Is this absurd? (3 way with huge jump from Mid to Woofer diaphragms)

[CinamonRolls]

The SB Acoustics kit "Gema" uses a 1 inch dome, 4 inch mid, and a 12 inch woofer. How big the woofer can get with that tweeter/mid combo? Not certain, but I think 18 inch woofers start beaming around 700Hz. The low crossover for the Gema's 4 inch mid is 680Hz. Could you go as far as an 18 inch woofer? I typically think of three ways making smaller driver diaphragm increases rather than large leaps so I didn't know if it's bonkers or not. Oh, and no waveguide. Just a simple driver on baffle arrangement.


SB Acoustics DIY Kit, Gema: https://sbacoustics.com/product/gema/

I was thinking something like this:



23in x 48in x 14in (W x H x D) - a little wider than a JBL M2 to account for the larger woofer.

SB Acoustics 1" SB26STAC
SB Acoustics 4″ SB12MNRX2-25-4 (or BC Speakers 4" 4NDF34)
BC Speakers 18" 18PS100 (or 18PZB100) - a lower end 18 inch woofer

Crossover: 680/ 3000 Hz (same as Gema)

Side question, why does the Gema offset the tweeter and midrange rather than keeping it in the center? How do you determine how far to offset the tweeter/mid in a speaker style like this?

 

[ChrisG]

Some notes from me:
- If your beaming-related calculations are correct, an 18" driver is probably fine in terms of off-axis response.
- The difficulty you'll have is finding an 18" driver that's smooth/linear/resonance-free up to the approx. 1kHz area. Your crossover slopes are unlikely to be brick wall (and if that was the plan, I'd discourage it for a few reasons), and just-outside-of-band mash can still be quite audible.
- Unless you're planning on a sealed box and EQing it for extreme LF extension (search: Linkwitz Transform), it's a near-certainty that you'll never get close to the limits of the 18" driver. There'll be a lot of LF headroom, sure, but that ends up being an awfully large cabinet & driver that arguably isn't necessary.
- The horizontal offset of drivers can be useful to reduce the effects of diffraction. You can play with The Edge, a neat little baffle edge diffraction simulator.

If you do want to go ahead with this project, be careful of which 18" you select, and you'll probably be fine. I'd suggest sticking with a 15", though, since they're more often used as midbass drivers in PA applications, and are therefore more likely to have a useful response towards your intended crossover area.

I'd also recommend getting a measurement mic, and finding your way around REW.


Chris

 

[JPA]

The offset tweeter and midrange help smooth out the diffraction effects from the side edges of the front baffle. You would want to have mirror-image baffles if you do that.

 

[cavedriver]

any reason you don't want to lower the crossover point to ~500 Hz? That SB mid should be able to handle it to fairly loud volumes (4.4 mm Xmax) and it would give you a little more freedom in 18" woofer choice and crossover slope. I trust you aren't trying to recycle the exact crossover design from the Gema?

 

[Devnull]

Doubtful that the kits xover will be correct changing woofers. Just a quick glance shows the 18 being more sensitive. You may have to pad down the woofers response

 

[CinamonRolls]

any reason you don't want to lower the crossover point to ~500 Hz? That SB mid should be able to handle it to fairly loud volumes (4.4 mm Xmax) and it would give you a little more freedom in 18" woofer choice and crossover slope. I trust you aren't trying to recycle the exact crossover design from the Gema?

No, I'm fine with 500Hz. I used the Gema crossover frequency as a starting point for figuring out if a design like this has any merit. If I build this I'll use Equalizer APO to create the crossovers then convert to passive crossovers if I can get it to sound good with active.

 

[CinamonRolls]

There'll be a lot of LF headroom, sure, but that ends up being an awfully large cabinet & driver that arguably isn't necessary.

Chris

I don't know how to choose a cabinet size for a project like this. Out of ignorance I set it about the same size as a JBL M2 as a starting point. If you built it as a sealed cabinet what dimensions would you consider as a starting point?

 

[CinamonRolls]

- The difficulty you'll have is finding an 18" driver that's smooth/linear/resonance-free up to the approx. 1kHz area. Your crossover slopes are unlikely to be brick wall (and if that was the plan, I'd discourage it for a few reasons), and just-outside-of-band mash can still be quite audible.

Chris

I forgot to ask how to identify a smooth/resonance free driver to apprx 1kHz. For example, consider the BC Speakers 18PS100. It lists a frequency response of 30 - 1000 Hz but how would I know if it's actually smooth and resonance free up to 1kHz? Is the only way to know that information through physical measurements after it's in the enclosure?

18PS100 8 Ω - LF Drivers - B&C Speakers

www.bcspeakers.com

 

[CinamonRolls]

Just discovered I misunderstood beaming and how to calculate when a given diameter starts beaming. An 18 inch woofer basket with a 15 inch diaphragm should "beam" around 1250Hz (see this link: https://acousticfrontiers.com/blogs...is-response-theory-and-measurement-techniques). It's the point where lobes start to form due to phase cancellations inside the cone.

I mistakenly thought beaming was when a diaphragm started to rapidly switch from omnidirectional to narrowing. And now I'm confused about where you can safely cross a woofer after it starts to narrow from fully omnidirectional.

 

[DWPress]

Speaker driver data sheets give you the TS parameters that you will give you optimal cabinet size. Parts Express usually gives you these suggestions for sealed and ported boxes in their specs. There are many online calculators for figuring this stuff out too and I suggest you learn to use them before purchasing anything. I have DIY speakers with 15" woofers in sealed cabinets and 180 liters is the minimum volume needed for the frequency range I need them to perform in AND I'm still ditching everything below 65Hz to subwoofers.

 

[CinamonRolls]

Part of the reason for the larger than necessary size with regard to low frequency woofer efficiency (if that's the correct term in this context) is directivity control. However, I'm trying to learn so I don't have a hard and fast opinion on the best approach to balancing box volume for low frequency output and enclosure impacts on directivity. And I'm a long way from understanding the implications of floor bounce (e.g. should I move the 18 inch woofer lower or keep it in the middle of the enclosure because it creates better vertical directivity that way.)

 

[ThatSoundsGood]

Just discovered I misunderstood beaming and how to calculate when a given diameter starts beaming. An 18 inch woofer basket with a 15 inch diaphragm should "beam" around 1250Hz (see this link: https://acousticfrontiers.com/blogs...is-response-theory-and-measurement-techniques). It's the point where lobes start to form due to phase cancellations inside the cone.

I mistakenly thought beaming was when a diaphragm started to rapidly switch from omnidirectional to narrowing. And now I'm confused about where you can safely cross a woofer after it starts to narrow from fully omnidirectional.

That B&C woofer says that it's useable up to 1KHz because the cone breakup is above it. So, it's "useable" up to 1K, but it's going to beam lower than that. Also, that woofer wants to be in a ported cabinet of 9 cubic feet. That's massive. That's a concert subwoofer.
Hopefully I can help you clear your confusion. Beaming is when the horizontal (and vertical) dispersion narrow off-axis. So its the frequency where the off axis response begins to fall off. The larger the cone, the lower the frequency. AN 18" cone is generally in the 700Hz region. You want uniformity between drivers in your off-axis frequency response in order for your speaker to sound good in a room. For example, you don't want the frequency response to get narrow around 700-900Hz then widen from 900-2500Hz then get narrow again above that. This creates a situation where the room reflections are imbalanced and you will hear that at the listening position and it will sound poor. So, you want to select a combination of drivers and crossover frequencies with a certain uniformity in the on and off-axis response. Some speakers get narrower than others off-axis, but as long as they do it in a consistent fashion then it will still sound good. This means that you want to cross low and mid woofers below (or near) the frequency where they begin to beam.
Another point here is about when a speaker becomes omni-directional. This is often referred to as the baffle step. This happens when it's mounted into an enclosure and the width of the baffle causes frequencies to wrap around the speaker. You have a 6db in sensitivity loss around this region. It's a gradual transition.

I would say that using an 18" woofer for this would be quite ambitious and is gonna cause a lot of problems for you. Most 18" woofers are subwoofers and need massive ported enclosures. They are meant for enormous amounts of power and excel at producing concert levels in large spaces. But of course, have at it if that's what you want to do.

 

[CinamonRolls]

Thanks for that information. I discovered you can use the Diffraction tool in VituixCAD (with the export option) to see where a given diameter will narrow and eventually beam. This is what it looks like for an 18 inch basket frame with a 15 inch diameter woofer.

In VituixCAD when you select the Directivity checkbox (and Vertical if you want to include it) then select the Export button it will export FRD files you can then import back into Drivers/Frequency response just like you import REW measurements.

 

[CinamonRolls]

It looks like you'd get the best sound out of them by crossing under 450Hz and especially under 250Hz. Though I don't mean crossing above that would be bad. Just that 450Hz and 250Hz are guaranteed to work well.

With regard to enclosure size for the 18 inch BC 18PS100 you could build a cabinet around 60 inches tall, 23 inches wide, and 15 inches deep to accommodate a bass reflex alignment with an F3 of 29Hz. Yeah, it's big. But you don't need a speaker stand and you can perch gargoyles on the top to ward off evil spirits.

 

[tmuikku]

Hi, you have lots of good questions, very fundamental ones, and it takes helluva long post to explain so best advice is to take your time and study the concepts yourself. Few tips:
- only good reason to put woofer to same box with mid and tweet is ease of built and cost savings, which is mostly a commercial reason not reason from audio quality.
- big woofer could be in small box as well, if you have big power amp and suitable processor. Big box is needed if you beed all the possible output, and manage with passive filters that cannot boost lows (EQ).
- check what floor bounce actually is. There are 6 boundaries in a typical room, not just the floor, all of which make a first specular reflection, all of which are equivalent of "floor bounce". Specular reflection from floor typically has shortest path length difference to ditect sound of all of the 6, and is thus a problem in mid frequency, while ceiling and wall bounces are lower in frequency and thus possibly worse "bass issue". Tupically few of these are close enough each other and make huge dip in low mids, and floor bounce might be part of it.
- size of physical things like drivers and boxes directly relate to sound wavelength as you can see by playing with diffraction tool as you already have. Directivity and efficiency is higher to lower frequencies with bigger drivers.
-SPL from a driver at any frequency is related to volume displacement, and since the driver Sd stays ~same excursion must go up as frequency goes down. To aiintain flat frequency response excursion must quadruple for every octave down in frequency, thats why you need very big woofers to get any serious output on lows. 20kHz is 1.7cm long, fits on a thumbnail, while 20Hz is 17meters long and bigger than average detached house! All drivers, boxes, Floor bounces and most problems are mid frequency sized. So, you'd want separation of concerns to optimize for these very different wavelengths, hence multiway speaker NOT in one box would make sense because lows and highs can have different optimal locations in room for example.

To make any sense of these stuff you must have a context to reflect against, because you get all kinds of advice which is mostly contradicting and confusing because it's for different contexts. For one small speaker is good, for another a bigger one, and so on, so all the advoce might be good it's just for different contexts. A lot of stuff is just from business and marketing reasons, so you need to have some sense for these as well to be able to determine what is relevant to you and why, a lot of it isn't for DIY person. Commercial or any other reasons could be relevant to you, or not, you just have to figure out in order not to be confused about it.

Very solid foundation for speakers, context, is just to figure out what bandwidth at what SPL you need. Calc say 15db of headroom for transients, conpensate for listening distance to get figure for 1m system sensitivity. Now you have some concept for size in general. Then, what kind of directivity is good? depends on listening distance and acoustics how much you listen room sound and direct sound, how much you move or is it just for one listening spot. How about room modes? and so on. And you'll start to get an idea what kind of structure might deliver.

Most important thing is to have fun, so have fun with you project!

 

[CinamonRolls]

Not Sure if I did this correctly with VituixCAD but this is what I get when I use the Diffraction Tool and add the three drivers listed above on a 23 x 60 x 15in (W x H x D) enclosure. Not sure why it widens directivity at 1000Hz because I haven't moved the drivers off center to see what happens.

@tmuikku I have DSP and can boost lows with amplifiers. I'm trying to see what a large enclosure does and if it's worth building. In this case it's a really large box when using an 18 inch woofer. So adding the tiny 4 inch driver and 1 inch tweeter is nothing compared to the enclosure size. It's basically a subwoofer enclosure with mids and highs attached just because the box is so huge. Although, maybe that's how they used to think of it with the old speakers the SB Acoustic Gema is based on.




Normalized Horizontal/Vertical


Raw Horizontal/Vertical



Is directivity like this desirable or is it too wide? Is the extra width around 1,000Hz an advantage or a disadvantage? I could place the mid/tweeter into a waveguide to make everything smooth but does that make the speaker "dry" and boring?

 

[CinamonRolls]

Also, it almost looks like it's behaving like a coaxial. The crossover null at 3,000Hz is non-existent in the listing window. It looks like you have a huge vertical listening window. Which makes me think I messed up the VituixCAD model. I didn't separate the drivers because I wasn't sure if you're supposed to do that when you import your FRDs from the diffraction tool. Pretty sure it's wrong so I'll try to correct it.

 

[tmuikku]

Hi, do diffraction sim per driver, put mic on axis to that driver. Then use x,y coordinates in main program to position the driver (measurement data) in relation to intended listening axis. Just like you would do if you were measuring a real box with mic, as per Vituixcad measurement manual. You can do ~same thing in many ways, but it helps to reduce humanly errors if you stick to the same procedure with sims as with real data and measurements. Think the diffraction tool is just virtual spinorama.

 

[CinamonRolls]

When I adjust the drivers to the same spacing as the Diffraction tool it looks like this. So I'll need to keep working on it.

 

[AnalogSteph]

Doubtful that the kits xover will be correct changing woofers. Just a quick glance shows the 18 being more sensitive. You may have to pad down the woofers response

If you have to pad down the woofer, something has gone very wrong. Doing this is undesirable for multiple reasons, from plain resistor power handling required to poor damping factor (keep in mind that the fundamental resonance will be well within the desired passband, unlike the other drivers). In general, the woofer should (and generally will) be the least sensitive part. Now, when dealing with odd combinations like 18" woofer + 4" midrange, this is not at all guaranteed. The midrange may need to be a high efficiency type and placed in a waveguide to boot. You can give the midrange and tweeter a 3 dB advantage by going 4 ohm with an 8 ohm woofer, but that may already have been the case to begin with.

These are the kinds of things you have to deal with in a passive speaker design. When going active, you can combine drivers to your heart's content.