Dipoles Old & New - Show & Tell

[gvinson]

Thanks for the responses.

I'll go back and work with SL's spreadsheet some more.

By the way, I like the red widgets.

[oneoldude]

Then I get to thinking about a dipole bass. I'd absolutely love it if I could get bass from 300-500 cycles with a single driver down to a sealed sub, but that doesn't seam possible using SL's speadsheet. At least not being able to get a safe 105db peak (which is the design goal).

I think you might be wrong. Take a look at what can be done with one RSS315HF per side. If my calcs are right you should safely get there (with room to spare) with a Xover at 40 Hz.

Ok, I am going to try to do a S.L. analysis of what can be expected with RSS315HF drivers in an H-frame similar to the Orion’s. Please check my work to be sure I have not screwed up too badly. Ok?

There are three inputs necessary to use the Linkwitz spread sheet. They are:

Effective Piston Area in cm^2 = SD
Xmax in mm = Xmax
Effective Path Difference in mm = D

The RSS315HF has piston diameter of about 10” (25 cm). Therefore, SD = 490 cm^2. Xmax = 14 mm.

Effective Path Difference is deceptively long. The image shows how S.L. calculates it for an H-frame.



So D = 711 mm.

Plugging the numbers into S.L.’s spreadsheet gives the following numbers for one RSS315HF into free space.

40 Hz = 97 dbSPL
28 Hz = 88 dbSPL
20 Hz = 79 dbSPL

But wait a minute, that driver is not looking into free space in your listening room. It is on the floor and is effectively in half-space. So we must add 6 dbSPL to the outputs above. It looks like this. Not too bad so far, is it?

40 Hz = 103 dbSPL
28 Hz = 94 dbSPL
20 Hz = 85 dbSPL

But wait a minute; this is a stereo system so there will be two sides. At these low frequencies the signals will almost certainly be in phase. So we add another 6 dbSPL. Now the numbers look like this. givinson, I guess this is as far as you need to go. But there is so much more to the story.

40 Hz = 109 dbSPL
28 Hz = 100 dbSPL
20 Hz = 91 dbSPL

But wait a minute. If we double up the 12” RSS315HF’s like an Arvo (or an Orion on steroids) we must add yet another 6 dbSPL. This is getting interesting. The numbers now are:

40 Hz = 115 dbSPL
28 Hz = 106 dbSPL
20 Hz = 97 dbSPL

Wow! These numbers are right respectable aren’t they? And each additional doubling of the number of drivers will add another 6 dbSPL. For example, imagine a pair of columnar H-frames with 8 RSS315HF’s per side with one pushed up against the L and and the other against the R side walls of your listening room. They should fit nicely in a room with a 9’ ceiling. And, my oh my, the output would be as follows:

40 Hz = 127 dbSPL
28 Hz = 118 dbSPL
20 Hz = 109 dbSPL


I guess at this point all you need to do is decide how much grunt you really need, can afford and reasonably fit into the listening room.

[JonMarsh]

You get the numbers pretty well, though though in my experience the room gain with a diopole is less because you don't have nearly as much boundary re-inforcement. Part of the "traditional" 6 dB of reinforcement is due to more than one boundary - at low frequency it approaches quarter space at some frequencies, depending on placement. I'd suggest a figure more like 3 dB. Still, the rest all adds up. Which is why I think 2 12's with a moderate path difference (500 mm+) can do fairly well.

I figure 2-12's on main panels, then auxilliary woofers added as needed in W cabinets (push pull) will do fairly well.

[AJINFLA]

One slight problem. Using a 16" deep H-frame would put your resonace peak smack in the middle of 3-500hz (upper) range. Not good.
I would suggest a fairly large high x-max driver on a small as possible baffle, even though that will not be without problems either. Getting a dipole "woofer" up to 500hz is pushing things. Maybe a TC based 12 or 15 can go that high and sweep enough volume so that it will provide 105db output down to maybe 60hz-ish (on a smaller baffle) below which your sealed sub can take over. There is simply no way to get 500hz and 105db @ 30hz from a dipole woofer.
You need a small baffle for 500hz and a large for 20hz (high spl).
Thats the dilemma.
Do you really have to sit 18ft from your speakers 8O ?
Can't they be pulled far out into the room to bring them closer to you? That may be a large part of your problem also.

cheers,

AJ

[Dennis H]

Oneoldude, the problem with the deep H-frame is on the high end. You get the cavity resonance peak and a dip a bit higher. I think it was Monte Kay that did some experimenting and didn't like crossing much above 100 with a 16" H-frame.

Gvinson, have you considered a line array of smaller midbasses to mate to your ribbons? Pete Mazz tried a single 10" woofer and didn't like the sound as well as a line of inexpensive 7". The small dipole woofers will get you low enough to cross to a monopole sub.

[oneoldude]

Thanks Jon,

Now let's get back on topic.

How about this little guy?




This dipole is the Gradient 1.3 and was Jorma Salmi's first offering at Gradient. You can read about it here: http://www.regonaudio.com/Gradients.html

[gvinson]

Oneoldude,

Thank you for uncovering my mistake in the spreadsheet. I too added 490 Sd and 14mm of travel. The results gave me:

40 Hz = 89 dbSPL
28 Hz = 80 dbSPL
20 Hz = 71 dbSPL

I too corrected for 1/2 space, and got:

40 Hz = 95 dbSPL
28 Hz = 86 dbSPL
20 Hz = 77 dbSPL

I also corrected for stereo and had:

40 Hz = 101 dbSPL
28 Hz = 92 dbSPL
20 Hz = 83 dbSPL

The end result, as you can see, fell short of the goal. What I didn't do correctly is the path length ( I used 300mm for a shorter 12", not 711mm). I got started down the wrong road on path length from the very beginning and never caught my mistake. I took the somewhat simplified quote below too literally:

For a baffle of D = 20" (0.5 m) length and with L = 10" (0.25 m) estimated, the resonance peak in the dipole output is at F = 0.25*v/L = 343 Hz. Even when the peak is removed by equalization, the H baffle should only be operated below this frequency. It is a compact baffle for woofer applications and I use it with slightly different driver arrangement for the PHOENIX.
Output equal to a closed box occurs at Fequal = 0.17*v/D = 117 Hz with the 20" path difference D between the positive and negative polarity sources at the H baffle openings. -linkwitz labs

..and did not notice the "deeper H frame analysis" that shows the correct measurement technique for D. So, I thank you for straitening me out. Also, I know this was not the intended purpose of this thread, so thanks just that much more.

-Gary

[oneoldude]

gvisom - that is why we are here. :T

AJ and Dennis,

Good points. You are absolutly right. The resonance peak will be around 425 Hz.

From S.L., "The resonance occurs when L = (wave length)/4 = 0.25*v/F."

But a couple thoughts:

Hopefully you can cross substantially below the peak.
And if you can't, the analysis I presented can be used to see what can be achieved with a shorter D.

BTW anybody have a URL for the experiments on this issue done by Monte K?

[gvinson]

quick reply...

AJ - A rearrangement has put me at the edge of nearfield response, Oh what a glory. I'll look to get to 80-300Hz on a smaller baffle with a bigX driver (and stereo subs). Thanks a bunch.

Dennis - I have always had a line of smaller woofers as a secong choice, but until I get past the original design, it has remained a second choice.

Oldude - thanks again.

[Paul H]

gvinson,

The spl claculations noted above make sense, but don't forget they're based on a 1 metre distance, and you lose 6dB with every doubling of distance.

However, with an 80Hz crossover your design goals are much easier to achieve, & the RS driver should do it. Have fun :T

onoldedude,

That Gradient driver is intriguing. It seems some major modelling and experimentation would be required to get 4 tweeters with different crossovers working together properly.

Paul

[Dennis H]

From S.L., "The resonance occurs when L = (wave length)/4 = 0.25*v/F."

Yeah, that's the theory but SL is very clear that the theory doesn't work in the real world. The interaction of the baffles, driver baskets, etc. is too complicated for a simple formula and you need to measure nearfield to find the resonance(s). For example, you'd think the Orion, with the H-frame depth varying between 12" and 16", would have a resonance much higher than the measured 200 Hz where SL sets his notch filter. Using the old rule of thumb to stay an octave away from problem areas, SL's 120 Hz crossover is about as high as you can go with the Orion.

[oneoldude]

Now here are some pics from Slobodan P. This is a guy with great taste. And he builds beautiful dipoles too.

[maximumshow]

I hate all of you guys! The extent of my hate is directly proportional to the size of your sound systems.



I'm only 23, but I can't imagine ever having anything like these :cry:

[Beau]

What sort of woofer is that in the second picture?

[ThomasW]

Slobodan was supposed to send me info about all his designs for the B&G gallery pages but he never did...

Here's another one of his designs

[ThomasW]

Gee I suppose I should post a picture of my mains. There's a newer 3-way version with a RD75 as the midrange but it's so kluged together I embarrassed to show them...