Some design ideas for Dipole/OB needed?

[Mayhem13]

Re: Some design ideas for Dipole/OB needed?

I should just ask Zilch to mount the Ewave deluxe on an open baffle for me and see how he likes it! and of couse how it measures. I'll work about the low end later.

[ocdSCHACK]

Re: Some design ideas for Dipole/OB needed?

Clearly you know that ORION uses two Peerless 10" XLS per speaker, and has no drop in bass . . . they're flat to at least 30 Hz in my room (though they run out of excursion on explosions and dinosaur footfalls). Since your "H" frame has roughly the same dimensions I'd guess the drop you see is an equalization issue, or a room issue.

When I say the drop at 100hz, I'm referring to the speaker and baffle's natural tendency to roll off. I didn't mean it couldn't be eq'd to play lower, only that the efficiency of the speaker starts to erode much faster below 100Hz. Cutting off the speakers closer to 100Hz(instead of 30Hz) and using a sub allowed much higher SPL levels and no excursion worries.
My point was: using a sub helps eliminate the need for an expensive EQ with multiple filters and can simplify the build without losing the greatest benefits of an open baffle.

[martin]

Re: Some design ideas for Dipole/OB needed?

If i might ask.....consider your explanation of a midrange/fullrange driver on top.......as a comparator or equivilent would a waveguide loaded dome tweeter that holds directivity to 1khz or so do the job?

I really don't know the answer to your question. I have never used, measured, or even spent much time listening to a waveguide. It is something I would like to look into but just have not had the time.

My approach, as I am sure you are aware, is to mate a large high Qts woofer with a small full range driver crossing at about 200 Hz. This has the advantage of low cost (< $400) and ease of design and construction. The full range driver is open at the back so the dipole pattern should extend high in the audio frequency range. Recently, I have been using smaller and smaller full range drivers. It looks odd to mate a 15" or 18" woofer with a 2 or 3 inch full range driver, but it seems to work well in my music only sytem.

Sorry not much help.

[markk]

Re: Some design ideas for Dipole/OB needed?

My wife has promised a bit of free time for me on Father's day, so hopefully I can crank out a new set of baffles for the RS225 dipole/waveguide.

A couple of observations.

You have to make the distinction between "open back" and "dipole." The terms are often used interchangeably, but they aren't really the same. As Martin said it's about tradeoffs. There's more to open baffles than just dipole radiation, although that is very important. How important is for you to decide. Most speakers don't have perfect polar response, heck, the orions don't. If you really want to have dipolar response, you need a narrower baffle, active eq, etc. If you relax this requirement you can get the other benefits of open baffle and easier/passive construction.

I haven't heard a speaker yet that didn't sound worse right up against a wall, open or closed. I don't think the requirement of some clearance should dissuade you. Dipole/open back speakers still sound good even a bit closer to the wall. (Of course, you have to have some room. My flush mounted dipoles never quite got to market...)

The rear tweeter. This is a tough one. SL experimented with a rear tweeter and it didn't work for the phoenix. JohnK's mini doesn't have one, but the NaO does. Theoretically it makes sense to have a rear tweeter, but practically speaking the difference is small (imho). Again, don't let that dissuade you. The conundrum I have is that mounting a tweeter in a waveguide will improve directivity in the forward hemisphere but makes adding a rear tweeter problematic. You can just mount the tweeter without a waveguide and add a rear tweeter, but the directivity will vary more in the 1-2k range in both the front and rear hemispheres. So it's a tradeoff.

[martin]

Re: Some design ideas for Dipole/OB needed?

If you really want to have dipolar response, you need a narrower baffle, active eq, etc.

Why do you state that as a fact? The baffle lies on the plane of the null, the width will not determine of you have a dipole response. Equal sound radiation from the driver's front and back will make the theoretical dipole figure eight response in any size baffle.

[Mayhem13]

Re: Some design ideas for Dipole/OB needed?

My wife has promised a bit of free time for me on Father's day, so hopefully I can crank out a new set of baffles for the RS225 dipole/waveguide.

A couple of observations.

You have to make the distinction between "open back" and "dipole." The terms are often used interchangeably, but they aren't really the same. As Martin said it's about tradeoffs. There's more to open baffles than just dipole radiation, although that is very important. How important is for you to decide. Most speakers don't have perfect polar response, heck, the orions don't. mounting a tweeter in a waveguide will improve directivity in the forward hemisphere but .

This is where i'm concerned...that the high directivity is a camp too far from dipole to walk to...and as such becomes oil and vinegar....on their own their fine but they don't play well together...from an SQ point of view. Thus the discussion as we're talking subjective opinion vs measurements. The waveguide tweeter can do flat FR and hold directivity to 1khz or so and looks great off axis out to 10hz when 0,10,20,30 degrees are averaged......theoretically excellent from a measurment standpoint. But i think you would want to match that directivity and i'm concerned that's where this discussion falls apart. In a box and on a baffle the 12" midwoofer's directivity can match the tweeter through the XO range with an LR4 slope but what happens when we release that rearward sound? Do we lose the directivity? And what about the dipole peak.....do we have to maintain directivity to that point?....i don't think a 12" could do that even in a box and large baffle........

[Mayhem13]

Re: Some design ideas for Dipole/OB needed?

I really don't know the answer to your question. I have never used, measured, or even spent much time listening to a waveguide. It is something I would like to look into but just have not had the time.

My approach, as I am sure you are aware, is to mate a large high Qts woofer with a small full range driver crossing at about 200 Hz. This has the advantage of low cost (< $400) and ease of design and construction. The full range driver is open at the back so the dipole pattern should extend high in the audio frequency range. Recently, I have been using smaller and smaller full range drivers. It looks odd to mate a 15" or 18" woofer with a 2 or 3 inch full range driver, but it seems to work well in my music only sytem.

Sorry not much help.

Actually, you've been a big help Martin...thanx. Your approach to some degree was a consideration before starting this thread. The first 'plan' was the use of a parallel pair of Alpha 15's....it's the top end i wasn't able to figure out. For my goals, i can't sacrifice all that available sensitivity and SPL to a small fullrange driver that'll simply explode when called upon for HT use....or am i missing something. There's a reason for choosing such a low XO point.......mind sharing? It's gotta be critical or i doubt you'd make the tradeoff either. In the case of the Alpha's they're good till at least 400hz with and 2nd order filter and higher with a steeper slope.

[Flint]

Re: Some design ideas for Dipole/OB needed?

Just to expand a little on MarkK's coment about a real dipole vs. an open back: this is what I was saying before about the "state of the art" in dipole design. Drivers operate in baffles not much larger than their own width. As Martin points out it's more expensive and complicated, requiring active EQ to make it work. As far as I'm concerned this shouldn't be a barrier for a serious DIY endeavor, because there are digital tools for all this.

Here's a suggestion for a "true dipole", and of course this is complicated and expensive: use a really small tweeter (Vifa NE19) and really small midrange (SS 10F) at the top of a 4-way system. If someone else is buying, how about an Acoustic Elegance Dipole 12 (or 15) to get down to a closed box subwoofer. The baffle needs to taper dramatically. You can't do much of an H or U baffle on the mid-bass because it needs to extend high, like 500 Hz (based on MarkK's testing of the 10F midrange).

Forget the rear tweeter, but you've already got rear output much higher than the original Orion, because the M/T cross will be 3-4 kHz. The only reason not to add the rear tweeter is the geometry. It just doesn't fit and having it stand out behind the baffle will not work well (I've tried this but with a different tweeter).

Just to upset Deward I'll make this bold claim: a speaker with this design could totally outperform the Orion system.

[martin]

Re: Some design ideas for Dipole/OB needed?

The first 'plan' was the use of a parallel pair of Alpha 15's....it's the top end i wasn't able to figure out. For my goals, i can't sacrifice all that available sensitivity and SPL to a small fullrange driver that'll simply explode when called upon for HT use....or am i missing something. There's a reason for choosing such a low XO point.......mind sharing? It's gotta be critical or i doubt you'd make the tradeoff either. In the case of the Alpha's they're good till at least 400hz with and 2nd order filter and higher with a steeper slope.

I am merging the electrical crossover slopes with the dipole hump. For the low frequencies, I set the electrical crossover low and then the dipole hump pulls the acoustic crossover up in frequency. The electrical crossover might be at 200 Hz yielding an acoustic low pass crossover at 400 Hz. Then I do the opposite for the mid range or full range driver, I set the electrical crossover at about 500 Hz and allow the dipole hump to pull the acoustic high pass crossover down to meet the woofer. Locating the crossover straddling the dipole hump is the reason for the lower than expected (as in a typical boxed two way) crossover point, you have to think outside the box. See this article for an example of the method :

http://www.quarter-wave.com/OBs/OB_Design.pdf

So the mid or full range driver is crossed electrically higher than expected at 400 - 600 Hz depending on the baffle size, the demands on the driver are greatly reduced. I have cranked the 1812 Overture into my Goldwood / Jordan system and the full range driver is not stresses at all when the cannons fire. Without the requirement for producing bass, a smaller driver can play loud and clean. If you are thinking two Alpha 15A woofers, then a 6 or 8 inch driver could easily be used for the mid or full range driver.

All of tis can be done with an active crossover and two amps or a single amp feeding a passive crossover. No extra EQ is required with the higher Qts woofers. Cost and complexity are both greatly reduced.

I like simple, I am not smart enough to do rocket science.

[Deward Hastings]

Re: Some design ideas for Dipole/OB needed?

Just to upset Deward I'll make this bold claim: a speaker with this design could totally outperform the Orion system.

You build it, I'll listen . . . and let you know . . . :D

[markk]

Re: Some design ideas for Dipole/OB needed?

Why do you state that as a fact? The baffle lies on the plane of the null, the width will not determine of you have a dipole response. Equal sound radiation from the driver's front and back will make the theoretical dipole figure eight response in any size baffle.

OK martin, you caught me in a bit of sematic sloppiness on my part. In a real world driver directionality as well as the different rear FR (from the basket, kind of an acoustic filter) causes the directionality to change. I guess I should have said, controlled directivity.

See SL's 8554 on baffle



You're right that the figure of 8 pattern is maintained, but directivity is not a consant. So I guess it depends on whether you think that is important.

[martin]

Re: Some design ideas for Dipole/OB needed?

You're right that the figure of 8 pattern is maintained, but directivity is not a consant. So I guess it depends on whether you think that is important.

I believe that the narrow and wide baffle OB systems might have slightly different directivity characteristics. I have seen the really nice SPL plots made outside showing the directional characteristics of some of the narrow baffles systems. But I also recognize that at a 3 m distance in a real room those plots would look a whole lot different, I am not sure that striving for the ultra high directivity goal has a big return on investment in the real world.

[dantheman]

Re: Some design ideas for Dipole/OB needed?

Read this prior to building anything:

http://techtalk.parts-express.com/showthread.php?t=214334&page=3

The 1-4k off axis bump is readily audible. A larger mid/tweeter could be helpful. To get excellence in OB doesn't seem too difficult. That said, the combination of a midbass/WG tweeter might be an incredible option for cheap, but I haven't tried. It would seem to be feasible. A MT or WG(preferably)/WW with the WW crossed over a la MJK could very well be excellent. Have a look at the graphs of just the woofer on that same baffle as in the thread:
on axis:

22.5 off axis:

33.75 off axis:

45 off axis:

Too bad I don't have more off axis data to see what's really going on further out, but it's clear that it's beating the RS100 performance over the same frequency range:
On axis:

RS100 45 degrees off in same baffle:

and another 45 degrees off with foam on the side and no smoothing:


That should help clarify some concepts.

Dan

[markk]

Re: Some design ideas for Dipole/OB needed?

I believe that the narrow and wide baffle OB systems might have slightly different directivity characteristics. I have seen the really nice SPL plots made outside showing the directional characteristics of some of the narrow baffles systems. But I also recognize that at a 3 m distance in a real room those plots would look a whole lot different, I am not sure that striving for the ultra high directivity goal has a big return on investment in the real world.

Hi martin,

This has been the million dollar question as of late. How important is controlled directivity? It seems to be more important than is appreciated. Sean Olive has some research out of Harman showing that controlled directivity is preferred by trained listeners, or at least, monotonically decreasing directivity over designs with a dip and rise in directivity. So I think it's a reasonable goal. I was fairly impressed with the sound of the ER18DXT which has better than average directivity (i.e. better than the typical 2 way 4LR) although my personal N of 1 experiments are of, ahem, limited generalizability...

It's not perfect, but much better than the typical dip, peak of a 2 way 4lr.



It's conceivable that a narrow baffle unit with a waveguide will have very nicely controlled directivity in the forward hemisphere if the wg size and the woofer driver directionality matches up well although the rear pattern above the crossover is a bit of a problem.

[Deward Hastings]

Re: Some design ideas for Dipole/OB needed?

although the rear pattern above the crossover is a bit of a problem

Or not . . . which is an unanswered question. Since the rear radiation is heard only after reflection (from a presumably diffusing reflector) the rear pattern should not much matter, as long as it's power distribution tracks with the front (uniform power response overall). Its phase matters, but only to not produce a null at crossover. I wouldn't worry about perfectly flat, either, since reflection is going to mess with frequency response anyway. I suspect that's also why we don't hear the basket of the midrange . . .

Around 25 years ago I had a Yamaha DSP-1 digital "reverb" processor (actually I still have it in storage somewhere) that synthesised four "surround" channels. I never much liked the overall effect, but I did notice that it sounded better with the front "reverb" speakers *not* on the wall behind the mains (as sugggested by Yamaha) but directly behind the mains facing the front wall. It never sounded particularly good with dipole mains. I didn't pursue it at the time . . . in retrospect it has become more clear why . . .