• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Small Subwoofer Drivers in Tower Speaker

Arnandsway

Addicted to Fun and Learning
Joined
Sep 13, 2018
Messages
548
Likes
510
I'd suggest, if you can stretch it, to go active and use the Hypex boards. You'll have more possibilities. If you don't have the most trustful measurements, the passive crossover might be incorrect. And that will cost you in the long run.
Besides, you then have the possibility to use the Hypex boards for a later project
 
OP
D

D!sco

Active Member
Joined
Jun 16, 2021
Messages
218
Likes
149
From the griup delay graph the mid range and tweeter have 180 degree phase difference, is this your intended design? The power and DI graph looks alright though the phase difference is 180 degree.
I'm still learning some of these aspects of speaker design. What can I do to change this in a passive design? The tweeter appears to always have 90 degree change from 20-1khz. I can only really seem to make this trend more quickly from 180 to 0. I'm gonna guess that's worse.

I have some real work again, finally, for the first time since the pandemic started that wasn't an ohshit emergency. I haven't been thinking about this project as much lately.
 

bigjacko

Addicted to Fun and Learning
Joined
Sep 18, 2019
Messages
673
Likes
323
I'm still learning some of these aspects of speaker design. What can I do to change this in a passive design? The tweeter appears to always have 90 degree change from 20-1khz. I can only really seem to make this trend more quickly from 180 to 0. I'm gonna guess that's worse.

I have some real work again, finally, for the first time since the pandemic started that wasn't an ohshit emergency. I haven't been thinking about this project as much lately.
If there are 180 degree phase difference, you can invert one driver. 90 degree difference you can't solve it like this. I am not good at crossover too, there are many people can tweak the phase response to match two drivers perfectly in phase. I really hope there is a way to calculate the crossover component value given the frequency and phase response, but I don't know if there is something like this. There is one thing that is close, which is Xmachina crossover, it uses AI to do crossover for you, you just need to provide raw driver response and passive component values. Another way is to use active crossover like FA123, you will not need to learn passive crossover and have even better frequency and time and phase match.
 

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
235
Are you loading trace data directly into vcad or running the traces through the baffle sim and exporting each driver's off-axis and loading that into vcad? I would use a 60-degree off-axis trace and use that to generate 90 and 120 degrees in the baffle sim. For the vert, just use the on-axis measurements and let the baffle sim generate the rest. Be sure to know the SD of the drivers.
 
OP
D

D!sco

Active Member
Joined
Jun 16, 2021
Messages
218
Likes
149
Just did a quick update to my box model. It is now 1250x500x200mm (7"Wx50"Hx20"D). Approximately 82L (2.89ft^3) with the same pair of 3"x12" ports, now pointed forwards. This gives it an estimated F3 of 26Hz and an F10 of 22Hz. I have increased the bracing due to moving the ports lower, and reduced the size of the midrange box from 9.5L to 7L. I'd like to change the shape (and size) but need to consider ventilation from the top and bottom chamber. If the midrange chamber starts narrowing halfway down it's length in a simple 45 degree shape, I may have enough ventilation to keep the top sub happy, and maybe squeeze a bit more space and rigidity from the box. 3D printing a sealed cone still isn't out of the question.
I'd like to do a better job understanding how to use the enclosure calculator in VCAD and baffle diffraction, now that I'm settling on a cabinet size I feel comfortable with. If anyone has any tutorials they'd recommend I'd love to watch or read them.

I'm not sure the depth of the four-sub tower shows in 3D. 26 vs 20" depth is a huge difference in my smaller living space.
Screen Shot 2021-09-19 at 17.22.19.png
 
OP
D

D!sco

Active Member
Joined
Jun 16, 2021
Messages
218
Likes
149
I ran some box simulations on the SEAS woofers, and they’re extremely underwhelming. Big ported boxes only go down to 45hz. The Sica + Tang Band may be interesting, or Sica + Sica/Scan-Speak.

Really, the biggest competitor for my money here is the Philharmonic BMR Tower. $2k for a proven design I can take active later. Since my goals are point source sound with deep extension, the new tower already fits that bill, albeit with a much wider front baffle.
 
OP
D

D!sco

Active Member
Joined
Jun 16, 2021
Messages
218
Likes
149
I think this weekend I’ll try iterating my Tang Band SCS (Sub-Coax-Sub) design with passive radiators. Currently I see the 6.5” Dayton PR but without any experience I feel like the 8mm xmax may be limiting. There’s plenty of space for 4-6 PRs on the front/rear baffle. If possible I’d want the sides to remain clean for now.
 

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
235
"but without any experience I feel like the 8mm xmax may be limiting."

VCAD can answer that-sort of. If the question is, what will my baffle/spl trace sim's EQ do to the T/S param sim?

If you have a model in the enclosure tool and vcad you can link them and see what your eq is doing to the xmax and what SPL will exceed it. In the enclosure tool, on the "Align" tab, check off "Crossover of driver". Pick the driver from your VCAD sim that has filters. Use the voltage in the XO sim to crank things up.

I attached a "no eq" vs "eq" graph of the spls from the enclosure tool based on a small mid. With the filters applied and the voltage set high, I can see where my mid runs out of Xmax. The SPl dotted line is the T/S param sim w/o the eq. I only wanted it to hit 100-104db so this is fine. Notice the excursion is pretty useless w/o the eq attached.


Eton SYMPHONY II 3-312-C8-25 HEX Six-pack.png
 
Last edited:
OP
D

D!sco

Active Member
Joined
Jun 16, 2021
Messages
218
Likes
149
I like this idea but I’ve never been able to get both the crossover and multiple drivers working at the same time. Mostly I have to adjust ohms and guess based on existing data
 

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
235
On some level, messing with the baffle sim and 3rd party data, is close to the same process as measuring a real speaker. It is all about generating the vertical and horizontal data for each driver. The difference is you use the mic and angles in the baffle sim to generate the measurements. Everything happens in the baffle sim- no merging of near and far-field. You can merge an export of your enclosure sim's SPLs to your baffle sim to improve the 500-1000hz range.

Remember this is all a fuzzy view. Not the same as measuring a real speaker.

Be sure to check out the "Baffle diffraction simulation" section:


Since you have a 3d model of your sim, find out the Y position of each driver. Use that to make a baffle sim file for each driver's location and put the mic over the driver (there are boxes to type in the X/Y for the mic). Each driver should have the correct SD. I imagine you could just model the inside ones because you really just want to see how high freqs. are going to get messed up. Once you have all the baffle sim files done, you can use those to load the trace data. I think this is a part I might need to document.

In the diffraction tool: load the on-axis trace into the baffle sim with the folder icon by "half-space response". Check "full space". Messing with the "Angle ver/hor" settings will let you change the angle. Use the export to save the file for vcad's XO sim and use the naming convention (see image). I like to use the "feed speaker" check box only for the vertical. When you have that check and hit save it will save all of the needed angles at once. Now load that data into vcad's XO sim. Be sure to set the Y of the drivers in vcad's XO sim. If the tweeter is the center, all the distances should be from there.

To break down the data:
Use the on-axis measurement to create the vertical files and on-axis files
Use each polar measurement trace to generate the same polar baffle measurement. If the trace is for 15 degrees, set the "Angle hor" to 15 degrees in the baffle sim.
Use the 60-degree or furthest polar to generate the 60-120 degree files. You can set the baffle sim to 80 degrees and save the file as the 60 degree to play with how things roll off. I made a high-low sims messing with this.

I hope this helps.


vcad_data.png
 
Last edited:
OP
D

D!sco

Active Member
Joined
Jun 16, 2021
Messages
218
Likes
149
It really means everything to have some good people to talk to like this. Thank you.
 

headshake

Active Member
Joined
Apr 20, 2019
Messages
297
Likes
235
Cool! I hate writing so I did not know if I wrote enough to be helpful.

One more pic, the baffle sim setup. My baffle has a 44m edge. The tweeter's SD is set. No reflections. The mic is over the tweeter. A 60-degree tweeter trace is loaded and the horizontal set to 60. This is all that is needed to export each angle for each driver.

To help see what is going on, I graph traced the 0,30,60 degree spl's from the baffle sim. I loaded the file I created with the trace tool and changed the angle in the baffle sim to generate each of the green lines. Exporting is just another step. The export will only create a file for the active SPL sim, not the graph traces. I find it very useful to save traces within a graph. You can compare baffle positions in a second. The feature is all over VCAD with a right-click on a graph.

baffle_sim.png
 
Last edited:

sarumbear

Major Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
4,445
Likes
4,003
Location
Southampton, UK
Hey all, I've been reading more and more about speaker design since I built my first set and I have an idea I'd hope to be able to execute in a few years.
The 4" Tang Band midbass drivers I have on my desk are impressing the hell out of me with their low end. Tang Band clearly knows how to produce a lot of subsonic with their smaller drivers.
I assume you used the word "subsonic" as an exaggeration as one will not expect a driver with fs=60Hz to generate sound below 20Hz.

The title of this thread is "Small Subwoofer Drivers". I expect a decent subwoofer to produce at least 30Hz.

The driver you mention will have an f3 at around 100Hz when mounted in a Butterworth aligned closed box. It means you need to use an equaliser that boosts 30Hz for 40dB! For a reflex enclosure it is even worse as the slope is 24dB/octave and the boost is stupendous. Not to mention in order to generate 30Hz at, say 96dBSPL you need to use nine of those drivers.

Thiele has written his thesis in 1972. There had been speaker software on the market since 80s. Internet is full of simple calculators for the job. It is not hard to try to use a bit of science...
 

sarumbear

Major Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
4,445
Likes
4,003
Location
Southampton, UK
I also would have assumed you could read given your industrious writing capabilities but clearly I hadn’t considered comprehension as a factor. Those are mid bass drivers, not subs. I’m not talking about using those. Take your pretentious horseshit somewhere else.
Why don't you behave properly in a public forum without using explicit language. How old are you: 12?

I commented on your post which says "Small Subwoofer Drivers in Tower Speaker". It says:
[...] I have an idea I'd hope to be able to execute in a few years.
The 4" Tang Band midbass drivers I have on my desk are impressing the hell out of me with their low end. Tang Band clearly knows how to produce a lot of subsonic with their smaller drivers. I've also been a fan of the 6.5" driver [...]
The object of your post is the 4" driver. You also considered a 6.5" driver.

It is OK to change one's mind but why refer to explicits when you are told your original idea is not workable? Or that you are told that you are using the word "subsonic" wrong.

You have been reported.
 
OP
D

D!sco

Active Member
Joined
Jun 16, 2021
Messages
218
Likes
149
Hey all, it's been a bit. Thankfully it's because I've been busy, which means I may get to build again one of these days. This has led me to some more practical thinking about the project. To fully let these TB 6.5" subwoofers breathe, they need an enclosure I may not be able to provide for my space. Their power handling and sensitivity may also not be conducive to living room applications. If I pursue the "forward facing subs in-baffle", my approach may come in phases, starting with a bass module. I already have some HiVi 3.1's in my living room on monoprice stands. I did some rough volume calculations, and the ~32 liter, 7.6x400mm (3x16") ported box could actually work really well as a stand for my HiVis...
Screen Shot 2021-10-15 at 13.36.06.png
I could build a sealed coaxial/fullrange and place the three boxes in a vertical sandwich as a first attempt at a tower. It would be 60.5" tall.
And if they're not great, I'll make a car sub out of them.

My real holdup on digging into this design in particular is the caution I have towards the Tang Band coaxial drivers. I still want to build a speaker that plays flat with point source imaging and quality dispersion. The added goal of full range is still something I want, but I think I'll have to get a separate sub until I own a house. In the meantime, I'm back to the drawing board with the SEAS reed coaxial and reed woofer. Not sure if I should just start another thread for that or rename this one to "idiot tries to make the perfect tower speaker with minimal practical experience please come help".

And, of course, I've already worked out the enclosures and some 3D mockups. They're not remotely finalized, but it's the minimum before I can take a project seriously. The coaxial should be in its own 7L sealed enclosure, with the paired woofers in a 28L with two 76x250mm (3x9") ports. Gives it an F3 of 55. I'm also considering raising the sensitivity by including a third woofer, in parallel with the coaxial woofer. Maybe not the best idea.
Screen Shot 2021-10-15 at 13.50.50.png

Anyways, I've really appreciated the help and suggestions along the way. This community is fabulous and I couldn't wish for more.
 

Wolf

Senior Member
Joined
Nov 21, 2020
Messages
336
Likes
372
Location
Indiana
If you do use PR for the 1139, the 6.5 and 8 inch PR from Dayton DS/SD series cannot apply. Use a 10" per woofer. The others either don't have enough Sd or xmax initially, or can't take enough mass when in single or multiples to get the tuning Fb low enough.
 
Top Bottom