• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

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

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.

View attachment 424138View attachment 424137
Make sure your tweeter has only tweeter measurement, iow no other drivers on the diffraction tool when you export, and that the microphone in diffraction tool is on tweeter axis, not middle of the baffle where it defaults. Do same for mid, export only mid driver in place, microphone at mid driver 0-axis. Same for woofer. tip: If your drivers are offset to side in diffraction tool you must also use the x-coordinate.

You could also put the mic in diffraction tool to intended listening axis for all drivers, and then leave coordinates in main program to 0,0,0, because measurement data now includes the relative positioning. This however, is less flexible solution, but perhaps bit faster to whip up as simulation. If you measure on axis of each driver, you can use coordinates in main window to fine tune the response bit better, because the data is more accurate to do that. When coordinates are adjusted to deviate, the plan initial of course changed, and you need to make sure it's still possible to build it, perhaps generate new data to verify it's as intended. It's just design freedom to be able to juggle all things, until satisfied.

Drivers in the main program are really just "sets of measured data", not drivers, and it's your responsibility to understand when the main program and data is setup to reflect "reality". Anyway, I hope this helps you to make sense how to do it.
 
Last edited:
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?
Too wide might be a situation, where your listening distance has to be far away compared to room acoustics and you want to perceive sound that is not dominated by early reflections. :D Weirdly written sentence, what I mean is that any width of directivity goes, you just have to position the system suitably. You can compensate for effects of wide directivity conpared to barrow directivity by reducing listening triangle size, which delays and reduces level of early reflections in comparison to direct sound. Basically it's possible to get ~similar perception of sound in a given room with various systems. In this sense, directivity is related to room acoustics and at which distance you want to perceive which sound.

If your positioning is nailed due to practical reasons like "speakers have to be at one wall where the TV is, then there must be carpet and then a sofa other side of the room" -situation, then you'd might want to reduce coverage in order to reduce effect of early reflections to perceived sound, if you want that kinda sound. However, most people seem to prefer the early reflections dominated sound for some reason, which might be the case with you as well so there is no preferred coverage unless you understand what you want to perceive, how is your room acoustics, and what the playback system directivity needs to be in order to provide the perception.

This stuff is not mentioned almost ever, but is cruzial to make sense of any of this. Until then anything goes as long as the directivity is smooth, iow no weird peaks and dips in the power response. There might be some special rules of thumbs here, that "narrow coverage on some kHz helps with something", but this I cannot advice to you as I'm proponent to just smooth directivity as it seems to work very nicely. You might have to study this stuff yourself, which means building several speakers and AB comparing, this is easily multiple year process. If in doubt, try to get smooth directivity, smooth enough response toward listening window, but don't sweat over it. I bet most people don't notice a difference unless pointed or directly AB comparing. And, in the end big part of it is positioning the listening triangle. Max freedom to position the systen is enabled when the speakers are problem free enough so that sound is nice to most important directions. For example bad edge diffraction related interference makes response change per angle, so basically prevents you from proper toe-in, if it matters or not depends on your ability to notice several db undulation in frequency response per angle.

Adjectives like dry and boring might be due to many reasons, which you need to figure out yourself. Personally, I much prefer very involving sound that takes my attention and goes deep within. This is not happening with strong early reflections so in normal domestic acoustics this means either short listening distance or high wide bandwidth DI and needs some care with positioning. I suppose someone might consider this dry and boring because it's not as spacious in standard living room as the early reflection dominated sound most seem to adjust for but it's about room acoustics more than the speakers, you'd need to arragne envelopment somehow to make it spacious. But, the early reflection sound cannot take any involuntary attention from brain so clarity just isn't there, and does not have envelopment but just the hazy spacious sound, so it's good for background music for example, a different thing basically not due to the playback system but howbyour own auditory system processes the information entering ear canal and what perception it provides to you. Early reflection dominated sound is boring to me in comparison what the sound can be when also auditory system is fully involved. Could be boring and not dry, or not boring and dry, depends on how you have positioned the system and how is the room acoustics.

So, you must figure this stuff out yourself what you prefer, and what these adjectives mean, and how your system needs to relate with room acoustics in order to have it. That said, I prefer both sounds and utilize them as I wish. I take a chair for close listen when ever I feel like being sucked into the recording, but most of the time I'm listening the far sound, so I've tried to optimize my system for both which means very smooth directivity and no on-axis issues to enable freedom / proper toe-in for example.

As disclaimer I'm a hobbyist and have not listened on all speakers in the world in all rooms, and the above is just my experience so far which I could change as my experience widens. My main point is to raise awareness on context and how important it is to be able to determine what works and what doesn't, and what matters and what doesn't, what's in between.

It's hard to write at suitable level to make it understandable and relevant, so please do as you feel relevant and perhaps read my posts later on :D have fun!
 
Last edited:
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.

View attachment 424122
View attachment 424125View attachment 424124View attachment 424123

Normalized Horizontal/Vertical
View attachment 424128View attachment 424129

Raw Horizontal/Vertical
View attachment 424130View attachment 424131


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?
That directivity is not desirable. I'm not sure why it's widening at 1Khz since it seems to be in the middle of the mid-range. What I can tell you is that this model doesn't look very good.
I encourage you to shoot for:
1) flat frequency response on-axis
2) Horizontal directivity that never widens after it narrows as you move up in frequency response..........you want it to gradually narrow or stay the same
3) Vertical directivity that does the same but it's understood that there will be variations here when you have vertically stacked drivers.

I encourage you to keep playing with the model. Keep in mind that the listening position is important on these models. Generally speaking, it's 1 meter on-axis with the tweeter. Play with different crossover settings. Maybe try a 15" woofer since that's certainly going to yield better results and is unlikely to sacrifice much or anything in the lower frequency range for a hifi box in most rooms.
Another note about some large drivers: Most of them are not meant to be put in sealed enclosures. You can do that and use DSP to compensate for low frequencies, but you have to be mindful of maximum SPL at these lower frequencies and you will usually be severely limited there when the driver is intended for a ported enclosure, which most B&C drivers are. You can put these speakers in any box that you want, but DSP is really just a band-aid in these situations and there really is no substitute for a properly designed enclosure based on Thiele-Small Parameters.
 
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?


View attachment 424079View attachment 424078

Take a look at the frequency response, and then the impedance curve.

Resonances are indicated by frequency response anomalies, and usually a peak in the impedance curve. The bigger the peak in the impedance curve, the stronger the resonance.

There's a peak/dip at 300-350Hz which shows up on the impedance curve
There's also the brutal spike in the 1-2kHz range, which also shows up quite well in the impedance curve.

There are other smaller wiggles between 400Hz-1kHz as well.

IMO, the 18PS100 should be used up to 200Hz as a maximum, and that'd be with a reasonably steep (at least 18dB/octave) lowpass. The Faital 15FH520 will get you another octave before there are any resonances, and even those are quite well-damped: the frequency response curve is much smoother, and the impedance curve only shows minor wiggles. https://faitalpro.com/en/products/LF_Loudspeakers/product_details/index.php?id=151060150


In general, PA drivers use paper-based cones. That's not the only option. Dayton Audio make subwoofers with hard aluminium cones. That means the cone will stay as a rigid piston until you get to its resonant frequency, and then all Hell breaks loose: https://www.daytonaudio.com/images/...ton-audio-rss390ho-4-specifications-46177.pdf

For my own HiFi speakers, I took the hard aluminium cone approach with 2x Seas H1252-08 per side. They're running up to around 950Hz, crossing to an 18Sound NSD1095N. The crossover is complex, since I only have approx 2x octaves before some truly hideous resonances begin.
I'd also like to note that some cones are made of softer aluminium, allowing controlled break-up in a similar way to paper cones. The Mark Audio drivers are an example of this approach.

When it comes to putting PA 18" subwoofer drivers in sealed boxes, there's no reason it can't be done. They're typically intended for ported alignments, so will often return either a very low Q-factor in a sealed box, or require a very small sealed box (with a correspondingly high system resonance) resulting in a rolloff that's at a higher frequency than you might want.
Since you'll be using 18" drivers, though, EQing the LF response is an excellent option. Again, reading about the Linkwitz Transform is a good move here.

Remember: the maximum output of a driver in a sealed box is fixed, based on Xmax, cone area and frequency: http://www.baudline.com/erik/bass/xmaxer.html
If you make the sealed box smaller, the driver will have to work harder to compress the air inside, meaning you have to put more power in there to get the cone to move by a given amount. The more power you drop into a driver, the more distortion is produced by the motor.
It's a balancing act: you could use a really big sealed box and have high efficiency and low distortion, or you could bring the box size down to a more sensible size and accept a little more distortion.


Finally, the interaction between a cabinet and the driver can be simulated, allowing excellent prediction of the anechoic LF response. The first thing you'll find is that there is no "right" cabinet volume for any given driver. There's just a bunch of different response curves (some of which will match classical "alignments") to choose from.


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?


View attachment 424079View attachment 424078
Look at the impedance plot at right. See 1.4 khz has a wiggle? That is a pretty big resonance. See how the frequency response at left has a peak at the same frequency? You will need your crossover to fight valiantly to make sure that is inaudible. I. addition to crossing over at around 400 hz, well before the peak starts rising, you would also look to add a notch filter to absorb some energy at the resonant frequency.
 
Corrected the VituixCAD model. I shifted the 18 inch woofer down to 150Hz so the 4 inch couldn't handle going that low. But at least I can see how the large woofer and enclosure work now. Keeping the 4 inch with an 18 looks like it would require going 4-way. Or reduce the 18 inch as has been suggested. Alternatively, resort to a waveguide with compression driver and use a larger midrange.

18_BigBoy XO-schema-1.png18_BigBoy Six-pack.png
 
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

Once again, you do realize how big this is? We're talking small refrigerator size speakers. My inside dimensions are 14x35x22" and with the bracing, drivers and reinforced baffles they weigh a ton and take up gobs of space. You'd better have a good size room, listen from 10+ feet on average and not plan on moving anytime soon. Mine are in a workspace studio and I fly them from the ceiling, there's a sweet spot but design is for an enveloping soundscape. Yes, they sound great but the trade offs....

Note their size vs the doors.

IMG_7233.jpg
 
  • Like
Reactions: MAB
Once again, you do realize how big this is? We're talking small refrigerator size speakers. My inside dimensions are 14x35x22" and with the bracing, drivers and reinforced baffles they weigh a ton and take up gobs of space. You'd better have a good size room, listen from 10+ feet on average and not plan on moving anytime soon. Mine are in a workspace studio and I fly them from the ceiling, there's a sweet spot but design is for an enveloping soundscape. Yes, they sound great but the trade offs....

Note their size vs the doors.

View attachment 424238

You say that like it's a bad thing.
 
  • Like
Reactions: JPA
You say that like it's a bad thing.
Not at all! Just making sure you realize what you're committing to! Your proposed box dimensions are nearly 2x the size.

The first version of these cabinets were built over 30 years ago and have had several refinements over the years with drivers, the front baffle and DRC for their unique placement. I've only had to move once and it was a 4 mile jaunt. No regrets but if I were to do it again I'd do multiple smaller woofers in a more manageable cabinet for versatility.

Edit: My initial idea was full frequency range in a pair of speakers but like I said, I also use 2 subs and that's mostly to fill in the room modes from the placement of the main speakers. They easily dip down to 27Hz at F3 but the subs do a better job and with less distortion taking care of things up to 65Hz.
 
Would narrowing the low-mid verticals like this help with SBIRs or is that something we shouldn't worry about? It's not a constantly rising DI but it rises at and below 500Hz. Horizontal remains consistent. (I converted the 4 inch woofer into a coaxial driver, that's why the 3000Hz crossover nulls are missing)

18_BigBoy XO-schema-1.png18_BigBoy Six-pack.png18_BigBoy CTA-2034 Early reflections.png
 
Vertical first specular reflections are at such low angle, only about an array is effective to reduce sound toward those, but it all depends on your room, listening distance and height of things. With generic 2.5m room height, 3m listening distance and listening axis about 1m from floor the specular reflections are about at 30deg to floor and roughly 45deg to ceiling. If sound toward these is attenuated for example 6db, it could help some. -20db would be quite substantial, but would mean an array.

The thing is, specular reflections are almost copy of direct sound and interfere with direct sound arriving few ms later, which means both destructive and constructive interference. Narrow bandwidth dip toward specular reflection does almost nothing to the resulting interference, it reduces one or two peak and dip, like one octave out of many, all the rest is still there. Visually there is almost no difference, audibly I can't say, haven't done AB test, perhaps there is some. Perhaps it helps if on some critical bandwidth where ear is sensitive to such things. Audiblity of early reflections depends on direction for example, level and delay.

If two of six early reflections have narrow bandwidth difference, how much is the end result changed? You can emulate this in vituixcad, add 1 sound source (driver without any data loaded) to emulate direct sound, then six other sources with bit of a delay to them to emulate first specular reflections. Now play with PEQ to introduce frequency response dip for two of them. Add any other EQ and attenuation blocks you think are necessary to more closely emulate what you have there, directivity, acoustic treatment and others. But, this is just academic stuff, in reality there is a lot of sound from all directions soon after, auditory system molds it all into a perception.
 
Back
Top Bottom