300 - 1500 Hz: Horn vs Open baffle mids

[slowmotion]

Hi all

Then it seems reasonable to me that a ~one octave larger Unity concept is the no-brainer solution if there's enough space for a 'good midrange horn'.

There's allways space for a good midrange horn

?! I didn't need any EQ with mine (as measured with an RTA/pink noise, not to mention my ears), so I guess I got lucky. Hypex is fine for midbass, but mids require low distortion and good pattern control, so IMO a large conical, or at least an expo radial with a large flare end correction ala' tractrix mouth, L'Cleach, or similar is required. That, or make the horn at least an octave larger than the -12dB XO point.

GM

I've found that I prefer hypex or expo to conicals, also for midrange, but that's just me.
I really like using LeCleach's spreadsheet for midrange and treble horns, it's a good starting point for when you want to design or try out a new horn.

Cheers

 

[rick57]

Hi

{Wasn’t the server down a while? Better donate to maintenance!}

(Starting to read more design material)
Where might I find a copy of LeCleach's spreadsheet?

Cheers

 

[GM]

Yes. http://perso.club-internet.fr/ndaviden/pavillon/lecleach.htm#ancre68820

GM

 

[rick57]

Thanks GM.
Better dust off my Grade 7 Franco-English dictionary, or see if Babel helps.

 

[Tim Moorman]

Exit Flare

GM

Bill's point is the geometry of the driver exit is something that must also be evaluated with regard to setting the dispersion angle of the horn/driver combination. The horn can do only so much to regulate pattern control and, I assume load the horn properly, if the driver exit flare is significantly different than the horn flare.

http://gallery.audioasylum.com/cgi/...session=&&moniker=RCA-fan&invite=&w=367&h=455

Widest exit flare is this Emilar driver.

If you view his other "throat" differences with the same horn and driver, it will give one pause at the amount of difference indicated.

Tim

 

[GM]

Tim,

I believe I'm saying the same thing: "..... flaring it based on the desired HF response first, then adjusting the flare as required as the horn marches down in frequency towards the mouth, so no abrupt transitions till fairly near the mouth.....". Of course Mantaray style CD design violates this, trading distortion for better pattern control.

I'm still curious about the phase plug mods you mentioned.

GM

 

[Tim Moorman]

Phase Plugs

GM

Two different areas of driver improvement are being discussed here: the design of the phasing plugs on Steve Schell/Rich Drysdale's new/old drivers, and the driver exit flare work and suggested matching coupling by Bill Woods(RCA Fan).

Here's a post by Steve Schell about the original WE/RCA designs, including a discussion of the phasing plugs.

The Oswalds Mill annual get togethers over a couple of years brought Steve together with Bill Woods for discussions concerning remaking the vintage drivers using modern methods(now in progress). There are other posts by Steve regarding his work with Rich Drysdale in their quest for the ultimate hi-fi by use of large format midbass compression drivers as well as other remakes of other early driver designs.

http://www.oswaldsmill.com/id29.html

Bill was involved early on and shares the same passion.

http://www.audioasylum.com/forums/hug/messages/81910.html

The measurements he took show large variations in response. In the "throat" post, as he discusses, a coupler is necessary to match the driver exit flare properly to the horn. Then he proves it.

I thought this was really pretty interesting since I had not seen measurements prior to these, with the coupler mentioned, much less installed. Great work on Bill's part.

Tim

 

[Tim Moorman]

Unity Concept w/phase plug

GM,

Here's a design heavily borrowed, perhaps even licensed, from Tom Danley's Unity design with a twist - the phase plug in the horn throat.

http://www.eaw.com/technology/proprietary/va4.html

I not sure I quite understand the reason for time alignment this design variation details, if the horn does indeed load back to the horn apex by the mids, as Tom Danley described in one of his early posts on the Unity concept. I did hear the Unity, but not this version.

Tim

 

[dwk123]

I'm not sure I have a lot to add here, but I'm doing a bit of struggling with where I want to go with my projects, and the horn-loaded/wave-guided mid is part of the problem..

I've been heading down the Unity path, and actually have some not-horrible initial results. I'm considering other approaches due to a) crossover complexity and b) increasing concern that a square-profile flare is not the best idea. Another potential factor c) is that I currently don't really have a system at all, and since available time is limited, something that gets me up and running quickly is attractive.

Despite Rick's deisre to use a compression driver down to 250/300, nothing I've seen indicates that it's a wise or even viable idea. Also, IMHO using a compression driver only for the range up to 1.5k is something of a waste - any driver that can go down anywhere close to 300 will be able to hit 5k or above, which to me would indicate a very different system design.

300-1500 is a classic application for a cone-driver mid horn. I've only skimmed this thread recently so I don't remember why rick57 doesn't want to use a cone, but if you really only have to cover 300-1500 it really is likely to be the best approach. IMHO a B&C 8PE21 on a set of the PI Speakers 90x40 conical flares is likely to be a very decent entry into the game for not a horrible amount of money ($400 total for the drivers and the flares).

Having said that, I have a couple questions.
- has anyone looked at how Earl Geddes' OS waveguide ideas may (or may not) apply to cone-driven mid horns? Since the thrust of the OS work seems to be exactly the task of smoothly mating the driver exit geometry to the waveguide flare that has been discussed here, it's not clear how much applicability it has to a cone-driven scenario. According to the frequency relationship Earl published in one paper, it would take a ~8" throat to properly load an OS waveguide down to ~300Hz.
- further to above. given the large throat of an OS waveguide which makes throat-cancellation a smaller problem, at what point would say an 8" or 10" driver transition from a 'waveguided' radiator to a direct radiator? This would seem to be the limit on how far up you could push a mid, as the pattern control would start to break down here.
- has anyone looked at how using an OS waveguide rather than a straight conical might change the Unity problem? Since the OS geometry will move the 'correct' midrange entry point (as determined by area) farther from the HF throat, it may really mess things up. OTOH, if you put the entry point at the 'right' spot according to the 1/4 wave restriction, then your loading will be all wrong.

 

[GM]

I thought this was really pretty interesting since I had not seen measurements prior to these, with the coupler mentioned, much less installed. Great work on Bill's part.

Thanks for the links! Again though, this is all covered in my previous comment, I just wasn't very specific in how to do it or able to show how much it affects response.

Since I had neither the time or interest in modding a driver to suit a particular app, I designed it all into the lens based on the driver and I made do with whatever BW/power response/eff. I achieved, which was always 'good enough', but I seriously doubt any of them would be considered optimum compared to what I could probably come up with using today's incredibly powerful measurement/simming tools.

GM

 

[GM]

Here's a design heavily borrowed, perhaps even licensed, from Tom Danley's Unity design with a twist - the phase plug in the horn throat.

Hmm, I don't consider this a Unity concept since the HF is so far out of phase with the mids/LF, instead being more like a horn loaded co-ax with the tweeter horn mounted on a crossbar in front of the LF, hence the need for TD.

WRT the Unity, the mids (and LF for his three way) do load back to the horn's apex, i.e. the beginning of the HF compression driver's throat (not exit), same as a Voigt pipe or compression basshorn with the driver(s) mounted along the side of the horn Vs at its apex.

GM

 

[rick57]

Dwk

It’s not that I’m opposed to using a cone, it’s that I’m told (and see why - lower mass) that a compression driver (say a 2nd hand 2 inch JBL) would be better.

The 1500 upper was a staring point, because I was intending to cross actively to a good quality driver that I already own, the largest (1/ 2 / 3?) Aurum Cantus ribbon (101 dB). Maybe if I changed that to say 5 kHz, it’d be better?

I was hoping to get the 2 inch JBL down to 300 or thereabouts, and was having trouble identifying a mid-bass driver. I’ve since found a few that are reasonably affordable that have mass rolloff > 500 Hz, so should easily mate to the 2 inch JBL.

B&C seem a very good brand, though maybe here in Australia nearly as dear as 2nd hand 2 inch JBL(?). If the JBL don’t work out, I’ll look into the 8PE21s, thanks.

You’re definitely right that to get ‘made’ flares could be a very good shortcut.

Cheers

 

[dwk123]

Ah, ok I see your dilema. Mating a horn to the AC is going to be tough no matter how you slice it due to the dispersion issues, but if you go for a 90x40 pattern on the mid horn it might give you a fighting chance.

I heard two systems with 2" CD midrange at the RMAF show, and thought both were excellent - the Edgar Titan (using a JBL) and the Exemplar (using either an Altec, or an Altec repro) So, I wouldn't necessarily disagree that a 2" CD is a better overall choice than a cone, only that getting one down to 300 will be tough. If you can get the lower limit up to 500, then I think you would be in much better shape. However, I am willing to bet that you'll find that running the JBL up much higher than 1.5k will sound better.

OTOH, I guess you don't know until you try. Coda, which is the systems division of BMS, uses the BMS 4592 down to 400 in a PA application, so it might not be too unreasonable for some drivers to make it down to ~300 at home-listening levels. (see http://www.codaaudio.de/codaaudio/products/ap25.html)

BTW - I *really* wish those elliptical waveguides that Coda uses on the AP24 and AP25 were available separately. A 4590/4592 on one of those would be almost ideal IMHO.

 

[GM]

Despite Rick's deisre to use a compression driver down to 250/300, nothing I've seen indicates that it's a wise or even viable idea. Also, IMHO using a compression driver only for the range up to 1.5k is something of a waste......

Not really, unless it's a Unity concept, and then only if it's custom made from a point source driver like the Unity's, which is limited to only 1 kHz IIRC. What's wise/viable depends on the app and/or implementation. That's the beauty of horns, they have so much flexibility.

- any driver that can go down anywhere close to 300 will be able to hit 5k or above, which to me would indicate a very different system design.

Agreed, and not one best suited for a HIFI app.

has anyone looked at how Earl Geddes' OS waveguide ideas may (or may not) apply to cone-driven mid horns?

I haven't yet.

Since the thrust of the OS work seems to be exactly the task of smoothly mating the driver exit geometry to the waveguide flare that has been discussed here, it's not clear how much applicability it has to a cone-driven scenario.

Depends mainly on the CR and desired compression loaded BW.

According to the frequency relationship Earl published in one paper, it would take a ~8" throat to properly load an OS waveguide down to ~300Hz.

Dunno, he's the acoustics expert/math whiz, but I get 7.99" regardless of shape. To me, this means a ~11.3" effective diameter point source driver is required, a non-standard size. Using a 15", I get ~184Hz, 12"/~240Hz. How do these compare with his math?

At least my way has worked well enough for me, but then I didn't have the gear to do really accurate measurements so maybe it's just a coincidence I get the same answer for this example.

- further to above. given the large throat of an OS waveguide which makes throat-cancellation a smaller problem, at what point would say an 8" or 10" driver transition from a 'waveguided' radiator to a direct radiator? This would seem to be the limit on how far up you could push a mid, as the pattern control would start to break down here.

Wherever the driver's DI is > the wall angle. If a point source driver is rigid enough to be pistonic then it can be calc'd, but unfortunately they tend not to be.

>- has anyone looked at how using an OS waveguide rather than a straight conical might change the Unity problem? Since the OS geometry will move the 'correct' midrange entry point (as determined by area) farther from the HF throat, it may really mess things up. OTOH, if you put the entry point at the 'right' spot according to the 1/4 wave restriction, then your loading will be all wrong.

What problem? I don't follow. I mean the mids are well behind the 300Hz 'entry' point. IOW the lens 'feels' as if there's a point source at the 'correct' point along its length since they are exiting in a ~uniform particle density zone for this frequency.

GM

 

[GM]

The 1500 upper was a staring point, because I was intending to cross actively to a good quality driver that I already own, the largest (1/ 2 / 3?) Aurum Cantus ribbon (101 dB). Maybe if I changed that to say 5 kHz, it’d be better?

WRT to mating to another driver or horn of any size, lower is better, though not being familiar with this ribbon I don't know how much XOing at the minimum recommended point/slope affects its tone.

I was hoping to get the 2 inch JBL down to 300 or thereabouts, and was having trouble identifying a mid-bass driver.

Not a problem, you just need to make a 150Hz expo lens (min) assuming at least a 300Hz/2nd order XO. If conical, it will have to be somewhat bigger depending on the wall angle. Of course if you don't have the room......

A waveguide will be required for the AC to acoustically match it, but I don't know how much, if any, loading it can handle.

Unless you get lucky, then done right this could turn into an expensive, time consuming, trial n' error project.

GM

 

[rick57]

Dwk

Yes, more mass produced waveguides in various flares, without the whopping profit margins, would be a very good thing.

Thanks for your suggestion of a 90x40 pattern. Any other views on a good pattern?

GM

> Rick's desire to use a compression driver down to 250/300
. . since changed to maybe 350-450.

> any driver that can go down close to 300 will be able to hit 5k or above . . and not one best suited for a HIFI app

The big AC ribbon is hi-fi, and could be crossed anywhere from 1500 Hz up. I must be missing something - could you explain how it wouldn’t be hi-fi in a horn??

> Unless you get lucky, then done right this could turn into an expensive, time consuming, trial n' error project.

I was only counting on two of those three ; thanks for the warning

 

[GM]

>GM

> Rick's desire to use a compression driver down to 250/300
. . since changed to maybe 350-450.
====
It need not be, it just depends on how much dynamic range/BW from the mid/HF driver is desired.
====
>The big AC ribbon is hi-fi, and could be crossed anywhere from 1500 Hz up. I must be missing something - could you explain how it wouldn’t be hi-fi in a horn??
====
Not the ribbon, but the high, though not high enough, XO point. Since I'm one of those firm believers in not having any appreciable XO overlapping BW in our acute hearing BW, neither 1.5 nor 5 kHz is an acceptable XO point IMO, so the critical mids/HF compression driver would have to be 1" to get the necessary HF BW, with the downside being either a higher Fc and/or settle for less dynamic range if a <500-600 Hz XO point is used. That, or use dual 1" drivers, with one being limited to LF 'fill'.

Of the two, a 300Hz XO point is preferred IMO, so either dual drivers or a Unity concept horn is required, then if the HF isn't extended and/or have a wide enough 'sweet spot' beyond ~12 kHz where the 1" begins rolling off, the ribbon can be capped off super high to add some 'fill' and still keep it attenuated enough so that the huge phasing errors won't be noticed due to our falling hearing acuity beyond ~7 kHz.

WRT polar patterns, 'sound is round', so ideally the room would be large/long enough and shaped like a waveguide so that round waveguides of a slightly less wall angle would cover the entire listening area. This is but a fantasy for most of us, so using square, or at worst low aspect ratio (< ~1.273:1) rectangular designs to keep distortion low is an acceptable trade-off. Drawing up the room layout and figuring the required angles will give you what you need to calc the horn depth/mouth dims required and if you need a high aspect ratio design, then consider rearranging the room to reduce it if that's an option.

GM

 

[rick57]

Interesting ideas GM!
To clarify, when you said:
“or settle for less dynamic range if a <500-600 Hz XO point is used”, at first I read this as meaning less dynamic range in the mids/HF compression driver.

But do you mean in the mid bass, or either way over 300-600 Hz?
I’m considering a JBL midbass, which has an Fmh of 588 Hz. I would have thought that (within it’s mass rolloff range) to extend the midbass’s range upwards eg to 500 Hz, rather than extend the mid/HF compression driver’s range downwards, shouldn’t compromise dynamic range.

While dual drivers and a Unity concept have other advantages, it’s also simpler. This being my first speaker, that’ pretty desirable – what do you think?


Though I now have doubts (raised by others off forum) & ideas about integrating the ribbon with the horns. The right dispersion pattern would help a lot.
If you look at the Aurum G1’s plots http://www.aurumcantus.com/
Vertically, it rolls off after 10 degrees above about 10-12 kHz; while horizontally at 60 degrees it’s 3-5 db down, it’s pretty flat. Could a mid/high horn be built - without other compromises, that’s say 40 V x 120 H?

Then draw up the room layout and figuring the angles, to give a room based ideal, and strike a compromise between flares for optimum tweeter integration, and flares for optimum dispersion for the room?

Thanks

 

[GM]

>Interesting ideas GM!
To clarify, when you said:
“or settle for less dynamic range if a <500-600 Hz XO point is used”, at first I read this as meaning less dynamic range in the mids/HF compression driver.

>But do you mean in the mid bass, or either way over 300-600 Hz?
====
Below this point dynamic range begins to suffer due to a 1" compression driver's limited displacement, so depending on how loud you're likely to ever 'crank' them will set the minimum XO point/slope. I ripped the VCs off my OEM alum. diaphragms in 511/808s using 500Hz/2nd playing mostly uncompressed pipe organ solos at ~ live levels. From this I learned two things:
(1) they aren't rated for this kind of abuse.
(2) I prefer their 'sound' even more after I patched them together with super glue as a quick fix, and haven't popped them again. As always though, YMMV.
====
>I’m considering a JBL midbass, which has an Fmh of 588 Hz. I would have thought that (within it’s mass rolloff range) to extend the midbass’s range upwards eg to 500 Hz, rather than extend the mid/HF compression driver’s range downwards, shouldn’t compromise dynamic range.
====
Hmm, for a 500Hz XO point, a ~356Hz mass corner is more appropriate for a horn going down to 70-80Hz.
====
>While dual drivers and a Unity concept have other advantages, it’s also simpler. This being my first speaker, that’ pretty desirable – what do you think?
====
Yeah, they are simple in concept, but the 'devil is in the details', so not a beginner's DIY project if realizing the most from it in a reasonably short time frame is a goal.
====
>Though I now have doubts (raised by others off forum) & ideas about integrating the ribbon with the horns. The right dispersion pattern would help a lot.
If you look at the Aurum G1’s plots http://www.aurumcantus.com/
Vertically, it rolls off after 10 degrees above about 10-12 kHz; while horizontally at 60 degrees it’s 3-5 db down, it’s pretty flat. Could a mid/high horn be built - without other compromises, that’s say 40 V x 120 H?
====
Not really, it would be highly 'colored' and require considerable EQ to smooth it out. It wouldn't be cheap either since you would have to use a large number of drivers to create a line source capable of being loaded to 70-80Hz. This makes for a large throat, ergo a very large horn. There may be a way around this, but nothing comes to mind at the moment.

Really, you'd be better off either getting more ribbons to stack up to square off its response to more like 60 deg. vertical and make a foam waveguide to limit the horizontal response to whatever it needs to match your mid/HF horn (usually between 60-90 deg), or if the one ribbon has enough power handling for the app, just use the 'WG'. This is what I did when I used the late lamented Pioneer ribbons since there's no typical HIFI app that needs this much horizontal coverage IMO. Even Altec's 105 deg horns are too wide, so I would stuff foam into the two outer rows of cells to roll off their HF BW.

GM

 

[rick57]

GM

> for a 500Hz XO point, a ~356Hz mass corner is more appropriate for a horn going down to 70-80Hz

I thought:
- for a 500Hz XO point, >500 mass corner is nneded
- for a horn going down to 70Hz, as long as Fs < 70Hz (it's 50 hz), that's ok. Why does it matter if Fmh is high or low?


The idea of the 40 V x 120 H to integrate with the ribbon, was for the mids (500 Hz up). Though it souns like that idae should be dropped.
I imagine that 90 * 40 is a reasonable option.

> or if the one ribbon has enough power handling for the app, just use the 'WG'
What's the 'WG' - a particular waveguide?

Cheers