Bullet-4W3's: 4x W3-881SI per Tower build

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
My first experience with these little 3" full range speakers was for a friend who wanted a portable kareoke machine with better sound quality than what you buy @ wal-mart. I used a small amp, ran through a netbook, and the little buggers sounded REALLY good for a small sealed enclosure!

So, I bought a handful more for future projects. This winter, I decided to build a pair of the "Nanotowers"...
http://www.diyaudio.com/forums/full-range/200912-nanotower-tang-band-w3-881si-9.html#post3293089

The results were stunning... for $8 a speaker, it was just dumb how well they sounded! and at 3"? Get outta here...

So, I got 8 more. Have been working on a design for towers to house 4x each. I went with the "go big or go home" idea, but obviously, keeping the 3" drivers... (insert reference of turbocharged 4 cylinder engine VS big V8 here).

I liked how the single driver sounds in the nanotowers, so I wanted to keep that same formula... roughly 1/2 a cubic foot (14L) for the single driver. That means these towers would have to be massive when you quadruple that!

The hardest part of designing this enclosure was simply to get enough volume inside, while keeping it from being too large in any one direction. I ended up designing in some 45* angles on the front to keep the baffle as narrow as possible, while still getting the maximum volume out of the enclosure. This makes to the top-down profile kind of look like a bullet. I'm definitely not as experienced as some of you, so if you see something wrong with this design, please point it out! (see attached)

The design puts this right about 1.95ft^3 after displacement (55.2L). Two slot-ports in the back of 3.25" wide by 1.5" tall and 3.25" deep to get a rough tuning of 51.5Hz. A little EQ to boost 70-80hz range and this thing should see the low end drop off a predicted F3 of 46.5Hz.

The enclosure will use 1/2" MDF (except on the 45* angled pieces, those will be 3/8" for ease of installation and design). This should maximize internal volume and should be strong enough to take these small speakers (I've done the RS180's in large ported, well braces enclosures... these have nothing on those speakers, so it's gonna be alright! ;) ). Enclosure will be the full 48" high, simply for ease of cutting the sheet. I will be putting in a full brace (reaching all sides, back, and front panels) every 1' (so 3 internal braces, in quarters of the enclosure: first one between speaker #2 and #3, one centered below speaker #4, and one between there and the bottom). I will also put some small bracing between the sides and rear and the tops and bottoms, likely triangular chunks. Depth overall will be 12" and width overall sits at 8".

Still no idea what I'm building these for. Just have the drivers sitting on my desk, screaming at me to make them sound as good as possible. Likely gift/sell ultimately, but should be fun regardless.

I'll get some REW graphs in my treated room after the build, see what these things do on paper for ya.

Let me know any opinions, etc. Not sure when I'm going to start these, gotta find time to get the wood, cut it, and start the assembly.

I had the idea of painting the front 3 panels piano black and veneer the top, sides, and rear.... but that's still in the air.
 

Attachments

  • Bullet4w3.jpg
    Bullet4w3.jpg
    26.8 KB · Views: 217
Founder of XSA-Labs
Joined 2012
Paid Member
Keep c-t-c spacing as tight as possible to reduce comb interference effects. Instead of making such a voluminous cabinet, if the drivers are all wired 2x2 (2 series x 2 parallel for equiv 8 ohms) they will act as a "single" driver with its 'center' located between 2nd and 3rd driver and act as a single driver with an equivalent Sd that is 4x the size of one the drivers (equiv to a 6 in driver). Now, design a mass loaded transmission line such as a folded TABAQ, or a TQWT, etc of your choice with the new driver. I think your cabinet will be smaller with a MLTL, have deeper bass extension, and sound better than just a design for a plain BR for 4 drivers. Although the design you have sketched will probably have some MLTL effects going for it as it is long, and provided you put the port at the bottom. Good luck.
 
Member
Joined 2011
Paid Member
Yes, I think designing a straight BR is not the best approach.

The reason the nanoTower works so well is that there is a lot of accidental transmission line action happening there. I have no idea how to actually design a transmission line, and I actually designed the nanoTower as a simple BR, but I got lucky because I wanted a very tall enclosure, so everything just fell into place and I ended up with a TL.

If you can (I still haven't been able to) figure out M.J. King's worksheets, design an enclosure as a TL from the start, as per xrk's advice.

That way you won't have to rely on luck.

Glad you like the nanoTowers! I still love mine, too.
 
Keep c-t-c spacing as tight as possible to reduce comb interference effects. Instead of making such a voluminous cabinet, if the drivers are all wired 2x2 (2 series x 2 parallel for equiv 8 ohms) they will act as a "single" driver with its 'center' located between 2nd and 3rd driver and act as a single driver with an equivalent Sd that is 4x the size of one the drivers (equiv to a 6 in driver). Now, design a mass loaded transmission line such as a folded TABAQ, or a TQWT, etc of your choice with the new driver. I think your cabinet will be smaller with a MLTL, have deeper bass extension, and sound better than just a design for a plain BR for 4 drivers. Although the design you have sketched will probably have some MLTL effects going for it as it is long, and provided you put the port at the bottom. Good luck.

Gotcha. Will move those drivers a bit closer together. Was planning series/parallel wiring for 8 ohm per speaker, so that works well.

I'll have to do more research on the enclosures you suggested. My current design puts the port on the lowest possible point of the enclosure (slot port, using the bottom of the enclosure as part of the port.)

Yes, I think designing a straight BR is not the best approach.

The reason the nanoTower works so well is that there is a lot of accidental transmission line action happening there. I have no idea how to actually design a transmission line, and I actually designed the nanoTower as a simple BR, but I got lucky because I wanted a very tall enclosure, so everything just fell into place and I ended up with a TL.

If you can (I still haven't been able to) figure out M.J. King's worksheets, design an enclosure as a TL from the start, as per xrk's advice.

That way you won't have to rely on luck.

Glad you like the nanoTowers! I still love mine, too.

Wish I had more time w/ the nanoTowers. I made them as a gift and had to ship them off for Xmas. I'll definitely look further into the design of the TL enclosure... but I have a feeling it's gonna make me go cross-eyed! :)
 
Founder of XSA-Labs
Joined 2012
Paid Member
The MLTL can be designed following some available designs from the TABAQ and (WIBAQ - not proven yet) threads. Bjohannesens's sims for the 3 in and 4 in drivers typically have a 30 in length with a TL cross sectional area of 4 x Sd. I think as Sd gets bigger, this 4x ratio can get smaller, for example it is less for the 8 in BetsyK driver in the WIBAQ. One thing about using 4 drivers in unison is that their Fs is still that of a 3 in driver. So that makes setting the length easy = 30 in. You could probably push it a couple of inches to 32 or 34 inches no big deal. Now set the cross sectional area as perhaps 3 x Sd_of_four_drivers. Sd of one driver is 32 cm^2 x 4 = 128 cm^2. 128 cm^2 x 3 = 384 cm^2 or 59 in^2. If you set the front baffle width to say 5 in then the depth of the cabinet is 12 in (all internal dim). Mount the drivers with the intersection between the 2nd and 3rd driver located at 1/3 from the closed end or 10 inches in this case. Without running an MJK sim to calculate the port length and dia, resort to WinISD BR modeling program to calculate a round port. The TABAQ typically uses a port area equal to about 0.25 x Sd, or in this case, the same as Sd for one driver. So you need a 2.5 in dia tube to serve as a port. Your cabinet volume is 30 in x 59 in^2 = 29 liters. Set tuning freq to 67 Hz which is typical WinISD, the real freq with mass loading will be around 50 Hz. WinISD says that a 2.5 in dia tube port should be 1 in long with 4 of these drivers in a 29 liter box. WinISD also says that the system is now a 91 dB system (not bad for equivalent of a $40 driver) Get one of those adjustable length port tubes and you can play with the tuning to optimize it. I know this is a round about way to do it but it will at worst case give you a BR box, and with the MLTL kicking in, you will get nice lower freq bass extension. You need to put stuffing in first 20 in of the line from the closed end. Good luck!

Bjohannesen may cringe at how I did this, but if he runs sims I wonder how far off they will be? ;)

To recap, the design is a 30 in tall x 5 in wide x 12 in deep box with 4 drivers centered at 10 in from top, use a 2.5 in dia x 1 in long port located at the bottom, stuff top 20 in of line with polyfill (pillow) stuffing. If you are using these near a wall, I would aim bass port at wall to get additional wall enhancement of bass. If not, aim bass port to front. You may need BSC circuit, but can be EQ'd if you are driving digitally with computer or mp3 player.
 
Last edited:
Thanks xrk. I've been trying to do some research on TL/MLTL... I'm very visual and there are plenty of articles, but man... my ADD-like brain can't get through a few sentences w/o feeling overwhelmed!

Looking at my current design (with drivers pushed closer together and to the top), there is ~40" to the port at the bottom. Just to get an idea of what's going on with that, it looks like it would be a little too long for the "proper" 1/4 wave of these driver's FS of around 100Hz? In a design like a MLTL, are you basically building a ported enclosure while factoring in the distance from driver to internal port opening? Sorry if I seem to be asking some pretty low-level questions, but I want to be sure I understand how it works. I've build a few hundred boxes in my days, but all were designed with sealed and BR. Have build a folded horn for some PA subs, but definitely didn't design em.

Something I don't understand about these LT/MLTL setups are the airspace within the enclosure. I get the length requirements (I think), but what actually determines what the enclosure volume should be?

I'm definitely down to learn more, but reading some of this stuff is like reading Latin. ;) I do appreciate the help.
 
Short version:

-Sd has zero correlation to the CSA of an MLTL. Martin merely used it as a convenient way of expressing an area, which with hindsight was a mistake since it merely served to perpetuate myth. In later versions of his worksheets, he changed the data input to physical WxD dimensions.

-An MLTL is not necessarily smaller or larger in volume than a bass reflex speaker. It conforms to the laws of physics and there remain certain minimum volume requirements. It simply takes the eigenmodes (standing waves) present in an enclosure into account and actively uses them, usually to allow a lower Fb than would be possible in a BR while maintaining a reasonably linear response. In practical terms, the crossover point between a BR & an MLTL can be said to occur when one dimension is stretched sufficiently for the standing waves to change the vent tuning from what Helmholtz maths (which assumes a uniform air-particle density within the enclosure & no standing waves to be present) says it should be.

-There is no one alignment for an MLTL, just as there is no single alignment for any other type of speaker. It depends what it is you are trying to do. You can use them to force additional LF extension, but as a simple ROT, avoid tuning below 0.707Fs. The drivers won't thank you for it. These days, I won't vent anything smaller than a 4in driver. IMO, it's not worth the effort. You can sometimes obtain impressive extension, but the dynamic range is usually crippled in the process as HD rockets, rendering most such designs only suitable for very small spaces & undemanding material.
 
Founder of XSA-Labs
Joined 2012
Paid Member
Thanks xrk. I've been trying to do some research on TL/MLTL... I'm very visual and there are plenty of articles, but man... my ADD-like brain can't get through a few sentences w/o feeling overwhelmed!

Looking at my current design (with drivers pushed closer together and to the top), there is ~40" to the port at the bottom. Just to get an idea of what's going on with that, it looks like it would be a little too long for the "proper" 1/4 wave of these driver's FS of around 100Hz? In a design like a MLTL, are you basically building a ported enclosure while factoring in the distance from driver to internal port opening? Sorry if I seem to be asking some pretty low-level questions, but I want to be sure I understand how it works. I've build a few hundred boxes in my days, but all were designed with sealed and BR. Have build a folded horn for some PA subs, but definitely didn't design em.

Something I don't understand about these LT/MLTL setups are the airspace within the enclosure. I get the length requirements (I think), but what actually determines what the enclosure volume should be?

I'm definitely down to learn more, but reading some of this stuff is like reading Latin. ;) I do appreciate the help.

The 40 in length could work, but if you start to apply mass loading (constricting the output vent) the frequency of the resulting MLTL may dip too low for the driver to handle. As the TABAQ simulations have shown that only a 30 in length is needed to reach 55 Hz with mass loading. The point of the MLTL is to push the tuning freq below the driver's Fs. This can be done for moderate to high Qts drivers. Bjohannesen's simulations (and my measurements) indicate that 50 to 55 Hz is possible with 3 inch class drivers in a MLTL. Scottmoose has very good points regarding the tuning to be no lower than 0.707 x Fs - there may be some harmonic distortion at these lower frequencies - but you may not mind it. The cross sectional area of MLTL is independent of the driver Sd to a certain extent (but it would not make sense to have CSA of only 5 cm^2 for a 50 cm^2 Sd), but if you can't use the MJK model yourself to run a simulation, then using the existing TABAQ designs as a basis to start with is not a terrible way to go.

You should play around with an easy to make foam core speaker to try this first before cutting any wood. A 30 in x 6 in x 8 in deep box with 4 cutouts and vent is very simple to do with 2 sheets of $1 ea foam core. I might try this design myself with 4 Vifa TC9FD drivers just to see if the concept works.
 
The ultimate length of the box isn't too much of an issue for FB per se, since you simply apply greater levels of mass-loading with a shorter pipe. I really wouldn't want to tune so low though, irrespective of the the type of enclosure.

There will be rising HD & reduced dynamic range, because that's just a matter of simple physics. On highly compressed music, you may not notice it since it's already been ruined in any event. With multiple units, these issues are compensated for to an extent, but these things are still relative, since if used in a short array, the temptation is to make a physically taller box & tune lower, thereby countering the gains you made previously.

No, the CSA of an MLTL is not independent of Sd to a certain extent, it has no relation to it, period.
 
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
No, the CSA of an MLTL is not independent of Sd to a certain extent, it has no relation to it, period.

Maybe what I am asking is not the physics-based relationship which there is none as you say, but in designing an enclosure, what do you use as a guideline or rule-of-thumb to size the CSA to match a driver? Something along the lines of the formula used to size a BIB: a function of Vas and Qts perhaps?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.