Big waveguide, small midrange

[TulseLuper]

Does it makes sense to use a tweeter loaded in a large waveguide (for example 170mm diameter like the Satori TW29BNWG) with a small (~4") midrange? It seems that tweeter waveguides are always either close to the width of the mid/woofer they cross to or smaller. The Perlisten S7t made me think of this - it has a full-width waveguide with small dome midranges (I realize in this unconventional speaker the tweeter plays all the way to the low end of the mid spectrum so it is not an example of what I'm asking about).

I hope this question isn't totally illogical. I'm familiar with the list of benefits that come from tweeter waveguides, but I do not fully grasp how it works so was hoping to gain some understanding here. I'm less concerned with the efficiency/potential diffraction benefits of the waveguide, and more focused on the best way to achieve smooth directivity. I'm planning an active 3-way speaker with a 4" midrange, and had planned to go with the SEAS DXT tweeter, but am curious to know if the larger Satori waveguide would offer any benefits.

 

[Zvu]

If we start with basics, you need a waveguide to be as wide or a wee bit wider than cone of used midrange beneath it. You get narrowing directivity from direct radiator up to the crossover point and as the waveguide takes over, the directivity becomes constant.

What that doesn't say is that, as you are aproaching crossover point (which is substantially higher than what you'd use with direct radiator tweeter - unless you use waveguide with reeeally wide horizontal dispersion) breakup is also getting dangerously near and all the Scheiße that comes with it.

Unless you are using dsp and pretty steep acoustic slopes, i would say that using smaller diameter midrange with larger diameter waveguide is less of a compromise than it would seem to an untrained eye.

Since you will be using dsp, depending on the midrange you've chosen, you will be fine with 4" midrange and Seas DXT.

 

[TulseLuper]

Thanks for the note.

What that doesn't say is that, as you are aproaching crossover point (which is substantially higher than what you'd use with direct radiator tweeter - unless you use waveguide with reeeally wide horizontal dispersion) breakup is also getting dangerously near and all the Scheiße that comes with it.

Are you saying with the bigger waveguide I'd want to use a higher crossover point? That's the opposite of what I expected. I do intend to stay a long way away from the midrange breakup, and expect to try steep (48dB/octave) slopes.

I expect the DXT to work well, just wondering if there's any benefit to the Satori Be w/ waveguide. Ignoring cost/value at the moment.

 

[Jim Matthews]

I was under the impression that waveguides are used primarily to control directivity with the added benefit of lowering the required crossover point (typically for HF drivers) an octave or so.

I prefer a lower order crossover with Waveguide "loaded" tweeters.

I'm after a smooth transition, where the woofer still contributes higher frequency energy, so the handoff isn't obvious.

Given the size of the woofer described by the OP, I'm not sure it's altogether necessary - and it will dominate what could otherwise be compact.

 

[TulseLuper]

I prefer a lower order crossover with Waveguide "loaded" tweeters.

I'll try several slopes/points - benefits of going with a DSP crossover.

Given the size of the woofer described by the OP, I'm not sure it's altogether necessary - and it will dominate what could otherwise be compact.

Yes, proportions would look better with a smaller tweeter. But I'd get over it if the big waveguide came with big benefits.

 

[Jim Matthews]

Yes, proportions would look better with a smaller tweeter. But I'd get over it if the big waveguide came with big benefits.

I suppose it would be more noticeable with a larger woofer, with a broader range. I like the waveguide design in that it helps "bridge the gap" between a genuine midbass driver and tweeter - eliminating a midrange driver.

Most 4" woofers are designed for midrange, not mid bass.

 

[Zvu]

Thanks for the note.
Are you saying with the bigger waveguide I'd want to use a higher crossover point?

No. Quite the opposite. I'm saying that waveguide forces you to use larger direct radiator midwoofer and cross it higher to be able to exploit full benefits of waveguide loaded loudspeaker and to make directivity transition from midwoofer to the tweeter mild. You could never cross 8" to a 1" tweeter at 1700Hz and achieve good directivity without a waveguide.

If we look from one side, use of a waveguide in that loudspeaker enabled the tweeter to be crossed over low.
If we look from a woofer side, waveguide enabled us to use that woofer mighty high.

DiscoM by Heissmann does exactly that because it uses waveguide loaded tweeter.

https://heissmann-acoustics.de/disco-m/

1700Hz is quite high for an 8" woofer and very close to breakup region for majority of 8" woofers. In this design the fantastic recipe for cone made of glassfiber saves the day.

If you would try to make 8" and 1" tweeter work good without a waveguide, crossover point would be much lower. Probably close to 1kHz or so. That is possible with only few tweeters (that are actually 1.2"). One of them being Peerless DA32TX.

 

[TulseLuper]

No. Quite the opposite. I'm saying that waveguide forces you to use larger direct radiator midwoofer and cross it higher to be able to exploit full benefits of waveguide loaded loudspeaker and to make directivity transition from midwoofer to the tweeter mild. You could never cross 8" to a 1" tweeter at 1700Hz and achieve good directivity without a waveguide.

Thanks for clarifying - this is how I thought it worked.

The tweeter in my design would be crossing to a 4" midrange, so avoiding that driver's breakup would not be hard - I hope to cross more than an octave below it. If crossing a 4" midrange to a 1" tweeter around 2kHz-2.5kHz, is a waveguide necessary for good directivity? Or does the waveguide only become necessary if crossing to a midrange above the point at which the midrange starts to beam? I know this is basic, just trying to confirm that I have it right.

 

[Jim Matthews]

If crossing a 4" midrange to a 1" tweeter around 2kHz-2.5kHz, is a waveguide necessary for good directivity? Or does the waveguide only become necessary if crossing to a midrange above the point at which the midrange starts to beam? I know this is basic, just trying to confirm that I have it right.

That's a good question. There are many perfectly acceptable two-way designs that use this sort of arrangement, mainly with 165 mm diameter woofers. For most, if the front baffle is small and the corners are rounded, the polar response is pretty even.

I beg to differ with Zvu, the waveguide influences only the behavior of the driver it contains. The woofer below will *not* play above its breakup point without distortion regardless of the crossover, tweeter or waveguide above.


Here's the variant I made:
http://troelsgravesen.dk/QUATTRO.htm

More articles from the designer:
https://techtalk.parts-express.com/...speak-waveguide-articles-from-troels-gravesen

 

[Duke]

Does it makes sense to use a tweeter loaded in a large waveguide (for example 170mm diameter like the Satori TW29BNWG) with a small (~4") midrange?

Imo what makes sense would be to choose a waveguided tweeter + midrange cone pairing whose radiation patterns match in the crossover region. I suggest eyeballing the off-axis curves to find potentially good combinations.

 

[Zvu]

Is crossing a 4" midrange to a 1" tweeter around 2kHz-2.5kHz, is a waveguide necessary for good directivity? Or does the waveguide only become necessary if crossing to a midrange above the point at which the midrange starts to beam? I know this is basic, just trying to confirm that I have it right.

It isn't necessary with some clever design of the front baffle. Here's my threeway with 3" midrange and 0.75" tweeter crossed at 3.2kHz. 0 and 90 degree curves provided, i have other curves in between if you want to see them but i warn you, they are borring.



What Duke wrote is essentially the first step in how it's done if you don't have your own measurements of the drivers and assessing the driver compatibility so you wouldn't have to buy them. That's the first stage of what i do when i want to start designing. You should be looking at off axis response, provided that on axis is flat in the intended passband. More precisely, at which frequency the midrange gets -6dB in regard to on axis response. Then you look at tweeter off axis response at that same angle for the exact same thing. This will get you in the ballpark to see if it is doable or not.

Crossover has immense influence, of course. There are tricks. If you use a little bit larger midrange and it starts to beam sooner, adjusting the crossover slopes will give you some leeway. Woofer with steeper acoustic slope and tweeter with shallower slope can lead to better overall directivity in this case.

 

[bigjacko]

The 3" driver's dispersion is pretty close to the tweeter, at 90 degree extreme the response is still close to tweeter. Can you show more angles in between? Also can you provide some explanation on clever design of baffle and why steep woofer slope with shallow tweeter slope makes directivity better? If two slope are not the same how do they merge together?

 

[Plcamp]

There was a bunch of work done (I found it at diyaudio) to match SB tweeters with elliptic waveguides to match dispersion and permit lowered cross frequencies…one such picture below. I printed one for an sb26 tweeter to match an 8” mid but have not tested it yet. There are files for 3d print and for routing.
.

 

[TulseLuper]

I suggest eyeballing the off-axis curves to find potentially good combinations.

Will do, thanks for chiming in @Duke

at which frequency the midrange gets -6dB in regard to on axis response. Then you look at tweeter off axis response at that same angle for the exact same thing.

Here is ScanSpeak's FR data for the 12MU midrange (there's also AudioXpress data from Vance Dickason available which looks just a little different). It shows -3dB 30-degress off axis @ 4,000Hz. -6dB 60-degress off axis is also near 4kHz. -3dB at 60-degress off axis is ~2,500Hz.

So am I looking for a tweeter which has a similar spread at these same frequencies, or at lease one where the spread is similar at the crossover frequency? Here is the Seas DXT graph - at 4kHz, off-axis something like -2dB @ 30-degrees and -5dB @ 60-degrees. And we're also something like -3dB @ 2,500Hz, 60 degree.

 

[Zvu]

The 3" driver's dispersion is pretty close to the tweeter, at 90 degree extreme the response is still close to tweeter. Can you show more angles in between?

This is a passive crossover loudspeaker. Here are horizontal measurements 0-90 degrees in 10 degrees increments. Grid height is 20dB in 1dB increments (you've read it correctly, not 5dB, not 10dB but 1dB increments). I believe that highest frequency response measurement resolution i saw on this website is provided by @napilopez that shows his frequency response measurements with 25dB grid height, so don't be surprised if things look less linear Measurements are done from 1m distance, quasi anechoic and are valid down to 120Hz. Bellow that groundplane must be done to achieve best precision.

0-10-20


30-40-50


60-70-80


You have 90 in my last post. Here is the crossover:

Also can you provide some explanation on clever design of baffle and why steep woofer slope with shallow tweeter slope makes directivity better? If two slope are not the same how do they merge together?

Well, waveguide isn't used just for directivity matching but for avoiding diffraction too. Baffle design influences tweeter directivity, especially true for direct radiators. Here is this tweeter measured on the square baffle and then on faceted baffle



As for the asymmetric crossover slopes, look at Dynaudio LYD5 Amir measured.

Point is that tweeter has mild slope toward low frequencies and since it is just 25mm in diameter, it has wide directivity and plays significant role in overall response at overlap region.

 

[Zvu]

Will do, thanks for chiming in @Duke



Here is ScanSpeak's FR data for the 12MU midrange (there's also AudioXpress data from Vance Dickason available which looks just a little different). It shows -3dB 30-degress off axis @ 4,000Hz. -6dB 60-degress off axis is also near 4kHz. -3dB at 60-degress off axis is ~2,500Hz.

View attachment 143557

So am I looking for a tweeter which has a similar spread at these same frequencies, or at lease one where the spread is similar at the crossover frequency? Here is the Seas DXT graph - at 4kHz, off-axis something like -2dB @ 30-degrees and -5dB @ 60-degrees. And we're also something like -3dB @ 2,500Hz, 60 degree.

View attachment 143558

To me DXT and 12MU look like a perfect match. It will look a bit different in cabinet but it shouldn't be too affected given the intended crossover frequency. Look also at Morel cat 378 and Wavecor TW030WA11 if you realize you don't like the sound of DXT tweeter.

 

[TulseLuper]

Many thanks. I had my fingers crossed that there would be a good beryllium option, but I hear all good things about the DXT.

 

[Zvu]

If i was in Beryllium mode, i'd use BliEsma T25B with one of Augerpros waveguides.

https://www.somasonus.net/bliesma-t25b

 

[PKAudio]

Such a small midrange like 12MU will have diffraction 2-3kHz (wide dispersion), this will make dispersion matching at Fc more difficult.
From my own experiences with 12MU, DXT, Bliesma, 4inch WG etc.....I would go with 12MU and Bliesma, tilted baffle and symetric LR2 at ~3000. This will yield very nice horizontal off axis responses and most probably no dip in power response at Fc when vertical measurements are included.
DXT measures well, but sonically there are better options.

 

[TulseLuper]

Such a small midrange like 12MU will have diffraction 2-3kHz (wide dispersion), this will make dispersion matching at Fc more difficult.
From my own experiences with 12MU, DXT, Bliesma, 4inch WG etc.....I would go with 12MU and Bliesma, tilted baffle and symetric LR2 at ~3000. This will yield very nice horizontal off axis responses and most probably no dip in power response at Fc when vertical measurements are included.
DXT measures well, but sonically there are better options.

Thanks for the input - to be clear, you recommend the 25mm BliEsma as is, or with a custom waveguide?