[Graham Maynard]

Hi David,

Yes Mr Briggs went on to use 10+12 on his commercial sand-baffle loudspeaker, with the upward facing rear tweeter.
His drivers had reasonable LF for their time and the long baffle dimension was directly along the floor, so probably a reasonable performer.

I run my drivers in parallel.  The larger one has a higher inductance voice coil plus slightly lower Fs, so it is as if the large driver has an additional series choke network with roll-off in the upper bass.
Yes both my LF drivers are driven by a single T-bass per channel, but then my amplifier is one ohm capable for driving both into X.lim.

I too have thought about a boxed or tuned LS for sub only, but I am getting such good output down to circa 25Hz without, so do I really need something for the 20Hz region which will only end up cracking my ceiling again ?
(I might have reached the point where subs are not essential because they cause more problems than they solve, but not had a chance to work on my speakers for a while now, so I cannot say whether this is a final decision; besides I am driving from a flat (amplitude and time) amplifier output without EQ anyway and thus I still have room for a little extra lift circa 20Hz !)

Your LF OB might stand some rear 'U'ing because there will not be any reflective planes perpendicular to the baffle.

Good Luck with the project.

Cheers .......... Graham.

[D OB G]

Hi Graham, 

Thanks for your input.

There is another reason for the subs, an attempt at some sort of version of Geddes distributed subs, admittedly a poor one (I've been reading to much!!!).  There is always scope for more!!  And I still reckon, from walking behind them, that I'm getting some sort of cardioid response i.e. exciting rear wall reflections less. 

And also, because I can dial in the relative contribution, organ pedal notes can be felt in the chest, and they play tunes, but there is not that same "cracked ceiling" type pressurization that you talk about.  (If that makes sense!).

I have tried "U"ing in the past, and for the small extra extension (really just an improvement in efficiency from lower Fequal), I think there may be a deteriment to the sound???  (It's not that need I them for structural integrity).

The biggest improvement I have made lately is to scallop out the open-cell foam so that it is parallel with the waveguide surface (the XT1086 is much smaller than Geddes' wavegfuide) i.e. about ~2.5cm, 1", thick all over, including over the throat covering the phase plug (which had already had the "bug screen" removed), rather than having it flush with the front of the waveguide, as Geddes does.

The missing treble detail that I had complained about has been restored, and then some!!, with, I belive, the HOM reduction still being effective.

I really recommend that anyone using the foam plug, who has the option of making a comparison, should give this a go.

David

[Graham Maynard]

Hi David,

I'm following all your reasoning there.
Just musing on related aspects here -

The Geddes distributed subs might reduce waves due to room dimensions but still represents a room volume pressurising system.

By definition cardioid is a point source phenomenon.
John K has studied audio LF cardioid, but I do not know if a genuinely useful working solution has yet been produced.

Two most important requirements for cardioid generation is matched pistonic displacement, and coherence of cone motion between drivers.  Due to falling OB radiation a genuine cardioid response cannot be generated having flat SPL.

Now that I have my OB working so well at low frequency I would be loathe to introduce something which is going to take me back to a situation I have made such effort to escape from, ie. pressurisation of the listening room itself which is the most significant problem at low bass frequencies.

A monopole pressurising source driven by an electrically phase coherent source having rising 12dB/oct amplitude characteristic with falling frequency;  separately driving a *small* sealed sub enclosure only at frequencies below system resonance in order to maintain phase coherence, ie. resistively loaded = no resonant transduction peaks nor any phase changes wrt to the OB driver already within its 12dB/oct falling SPL radiation characteristic;  with the air volume displacements between dipole and monopole matching at between say 20 to 25Hz, is something I might try in 2009 (health and circumstances permitting).

With a subwoofer arrangement working over such a narrow bandwidth frequency range, say 15 to 25Hz, it might well be possible to obtain good subjective results via a less than objectively perfect separately EQed plate amplified small sub behind or beneath the dipole.

Which might well be what you are already checking out, though with the system drifting between monopole and dipole about only one true cardioid frequency ?
(Like a Home Theatre OB!)

Cheers ....... Graham.

[D OB G]

Hi Graham,

Yes, I'd say that's what I'm checking out.  It's a pity that I can't roll the sub off any lower than 40 Hz (Need a couple of the old REL subs!), but the doubling up over nearly an octave might be an advantage- not sure, but I certainly see your point.

Also, what is happening to the discontinuity to the power response where the CD horn takes over from a truly dipole midrange? I tried "semi-permeable" draped carpet behind the midrange, so that by the time the horn freq is reached, minimum back radiation occurs.  That can be achieved, but it doesn't sound as good!!  At least there are less wayward high freqs to deal with. I have personally never had succes in any room of any size with back firing tweeters, even though the power response would be improved.

So I have got monopole for a few very low freqs (where OB cannot work), some sort of not truely cardiod response (I can dial in the best match of monopole with dipole- so I'm not really pressurising the room over that band) for less than an octave, merging into true dipole, until constant directivity.  According to Toole, apparently, the single biggest determinate in double blind preference for loudspeakers is very flat and extended on listening axis freq response, which I can achieve very well with DEQX.  The second is the the speaker with constant directivity over the widest freq range. I have some of that, and OB natually achieves that.  Never mind Toole, us OBafflers know the virtues of OB midrange and bass at least.

So for all the questions I have tried to answer, I still have a heap more!!

David 

[Graham Maynard]

Just laughed at your last comment - as if we expect our solutions to end all problems ?

Cheers ........ Graham.

[Graham Maynard]

Also, what is happening to the discontinuity to the power response where the CD horn takes over from a truly dipole midrange? I tried "semi-permeable" draped carpet behind the midrange, so that by the time the horn freq is reached, minimum back radiation occurs.  That can be achieved, but it doesn't sound as good!!  At least there are less wayward high freqs to deal with. I have personally never had succes in any room of any size with back firing tweeters, even though the power response would be improve

Good question David.
More *OB* thoughts.

A midrange is not dipole above say 500Hz anyway because front and rear driver radiation patterns are so different.
So can a rear facing CD/tweeter match the rear output of a cone driver?

'Wayward' high frequencies are cerebrally recognised as room, and not reproduced performance, characteristics.
Over damping of delay reflected rear radiation dulls enjoyment in the cerebrally recognised listening room due to sound characteristics clashing with the mental image.
Using a front CD gets rid of even more 500Hz to 5kHz rear radiation.

As much effort expended on correcting the coherence and dispersion of a cone driver as is on CDs could obviate the need for crossover to a CD from the cone driver, except for high frequencies >15kHz where cone driver voice coil impedance goes high.

This is why I stick to an open frame full-range as the main dipole driver, and I *empirically* correct the rearward shift of its acoustic centre and the increased beaming/HF focussed peaking with increasing frequency so that my tweeter needs no more than say 0.33uF of single capacitor crossover to make the top HF more fluid.

I attach a forward pointing paper (2.5") cone in place of the dustcap, this reproducing superiorly to any other centre or phase plug or whizzer, and use foam shapes in front of the cone until the centre beam is dispersed and becomes audible at wide listening angles, all this with the driver input held electrically (coherently) flat.

Cheers ........ Graham.

[D OB G]

Hi Graham,

A few thoughts/notes.

I belive the PHY may be dipolar above 500 Hz, due to it's very open basket, but as you say, the actual pattern will be different with higher freq. (at 1200 hertz, a wavelength is still almost one foot).  I hope to get Praxis out today!

When I have used a rear facing driver, it has been a rear esotar to match a front esotar, at the same crossover, but sometimes padded down (a la Linkwitz).  This is where the power response would approach equality, and Linkwitz called it a revelation, and others love it too, but for me it has never worked.

As far as CD versus OB crossover freq point is concerned, I discovered that the waveguide is constant down to 1200 not 1600 Hz.

So I have had the chance to play with crossovers between 1200 and 8 kHz now.  The PHY is very flat (no rising response) to almost 9 kHz.  The speaker sounds good crossed at 8 KHz, but there is something about the way it loads the room that makes 1200 CD sound quite a bit better.  (i.e. getting rid of 1200 to 5 kHz rear radiation is not a problem). Big difference between a beaming 8" (even though it has a pole-piece) and CD.  Maybe this is why Lynn Olson is prepared to go from OB to horn at such a (relatively) low freq??

I think I understand your rationale Graham, but correcting a cone for CD might require the sort of design that Romy the Cat has come up with!!, or maybe you are achieving it with your modifications?

Regarding full-range drivers, I did try it with a Fostex FE206e (which in my opinion is the worst driver I have have ever heard- so not a valid test), and an Audio Nirvana, which wasn't a whole lot better.  I think your avoidance of a whizzer cone is a good move.

David

[D OB G]

P.S.

I just remembered that the Hawthorne speakers lose OB above about 1.5 kHz???, and they seem to be well received as OB speakers.

[D OB G]

P.P.S.

I've just measured the rear radiation of the PHY up to the crossover of 1200 Hz.  It's about a perfect overlay of the front at the same distance. i.e. I do believe that I'm getting essentially dipole radiation up to the crossover (slight discrepencies might just as well be due to microphone inconsistency).

[JoshK]

I would guess an 8" speaker normally gets dipole radiation to about 1k and more with a more open basket.  Jon Marsh measured the SEAS w22 to be dipole up to around 1k, maybe a bit more.  I think SL did some measurements too.

[D OB G]

Hi scottw,

I think you were right about crossing the PHY-HP higher (I think!).

I have been concerned about a lack of detail with the DE250, and even been on DiyAudio for help.

I've tried a number of approaches with the stuffing of the waveguide, from Geddes' technique, to none at all, to a scrim bag loosely filled with long hair wool and held in place with one layer of the open cell foam.

I'm now crossing at 5.5 kHz (for the moment!).  It's now clear that in the range of 1200-5500 (at least) the PHY-HP has more detail than the waveguide/compression driver.  And there is no loss in dynamics.

I've gone to a degree of mismatch in polar response at crossover, but with a much more even power response.  That's the trade-off.

What were you saying Graham about 5000 Hz?

David

[Graham Maynard]

I'm convinced that constructors are going for a CD instead of cone because they are not making enough efforts to get the best out of their cone drivers first.

Cone driver radiation is affected by dimensionally modified characteristics, and these effects must be countered/controlled via as much effort as is used to develop horn shapes.

Maybe the PHY cone drivers are better from outset, for precious few cone driver manufacturers still have this start-to-finish understanding at hands-on level ?

Using an electrical crossover on any driver is tantamount to changing the dynamic shape of its cone because a med/high electrical drive split cannot be reconstituted after the individual components have been transduced by drivers having quite different characteristics. 

SPLs might be congruous, but that which is *musical* won't!

I firmly believe that the main OB driver should cover as much as possible from 100Hz to 10kHz , whether it be via a cone or a thin panel, then add appropriate and simply filtered augmentation both below and above. 

This precludes the use of a CD which needs an electrical crossover right in the middle of that range - the very frequencies where we are so sensitive to modified waveform dynamics and harmonic relationships due to different components being launched in fractionally shifted time frames so that they reconstitute differently within our ears.

Cheers ......... Graham.

[D OB G]

I've never liked rear facing tweeters.

While I wait for an 18sound NSD1095N to trial in the waveguide, I thought I'd try another experiment.

I have rear facing Esotars, in dipole polarity, running full out in parallel with the compression driver i.e. they are padded down relatively due to the lower sensitivity.

This will improve the power response, but to my surprise more detail is retrieved from the music. (Linkwitz described it as a revelation when he did it- 2 to 3 dB down I think).

I have measured the PHY-HP  as dropping off at the rear at 3 kHz (surprisingly high), so there is scope to try a lower crossover again (!!!), maybe optimising both polar response and power response at the same time!! 

Sounds good now, so we'll see.

David

[D OB G]

I'm now obtaining full detail from the DE250s (nothing wrong with them it seems) by operating a true dipole over the full freq range.  (The fault seems to have been an uneven power response).

I have a second DE250 in another waveguide (also constant directivity) pointing backwards and upwards, in dipole polarity.

With the PHY now operating to 2600 Hz, I am getting equal radiation front and back.  With the DE250s in parallel (no padding down works best), with a flat on axis freq response I am getting  a flat(ish) rear freq response, and something along the lines of an even power response, and good, smoothly changing polar response.  Seems all these factors are important (much along the lines Toole has found).

Certainly the detail is there now, but the imaging is also improved.

Speakers are two meters from the back wall, but one side wall is unfortunately too close.

Turns out the lower sensitivity of the Esotars just wasn't enough.

I've tried the same configuaration crossed over higher, but then there is a hole where the back radiation of the PHY runs out of steam, before handing over to the rear DE250.  It's noticable.

The PHY is good to 8 kHz (front radiation), but it just sounds better in this configuation, which surprisingly is more "integrated", and more "musical"!

David

[Viridian]

Hi David,
I am very interested in how you are integrating your EMI 1550's into your setup using Graham's TBass circuit. Have you been able to further tune the circuit? I would be very curious to see what component values you ended up with.

Also, perhaps you or Graham could kindly enlighten me: I had understood generally that a higher Qts driver with consequently higher Qes was more suitable for OB Bass. But in reading some of your comments on other threads I see that you have both very deliberately chosen low Qes/Qts drivers. Would you care to elaborate? I do understand from Graham that a lower Qes driver is capable of greater dynamics due to less stored energy. Correct? This lower Qts preference seems counter-intuitive to me in striving for lower FR OB bass.

Thanks,
Erik

[Graham Maynard]

Hi Erik,

Low Qes drivers allow an amplifier to exert good control upon cone movement.  Thus a SS amplifier will damp driver resonance even if it has a high Qms, also it will damp enclosure resonance.

A high Qes driver cannot prevent a high Qms driver from resonating no matter what the drive, and thus bass becomes louder but is less musical because it is mechanically amplified in driver 'time' !

Impedance in series with a low Qes driver will allow it to resonate too. 

The T-bass however has highest output during a first half cycle and then damps even though the output impedance becomes raised.  So the driver does not get to resonate and store energy as when it is directly connected.

Thus the T-bass driven driver has more punch and cleaner reproduction.

Cheers ......... Graham.

[D OB G]

Hi Erik,

Because I'm using DEQX, I have an "infinite" means of integrating the OB woofers with the boxed subwoofers.

Before I used the subwoofers, at one time I was just using one driver for bass.  The T-bass was very effective and adjustable.

Now that I'm combining OB with box, I'm experimenting! with driving them directly from the amp.  Because I'm going flat below 20 Hz the sound is different (ambiance, decay, impact e.t.c.) to when the OB went down to 35 Hz or so (basically another octave).

At the moment I'm trialling an extended listening period without the T-bass, just to go back some time to where it used to be, with the OB contributing, maybe, 80% of the sound at 35 Hz.

It's sounding very good, and I'm not ready to make a comparative decision yet. (But I emphasise that without the DEQX I would be using the T-bass).

David

[HiFiNutNut]

Taking out sensitivity, SPL (dynamic capability), frequency response, power response and polar response, how do you subjectively compare the Esotar to the compression driver in terms of SOUND, such as details, sound characteristic, speed, transparency, etc?

I am tuning my OB speakers with the Esotar T330D and found it no better than the Seas Excel Millenium in terms of sound. The Esotar is not very flat so it adds colourations to the sound, but fortunately the colourations are not offensive, such as the dip at 5-7kHz, mild elevation from 9kHz. The first one can not be removed from the crossover.

[Viridian]

Graham and David,
Thanks for your replies. I am still saving pennies to be able to buy the necessary toroids and will soon be posting my impressions in the appropriate thread.

Graham, what you say explains something for me which I haven't been able to sort out until now. The fellow who gave me my 4 Goldwood 18's replaced them in his OB's with Dayton 18's which were much lower Fs: 25Hz, Qes .23 and Qts than the Goldwoods. His impression, listening with B200 on top and with a Decware Taboo amp was that the Daytons had much cleaner and more impactful bass. Sound familiar? Of course those drivers were also much more $$. I always kind of thought he couldn't be quite right because everyone knows that a high Qts driver makes better bass in OB. Just learning here, see? I should now tell him about the TBass circuit, well perhaps after I can personally recommend its benefits.

Of course I will try my Goldwoods but now the two of you(especially Dave with his EMI 1550's) have me thinking along different lines. Like selling all my extra stuff to buy some 1550's.

Thanks,
Erik

[D OB G]

Hi HiFiNutNut, (better make that hello!)

I don't have any print-outs of my Esotar measurements, but I do have the factory data.

It shows a 1.5 dB dip between 5 and 6 kHz, and flat from 2-5 kHz, then flat, but elevated 2 dB, from 7-18 kHz, with a slight drop to 25 kHz.  (So pretty much exactly as you measure them!).

The dip is present at 30 degrees, but not at 60 degrees.  At 15 degrees off axis they are pretty much flat, with the exception of the dip.

When I was first told about the Esotar (a while ago now), I was told that they were very "electrostatic" sounding (which is a positive for someone who used to own a pair of Quad ESL57s!).

I found that to be a good description, but with a dynamic range missing from electrostatics.  To me, they sound open compared to other domes I have heard (Scanspeak, Vifa, Peerless), but I haven't heard the Millenium.

As I have mentioned on DiyAudio, my first experience with the compression driver was not favourable- not as open as the Esotars, and not as detailed.  The compression driver did have more dynamic range.  But this was solved when I used the rear facing driver as well.  I'm now convinced that Toole's recommendation to maintain as even a power response as possible is important, right up to the top frequencies, and that it does correlate with detail and openness (as SL found).

So, to answer your question, the dipole DE250s are no more open, perhaps less, than the Esotars, but are more "solid" (I don't know how to describe it, but it is a positive).  The off-axis response is obviously more even with the waveguide.

I'm building a pair of spherical (box) speakers for someone, and I'm going to use the Esotars (I've got a quite a few) as the best option, partly beacause of its "small" size (I have cut a slice off the faceplate), but also beacause I believe that the compression driver in a waveguide, only radiating to the front, is not as good overall.

Using a passive crossover, the compression drivers need much more complex equalisation (which isn't an issue with the DEQX). This could be a big factor in making a choice.  I am under the impression that commercial OBs with compression drivers, such as the Hawthorne e.t.c., aren't really "flat".

David