Another Sensitive OB Woofer Design

[cuibono]

Originally Posted by fwater

Dammit, somebody always beats me to it!

I bought a pair of the GW-210/8 a few years ago and messed around with a W frame with mixed but acceptable results. For a long time I figured that this arrangement was probably the best for low-end extension with these drivers, short of a very wide standard OB, and lost interest, finding it kinda a novelty.

I revisited these drivers armed with a little more knowledge, an inspiration from the ultra-exclusive "Whisper" (or the newer Helix http://legacyaudio.com/index.php?opt...=65&Itemid=195), and 8 of the 4 ohm versions. The results are surprising in the amount of output and extension (within reasonable expectations), and un-surprising insofar that the claims of the Whisper's arrangement and it's effect on bass are probably true. The sound is quite good and the low cost makes it very attractive for me to look deeper into this configuration.

I'm not the first to the plate, but I'll never be armed with hard numbers, either...

Whoa, thats a big speaker. I doubt its "13 inches" deep though..

Originally Posted by navin

From the pics what I see is 3 drivers facing in one direction and one driver facing in the other. Now unless the drivers are wired to compensate for this I am not sure how this works.

Maybe the angle the pictures are taken does not provide a complete view.

Any idea on what specs should the drivers be for this application? How does this system compare in Max SPL and F3 to a system using the same driver in a sealed/ported box? I would assume that a single driver in a sealed box would produce as much Max SPL as this combination of 4 right? What we are gaining is better quality of bass.

Also is the radiation pattern similar to that of an H frame? Can such a woofer be placed in the middle of a room or in a corner?

So many quyestions! :-)

Sorry its not clear, but two drivers face forward, and two face backward (and are wired out of phase). I'll take some more pictures today.

In terms of specs, its not too complicated, but I use MJK's mathcad worksheets, along with some help from John K's Dipole Design software, and a lot of background from Siegfried Linkwitz. The primary specs of interest are Qts (values around 1 mean higher efficiency in the low end), the volume displacement (Sd times Xmax), and Re, the DC resistance. There are others that matter too, but those were the first I was looking at.

Yes, a boxed loudspeaker would produce more sound. I feel the quality of the sound out weights the additional drivers. I don't know about specifics though (max output) - but it is definitely easier to make OB baffles than traditional boxes!

I assume the radiation pattern is similar to an H-frame horizontally, although I'd have to measure to know for sure. Vertically, its similar too, up to about 400Hz. Generally, OB speakers have to be at least a few feet from the walls (the more the better), and I definitely wouldn't put them in a corner. Remember there is an out of phase backwave that we don't want coming to us immediately.

[cuibono]

Next part of the adventure:

I've been using these a while, and am considering some changes. First, using separate amps is causing ground loop issues (high noise levels). I don't think the extra amp is justified anymore - a single 50W amp can get things pretty loud (should be easily able to do over 100dBSPL/1m at 50Hz considering both speakers and music as a source). Removing the second amp and wiring everything in series/parallel results in a hit of -6dB to the stated sensitivity. This could be avoided by using the 12" versions instead of the 10", and using 4ohm versions instead of 8ohm.

The other thing I'm considering is adding an inductor to the lower woofers. This should push up the upper possible crossover point some. I'm thinking of 100-120Hz as the roll-off point. But if the woofers are wired series/parallel, an inductor will be seeing a 16ohm load, and require about a 20mH inductor. This is relatively expensive ($30 a piece), almost the cost of the second set of amps. Maybe not part of a budget design. So I'm not ready to pull the trigger on that yet.

Keeping things how they are, I could use a smaller inductor, or just add a cap across the input of the amp of the lower woofers. Hmmm...

[cuibono]

Yes, four of the Goldwood GW-212/4 wired in series/parallel should give the same voltage sensitivity as the above designs at 50Hz, while using only one amp. The enclosure will be a little bigger, but that might be okay, the current enclosure is fairly modest.

[Bratislav]

I think this thread needs a bit of a bump, in light of how spectacularly successful Cuibono's dipoles turned out. Considering how small this section really is, yet reaching 30Hz -6dB in free space (perhaps even lower in room!) this is remarkable, if not outright unbelievable, especially if we mention the price !!!

[cuibono]

Thanks for the bump! I really enjoy sharing these designs.

I am pretty happy with the woofer design - it sounds really solid, when everything comes together. Although 30Hz/-6dB is pretty good, the nonlinear distortion begins to rise below about 40Hz, IIRC. I'll build sealed subs at some point, and probably cross them around 40Hz or so. But all in all, this seems to be a pretty solid design. It needs little EQ, which is nice too - particularly no dipole roll-off compensation, which is usually the big one.

I guess I didn't finish the design really - I never posted any dimensions or anything. I'll probably do a google sketchup one of these days. My impression is that the dimensions are not particularly critical - one could probably change things by an inch in any direction, without too much changing.

[CLS]

Originally Posted by cuibono

...

Generally, OB speakers have to be at least a few feet from the walls (the more the better), and I definitely wouldn't put them in a corner......

It'll be interesting if you try this:

[Andrew Eckhardt]

CLS, I don't know what you used to generate the graphic, but I think the reality of the situation will have more pronounced squashing of the central + lobe due to the increase in intensity of the - side lobes because of the near rear boundary. If the graphic was done by proper calculation using physics law, I eat my text.

[CLS]

Take it easy, it's just a rough sketch to show the idea only (Powerpoint saved as jpg). As you can see, there's no numbers / coordinate axis or the likes on that sketch

Oh, both the (+) and (-) lobes are 'loaded' by the near boundry, aren't they? I added a rear wall on the speaker (not only the 'brick wall'). It's not there in the original design.

[Andrew Eckhardt]

As an aside, I recently made a set a of dipole PC speakers using the square basket PE 10's, and they work great with some EQ. I know EQ isn't what the flea power crowd wants to hear, and these square drivers haven't been in fashion since the early 80's, but they spec. out great if you come to appreciate a more highly damped sound at the usual frequency / transient response tradeoff midpoint of natural Q=.7. They have bigger magnets, longer coils, and higher excursion compared to the Goldwoods at about the same money. The Fs is nice and low. They are really good drivers for moderate power and low crossover points considering the cost. I will say I had to add glue to large sections of the surround right out of the box, other than that build quality is slightly better than the GW215s I bought years ago, where the surround was glued well enough but quite a bit out of round. Now if only we could get someone to make drivers of similar design with shorting rings added, a finishing touch that only the OB crowd would want in such "wimpy" speakers.

[Andrew Eckhardt]

Originally Posted by CLS

Take it easy, it's just a rough sketch to show the idea only (Powerpoint saved as jpg). As you can see, there's no numbers / coordinate axis or the likes on that sketch

Oh, the (-) lobes are also 'loaded' by the near boundry, aren't they?

If you mean the driven impedance would fall then yes, but this works only because the intensity is higher for a greater volume around the driver, which is why it would help to better pinch off the desired lobe, a real -.