Some design ideas for Dipole/OB needed?

[Mayhem13]

Re: Some design ideas for Dipole/OB needed?

This is what I did for the 2007 RMAF - AirCirc 6600 in a custom milled 10" waveguide, W22EX open baffle. Crossover was 875hz ... bottom end is a Focal 10W.



I'm not the kind of guy that does hundreds of simulations, or tries to fit some theoretical target design model. I basically build things and see if they work, and I can tell you without reservation that this worked very very well. I've been to RMAF 3 times, and these are one of the best speakers I've ever heard regardless of price ... and I never once bothered to model dipole behavior ;)

IMO the two primary reasons they work so well is 1, the waveguide loading dramatically decreases the mechanical load on the tweeter allowing it cross low and increasing it's dynamic capacity greatly ... and 2, the W22EX is used in a band where it's behavior is perfect because of the lower tweeter cross. I've never heard a speaker that has the soundstage depth and width that these do, and for a big speaker they still disappear completely.

To many people like to obfuscate the design issues on the internet, and never build anything.

Well DT, thanx sooo much for this particular post as it lead me to try my current box design (WG loaded SEAS27TDFC/CA18RLY) over my RSS315HF boxes and all i can say is WOW! The soundstage opened up like the room quadrupled in size! Even the transition to the subs (125hz 2nd order) appears to be smoother as i've lost that Brrrr....pop with male vocals and bass guitar. Now this combo won't give me the SPL i desire...but it's a start...a very good start i must say. I'm looking forward to the 10" QSC guide over a 10-12" pro midwoofer soo very much now. I want to keep the efficiency of the pro driver, so a fairly efficient dome will narrow my tweeter choices to only a few. I don't think i'll be using a CD after all!

[Mayhem13]

Re: Some design ideas for Dipole/OB needed?

After reviewing all of the posts and linked information, for the sake of space and already available gear, my bottom end will be monopole sealed RSS315HF's in 2cuft....basically what i have now but i'll be changing the dimensions for a narrower yet taller and deeper box.

I'll use these with the passive SEAS/WgTDFC open baffle for now but at some point i'd like to use a more robust dome tweeter with a lower XO point to a 10-12" midwoofer. I'll have to wait to see what direction MarkK goes with his WG mounted RS28F but preliminary tests show controlled directivity down to 1.4kz......which is easily matched by a large,pro audio driver. I think the extra sensitivity of the SB29 would certainly make for a more efficient system but just how low can it go?.....who knows. We can assume from other waveguide experiments that the lower operating freq range distortion is dramatically reduced by the WG.

[markk]

Re: Some design ideas for Dipole/OB needed?

John,

Thanks again for the jamo/stereophile link. Amazing how the look of a speaker can influence your perception. And, really, amazing how mediocre is the crossover design. For 10k I'd expect better-most of us reading this board could do better...

Of course, Stereophile can't bite the hand that feeds it. Now don't forget to listen with the grills on

Wow...

[johnk...]

Re: Some design ideas for Dipole/OB needed?

How did you determine that the single sharp dip at about 1.2 kHz in the green trace is due to a wide baffle?

Would a baffle diffraction issue above the hump result in a series of small peaks and dips that are more rounded, a ripple in the response. A sharp null like that looks more like some artifact of a system or driver resonance.

First, it happens to be just about at twice the frequency of the peak. Second, when I run a dipole baffle simulation with dimensions representative of the speaker using either ABC dipole or Sound Easy and a 5.5" midrange I see very similar behavior:



Third, with a 5.5" metal cone driver I would not suspect a resonance "null" at that frequency. In fact I would never suspect a "null" at all prior to a metal cone breakup peak. At best I would expect at most a pre-peak dip. Look at the response data for any Seas metal cone in the classic or Excel lines, for example.

Lastely, I have measured such behavior with wide baffles.

Also, remember the Stereophile traces include the effects (roll off) of the crossover filters. Obviously they don't work very hard on the HP part and there is a clear resonance peak (at 55 Hz) which is most likely a result of resonance of the HP filter series cap with the driver impedance peak.

[DoubleTap]

Re: Some design ideas for Dipole/OB needed?

Nice Job D, love the built in waveguide look like Amphion dows on theirs.
Are these crossed actively to the Focal 10's?

Thanks, it pays to have friends with CNC machines

At RMAF they were actively crossed, but at home I ran them passive. I think at RMAF we crossed them at 180hz.

[Mayhem13]

Re: Some design ideas for Dipole/OB needed?

Nice Job D, love the built in waveguide look like Amphion dows on theirs.
Are these crossed actively to the Focal 10's?

[dlr]

Re: Some design ideas for Dipole/OB needed?

I should have stated my observation as a questions.

That single sharp null does not look like a wide baffle combined with a midrange driver issue, it looks like some form of resonance issue.

So using it as an arguement in support of narrow baffle theory seems a bit of a stretch. I use small drivers on wide baffles, measured a few, and have not seen single sharp nulls. I have observed the OB hump and rounded ripples of decaying amplitude with increasing frequency. My OB simulations show the same types of ripple behavior.

It does appear to be more of a resonance. It looks too sharp to be diffraction (dipole) related.

Small drivers on wide baffles are going to have a good bit of diffraction ripple. Hence my point about small baffles having much less of it, primarily due to driver directivity, especially those that are more appropriate for a dipole, large enough to have enough swept area for the required volume displacement needed.

John's point is well taken. Narrow dipole baffles coupled with a driver large enough for volume displacement and better integration for their directivity will need those larger diameter drivers, even for the midrange, if best advantage is to be taken of the dipole format.

dlr

[DoubleTap]

Re: Some design ideas for Dipole/OB needed?

I'm just trying to find out if the benefit of CD is worthwhile in a dipole design. I could just wait for MarkK to finish his RS225/RS28F dipole but what fun would tha be?:D

This is what I did for the 2007 RMAF - AirCirc 6600 in a custom milled 10" waveguide, W22EX open baffle. Crossover was 875hz ... bottom end is a Focal 10W.



I'm not the kind of guy that does hundreds of simulations, or tries to fit some theoretical target design model. I basically build things and see if they work, and I can tell you without reservation that this worked very very well. I've been to RMAF 3 times, and these are one of the best speakers I've ever heard regardless of price ... and I never once bothered to model dipole behavior ;)

IMO the two primary reasons they work so well is 1, the waveguide loading dramatically decreases the mechanical load on the tweeter allowing it cross low and increasing it's dynamic capacity greatly ... and 2, the W22EX is used in a band where it's behavior is perfect because of the lower tweeter cross. I've never heard a speaker that has the soundstage depth and width that these do, and for a big speaker they still disappear completely.

To many people like to obfuscate the design issues on the internet, and never build anything.

[martin]

Re: Some design ideas for Dipole/OB needed?

I should have stated my observation as a questions.

Wouldn't a baffle diffraction issue above the hump result in a series of small peaks and dips that are more rounded, a ripple in the response? A sharp null like that looks more like some artifact of a system or driver resonance.

That single sharp null does not look like a wide baffle combined with a midrange driver issue, it looks like some form of resonance issue.

So using it as an arguement in support of narrow baffle theory seems a bit of a stretch. I use small drivers on wide baffles, measured a few, and have not seen single sharp nulls. I have observed the OB hump and rounded ripples of decaying amplitude with increasing frequency. My OB simulations show the same types of ripple behavior.

[dlr]

Re: Some design ideas for Dipole/OB needed?

Would a baffle diffraction issue above the hump result in a series of small peaks and dips that are more rounded, a ripple in the response. A sharp null like that looks more like some artifact of a system or driver resonance.

For a midrange unit, a narrower baffle should tend to reduce diffraction issues due to the transition of the driver from omni-directional to directional above the step/dipole peak area. Even for a boxed midrange, all but the smallest units have minor diffraction influence other than the baffle step. The diffraction will be worse for a dipole midrange, but for most drivers I suspect that it won't be a significant issue unless a very small unit is used. As I see it, the dipole need for larger diameters for swept area pushes the use of drivers that won't have much of a diffraction problem other than the dipole peak region.

dlr

[martin]

Re: Some design ideas for Dipole/OB needed?

In particular look at the green trace for the midrange. It shows the draw back of using a small driver on a wide baffle. The driver remains fairly omni-directional above the dipole peak (around 600 Hz) and as a result there is a dip (first dipole null) around 1.2k Hz.

How did you determine that the single sharp dip at about 1.2 kHz in the green trace is due to a wide baffle?

Would a baffle diffraction issue above the hump result in a series of small peaks and dips that are more rounded, a ripple in the response. A sharp null like that looks more like some artifact of a system or driver resonance.

[johnk...]

Re: Some design ideas for Dipole/OB needed?

I would not try to clone the 907. I f you have not seen measurements look at this:



In particular look at the green trace for the midrange. It shows the draw back of using a small driver on a wide baffle. The driver remains fairly omni-directional above the dipole peak (around 600 Hz) and as a result there is a dip (first dipole null) around 1.2k Hz. The mid to tweeter crossover would ideally be around 800 Hz max. This is just another case of where a commercial manufacture pays more attention to what it looks like as opposed to what it measures and sounds like. Here is the measurement of the complete system:



Here is a link to the complete Sterophile article. Note that the elevated low frequency response is due to near field measurement of the woofer, uncorrected for the dipole eq.

[dantheman]

Re: Some design ideas for Dipole/OB needed?

Here's my measurement on the PE 10" WG w/ a selenium D220:

in 11.25 degree steps starting at 11.25 degrees off axis and going out to 90 sitting on top of a box.
This version shows the on axis dip and my first filter attempt:


Dan

[Mayhem13]

Re: Some design ideas for Dipole/OB needed?

http://yhst-93643794936312.stores.ya...p000106gp.html

Currently out of stock according to the page.

Superceded by this:

http://yhst-93643794936312.stores.ya...l000722ts.html

The only difference according to QSC is the addition of the gasket.:D

I know you'd like to the see the RS28F as well Pete!

[Mayhem13]

Re: Some design ideas for Dipole/OB needed?

I'm game to try one of the 10" qsc waveguides. IIRC these weren't that expensive. Is there a link to the qsc 10-probably just cheaper for me to order directly. Does qsc make an 8"?

PM me for the shipping info and i'll send you the 10" round one. I'm currently modding one for an AVS member with the SB29. I wonder if you wouldn't mind measuring up that combo as well?