciters.

And where should we place them?

The vibration element makes sound radiate from the entire (more or less) plane, which is a good thing for a wall mount but bad for a Synergy style horn, as the radiating elements can no longer be kept within 1/4 wavelength, so off axis response would be peaky.

[Nissep]

But if we take a solid ring that pushes on a ringmembrane instead of the whole area we get something like a bms 4590 om steroids.
The exciters pushes on the ring that transfers the energy to a flexible membrane that is of the right shape and size and at the right distance from the compression driver.

In this link there is a drawing of the bms 4590.
www.pa-forum.de • Thema anzeigen - BMS4590 an Vitavox 4cell Horn, Messergebnisse!(aktualisiert)

[weltersys]

Originally Posted by Nissep

But if we take a solid ring that pushes on a ringmembrane instead of the whole area we get something like a bms 4590 om steroids.
The exciters pushes on the ring that transfers the energy to a flexible membrane that is of the right shape and size and at the right distance from the compression driver.

An annular ring diaphragm has far less surface area than a cone diaphragm of the same diameter, and the small diaphragm area of the BMS 4590 is further limited by it's Xmax of less than 1mm.
Inexpensive cones mounted in an offset horn can far exceed the output potential of a BMS 4590 in the low midrange.

[Patrick Bateman]

Originally Posted by avincenty

Ok, so I ran another test to see the individual FR of the original paraline used in the array. Both single paralines show the bumps. The 3/16 paraline shows extended HF. 3/16 router bit vs 1/4 router bit.

Also, the interaction between the paralines in the array is lowering the HF and bringing up the mids. I don't have any experience with arrays so I can't explain what may be happening. What happens to the phase as you go from the the center of the slot to the ends?

People who have built Paralines have noticed that it's really hard to get a lot of high frequency output from the device. And this behavior is hard to predict, because hornresp isn't really designed to model loudspeakers past five or ten khz. (At these ultra short wavelengths the horn 'begins' way back inside of the compression driver, making it hard to model.)

While doing some research for my next set of speakers, I stumbled across a couple more factors which might be impacting high frequency response in a Paraline.

1) In a conical horn, the 'local flare rate' of the horn is changing as you move down the horn. At the throat the flare rate is very high, and it gets lower and lower as you get closer to the mouth. A Paraline is a radial horn, and a radial horn has a flare rate that's *similar* to a narrow angle conical horn, but it's NOT the same.



I think this could be a factor in the high frequency response of the Paraline. Basically at the throat a conical horn has a higher flare rate, which is better suited to high frequency loading. A few inches down the throat, the flare rate of the Paraline and a conical horn are getting closer - but the conical horn still has a higher flare rate. (IE, the conical horn is 'better suited' for high frequency loading than the radial/Paraline horn.)

The graph above shows this; the turquoise line is radial expansion, like in a Paraline. The dark red line is the flare rate of a 50x50 conical horn, like the flagship of the DSL line, the SH-50. See how the turquoise line is lower at all times? Basically, the Paraline is loading the lower frequencies with greater efficiency.

Now, I want to stress that I'm not using Paralines for their efficiency, I'm using them because they're compact and they give me a radiation pattern that works for my projects. But if anyone is wondering why the high frequencies on a Paraline are a bit muted, the reason may not be that the highs are muted; the real reason may be that the lower frequencies are just getting amplified more efficiently than the highs.

As they say in software, "that's not a defect that's a feature." Because one nice thing about good low frequency loading is that it reduces harmonic distortion by reducing excursion, which means we can add a lot of EQ in the top octaves without making the driver sound 'harsh.'


Here's a sim to illustrate what I mean. The sim shows the response of a B&C compression driver on a 360 degree radial horn, versus a 60x60 conical horn. In the sim we see that the lower we get in frequency, the radial horn produces more output on axis. On the downside, the radial horn is absolutely huge. (I used the same mouth size for both horns; since the radial horn has a much narrower angle, it's size dwarfs the conical horn.)

The main reason I did the sims was to illustrate the increased output of the radial horn - which may explain why the high frequency output of a Paraline seems 'muted.'

[Patrick Bateman]

In the last post, I showed some data that indicates that the high frequency loading of a radial horn like the Paraline isn't well suited to very high frequencies.



One idea I had was to 'fill in' the throat of the compression driver.
By reducing the area, right down into the compression driver, one might improve the high frequency output. (5khz and up.)

I think some of the VTC Paralines do this; there appears to be a cone shaped 'insert' in their Paraline.

[David_Web]

I thought the Paraline was parabolic due to parallel walls.
At least that is what I calculated.