DIY Centre Speaker advice

[bjorno]

here's a laugh for ya's put the woofers in a cardboard box…

No laughs, this is a DIY WOA model that might be very functional in the end. 🙂

Rec. In order to reduce the M to T distance; See picture; 1(2).
Decide for an appropriate even order crossover frequency (acoustic) between the M and T as low as possible.

Where you put your lower dividing depends on the c-c distance between the centerline of M/T and the W drivers, use Tolvans Xdir program to choose an odd (at Woofer level (dB) = - 3 dB, Woofer phase lag = 90 degrees)) or even order filter (at Woofer level (dB) = - 6 dB, Woofer phase lag = 0 degrees)) and simulate the MT as an T driver in the M-T-M configuration.

Maybe a surprise for you; Normal closed box for this driver is a rather large size and one driver requires for a Qtc of 0.7 about a volume of 42 L but in a Aperiodic box the volume can be reduced to a minimum of 20 L if there is sufficient headroom dimension for a 4” x dia 7” port.

I modeled your suggested VifaBC18SG49-08 in a minimal box that happens to be a Damped Box Damped –Vent System (Aperiodic) in order to avoid a too large box and a peaking low end.

The Stuffing density in the totally filled port is 32 kg/m^3 about 0.054 kg and the box should be filled with12.8 kg/m^3 = 0.125 kg stuffing.

A box with the outer dimension of 15.5”x11”x10.75” with the internal dimension of 14”x9.5”x9.25” will do and must be doubled for a WMTWW box i.e. to 31” + 5”(width of MT enclosure) = 36” in total.
See the simulation 2(2).

b

1(2)

 

[bjorno]

2(2)

 

[sunrise3500]

bjorno, i think it's pretty funny, although yes functional 🙂

although back onto serious matters now 😉
well now that I've got the woofers, I think I'll get the vifa dx25 tweeter, (unless opinions are D27TG or D26NC is better, but from the reviews I've looked at, they aren't)

and assuming I go with a W MT W design... for the midranges either
Vifa TC11WG-49-08
http://www.speakerbits.com/net/catalogs/showpic1.aspx?ID=VF110B

or Vifa C13WG-08-08, which is what I'm thinking of going with, appears better specs and better matched to the woofer?
http://www.speakerbits.com/net/catalogs/showpic1.aspx?ID=VF130C


ps if it helps, the woofers i've got are http://www.madisound.com/pdf/vifa/BC18SG49-08.pdf

 

[bjorno]

..or Vifa C13WG-08-08, which is what I'm thinking of going with, appears better specs and better matched to the woofer?..

A 4.7 L closed box for the Vifa 13WG-49-08 and the Vifa X25 looks good.

F3 is about 85 Hz/86.2 dB/Qtc=0.9,ref 8 Ohm and at 250 Hz the level for 1 W is about 89.8 dB and matches only the Vifa BC18SG49-08 at 91.6 if in series and with a slightly lossy low pass filter.

Maybe you can consider the Silver Flute 5-1/2" W14RC25-04ohm 5-1/2" in a 4.7 L box instead?

The Silver Flute is at 94.4 dB/250 Hz ref 8 Ohm and probably close to 95 dB at higher frequencies making it much easier to match the Vifa T driver and at the same time also with paralleled Vifa W drivers the overall Z will be lower, making more sensitive center speaker.

http://www.madisound.com/cgi-bin/index.cgi?cart_id=6214169.4678&pid=1075

b

 

[sunrise3500]

Thanks for the reply bjorno, just wanted to check my understanding to know where you're coming from.

Basically, you're saying i could go with that 8ohm midrange and it'll match fine, and the only reason you've suggested is that other one is that being 4ohm, can run the woofers in parallel with the midrange in series, so the woofers get half the power, midrange gets the other half. So with their Woofers sensitivty each at 87db, together they yield 90db (asuming im remembering what i learnt back in high school correctly), and the midrange matches this.

Opting to use the 8ohm midrange, and having the woofers in series, woofers get 1/3 of the power, mid the other 2/3, and together the woofers give 87db, so im losing out 3gb - half the power.

If i've got this correct, that makes the 4ohm a lot more attractive, and i've contacted madisound to find out how much shipping is to australia, assuming they can do that. 🙂

edit: and when you say, 94.4 dB/250 Hz ref 8 Ohm etc i have no idea what you're talking about.

 

[sunrise3500]

also, i think 5.5" too big for my liking, was thinking something with a cone more around 4" was better suited? i think the box would be too massive with a 5.5" mid in the middle lol
unless 5.5" is the with the chassis and everything ~

 

[bjorno]

Basically, you're saying i could go with that 8ohm midrange and it'll match fine, and the only reason you've suggested is that other one is that being 4ohm, can run the woofers in parallel with the midrange in series, so the woofers get half the power, midrange gets the other half. So with their Woofers sensitivty each at 87db, together they yield 90db (asuming im remembering what i learnt back in high school correctly), and the midrange matches this.

Two Woofers in parallel will yield a + 6 dB increase of the SPL level using the same reference where 1 W equals 2.83 V RMS over an 8-Ohm load, the reference for the suggested M driver but the sum sensitivity stays the same if in series.

This is only valid for low frequencies; typical for wavelengths longer than 8 x the W-W c-c distance,

The minimum c-c you can use is very close to (lamda/8=228 Hz) and the suggested crossover frequency at 250 Hz thus will work well.

Now a parallel Woofer sensitivity of 87 + 6 = 93 cannot match your suggested M driver = 88 dB.

For this case you must choose to put the Woofers in series = 87 dB sensitivity, but if you really want to use the higher sensitivity that comes from paralleled drivers, you must choose a more sensitive M driver that will deliver a matching SPL level as the one I suggested.

The rec. Silverflute sensitivity is 90 dB SPL/1W/1m/4 Ohm (1 W equals 2 V over a 4 Ohm load) will increase the SPL level 3 dB (10 log 2.83^2/2^2) if driven by a 2.83 V signal thus will match the paralleled woofers with a equivalent sensitivity of 93 dB.

…when you say, 94.4 dB/250 Hz ref 8 Ohm etc i have no idea what you're talking about……

This refers to the driver-in-a actual- box-sensitivity for the 4-Ohm Silverflute using not a 2 V reference Voltage but instead the 2.83 V over 8 Ohm.

….i think 5.5" too big for my liking, was thinking something with a cone more around 4" was better suited? i think the box would be too massive with a 5.5" mid in the middle lol....

Of course a 4” M driver is more space efficient, maybe you can find a driver with similar specification as the Silverflute?

b

 

[sunrise3500]

ah ok, so then the mid im looking at is actually fine - was thinking ill have to drop it its sensitivity, and i dont have to drop the tweeter sensitivty as much as i originally thought.... thanks for the reply, pain in the *** finding a 4ohm midrange but ill keep looking.

and got a reply for shipping on the silverflute, $50 US, lol, rip off.

 

[sunrise3500]

Looking at using the midrange i mentioned (C13WG), and if indeed I'll get a 6db gain (after the 3db loss), looking at the frequency response chart for the woofer, looks to me almost spot on with the midrange?
If I was to have the crossover at around 480hz, they look to be at around the same db (90db), and I'll get to miss the dip in response at ~400hz on the midrange.
Or would that frequency be too high to crossover at - at what frequencies do you get the problem with MTM's?
Also, if I drop the tweeter by 5 or 6db or so, it should match fine too.

So they seem great to me, other than the advantage with the woofers are in parallel if with a 4 ohm midrange instead - greater db sensitivty. Although my other speakers are 87db sensitivity i think anyway.. so i'll have to set -3 for the center anyways going with this.

 

[sunrise3500]

and another question...
am I better off doing something to drop the tweeter's sensitivty by 3db or whatever it is (on this subject, how would I do it.. i could use a resistor, but something tells me that's not the most efficient way?), or am I better off getting a tweeter with matching sensitivity, or is it really much of a muchness?

thanks

 

[bjorno]

Looking at using the midrange i mentioned (C13WG), and if indeed I'll get a 6db gain (after the 3db loss), looking at the frequency response chart for the woofer, looks to me almost spot on with the midrange?If I was to have the crossover at around 480hz, they look tobe at around the same db (90db), and I'll get to miss the dipin response at ~400hz on the midrange.

Hi, I’ve been busy for a couple of days….

I advise you never to compare FR curves at fractional dB levels especially before you haven’t decided on physical speaker dimensions and taken into account the room influences.

The data given is more reliable to use and the first approach rec. is to choose decreasing sensitivities starting from the tweeter.

It’s on your own risk to extrapolate by comparing different scaled measurements that looks suspicious.

The C13WG seems to be measured in a box as the lower end FR slope is more than12 dB/ octave and the about 420 Hz dip phenomena affects both on axis an off axis, it could be a box artifact not showing up later in your application.

Or would that frequency be too high to crossover at - at what

I recommended you to chose about 250 Hz for a W-W crossover for an obvious reason, see picture.

...and another question...am I better off doing something to drop the tweeter'ssensitivty by 3db or whatever it is (on this subject, how would I do it.. i could use a resistor, but something tells me that's not the most efficient way?), or am I better off getting a tweeter with matching sensitivity, or is it really much of a muchness?

The tweeter choice is excellent and a simple resistive divider is all what you need and in fact an asset when you should as recommended; Crossover as low as possible for the dispersion reason.

b

1(1)

 

[sunrise3500]

lol i like the "= Obvious!" comment in the picture

and well I'm still gonna try and chase up a 4 ohm midrange, and see what my alternatives are for my budget, but the C13WG is looking like will be my choice. I guess when I place my order, my selections are locked in (which will definitely be this week.)

 

[sunrise3500]

noticed these are available, and now considering them too, argh.

MP14RCY http://www.seas.no/seas_line/midrange/H422.PDF

and MCA11FC (which i think is a much better size, probably better crossed at 300hz though since it looks to only start to be flat at ~250hz, but i think the mp14rcy would suit better in terms of its sensitivty/response.)
http://www.seas.no/seas_line/midrange/H143.PDF

 

[sunrise3500]

Well I got the 25tg tweeter the other week
just need to lock in on the midrange
so which of the seas or vifa c13wg should i get? I'm currently leaning towards the mp14rcy, just don't particuarly like its size...

 

[bjorno]

...just don't particuarly like its size...

Hi,

Its up to you to decide, I vote for the MP14RCY anyway because of its quality, easy integration and by choosing the right volume, for a targeted Qtc value, behind you may also make savings on the otherwise mandatory HP filter when crossing over to the W drivers.

b

 

[sunrise3500]

Yeah well I figured either of the SEAS would be a better suit than the c13wg... and the mp14rcy looks like a *really* good match in terms of specs
screw it, on monday I'll just order the damn thing, hopefully I'll have it by end of week.

Once I get it, next will step will be the crossover.

 

[sunrise3500]

Alright, I've got all my speakers now. (the 2 vifa woofers, the vifa tweeter and the mp14rcy)
Now the crossover - this weekend is a long weekend, so would be good to do it then since I'll have some spare time, can pull out the soldering iron 😀

any suggestions on where to start? links? anything... cheers keeping in mind I don't know where to start, other than need to drop the tweeter 3db, and cross that high, and the woofers -> mids should be crossed low... heh

 

[bjorno]

I'll have some spare time, can pull out the soldering iron…

Hi, You could do that but then you probably have to help yourself and take the consequences when starting in the wrong end.

Who wants to help you with designing a speaker starting with a couple of drivers not knowing the exact built enclosure dimensions, the mean listening distance, the height where the center is placed and so on?

You also know that an added center should integrate smoothly with the main speakers and some further considerations concerning this matter must also be taken.

What’s first needed is information of the actual driver c-c distances, the amount of T driver recess and the baffle size.

So I advise you to start the actual box design.

The filter is important but much easier to design after the speaker is built.

Now the filter can be designed with the baffle-step, diffraction and last if necessary to include tonal corrective measures.

An imperative demand for seamless integration success is always that the driver inherent quality chosen for the center speaker must equal or supersede the mains.

The layman imagination that an optimal integrated center is a physical exact copy of the other mains and a necessary prerequisite for optimal tonal quality of a multi-channel set-up is nonsense and not backed up by any observation done or published in any research papers as far as I know.

I believe you have this prerequisite (of a good set) of drivers by now so I recommend you to spend some time at your drawing table visualizing your center and post your picture(s) later on for review of the content(s) in order to later aid your filter design.

The center tonal matching will rely on speaker design and the applied knowledge of the difference of listening at stereo phantoms or at real sound sources which the center speaker type belongs to and not in the first place the need for exactly copying the performance of the stereo placed mains.

The filter should ideally be compensated for the above-mentioned difference and if done right a perfect seamless soundstage is formed, invaluable when listening to music with all encoded formats, multi-channel distributed stereo included.

As an example of timbre matching making sense and free form DIY Urban Legends is taken from this paper:

J. Robert Stuart: THE PSYCHOACOUSTICS OF MULTICHANNEL AUDIO

See picture: 1(1)


http://www.meridian-audio.com/w_paper/multips3.pdf


But the main issue for incorporating a center speaker is for spatial distribution reasons even if closely linked to timbre matching.

The obvious apparent source size difference that can be experienced when moving a normal stereo dedicated speaker to for the hearing to an acute frontal position where the ASW (apparent source width) reflexes is quite different and normally shrinks mainly for the horizontal apparent size.

This even if the closeness of the visual screen is present augmenting the localization and expanding the dimensional experience, for this case not the size but from the multitude of stimuli’s affecting the hearing.

I’ve argued before that the size of a center should for that reason not be too small.

Further the auditory stream separation at the center and the perceptual dimension is also very different for a centrally placed speaker and too overlapping even though the close visual screen have separate objects the sound is less separated (to interwoven) than if stereo phantoms are produced.

Rec. read (at least) the ‘Discussions’ in: http://www.nue.org/~okuno/aaai99.pdf


b

1(1)

 

[sunrise3500]

looking at box design now. for the midrange, looks fairly straight foward.. the qas is gonna be the volume for the flattest response right?

when it comes to the woofers, I would like to try and extend the bass output, so I'm thinking of a ported design for that, but not really sure where to go with it 🙁

and tweeter is sealed, so dont need to worry about it

 

[bjorno]

Hi,

looking at box design now. for the midrange, looks fairly straight foward.. the qas is gonna be the volume for the flattest response right?

I guess you meant ‘qas’ = Vas.

No the flattest FR is for a volume of 1.5 L but then you must crossover at about 150 Hz using no HP filter for the MP14RCY and use a LP filter for the W drivers at that frequency. See picture 1(1) for more comments.

Make decisions and show your drawings.

The most important for this project is to make the T and M right, as you will later use the CSA to the left and right of the MT drivers and continuing the compartments for the W drivers.

For the MP14RCY you should pick a volume between 1.5-3.3 L (larger doesn’t matter as you then will use a 12 dB filter) that first can be done after you have decided how the W-M (T)-W integration affects the horizontal dispersion, i.e. if you follow my rec. to choose 250 Hz crossover (the ‘obvious’ statement) or a BW filter using the lowest possible volume = 1.5 L.

The minimum enclosure height is 228 mm and you should also decide the internal width for the MP14RCY compartment that cannot be less than 134 mm. See picture 2(2)

Make also decisions based on the fact that the W drivers should be placed as close as possible to the M (T) drivers.

when it comes to the woofers, I would like to try and extend the bass output, so I'm thinking of a ported design for that, but not really sure where to go with it

You’re becoming to vague, new requirements?

I’ve already shown some of the ported enclosure consequences for the bargain Vifa BC18SG49-08 in post # 62 and its up to you to make further decisions or simply post your new requirements and hope that somebody will help you.


I wrote:

Maybe a surprise for you; Normal closed box for this driver is a rather large size and one driver requires for a Qtc of 0.7 about a volume of 42 L but in a Aperiodic box the volume can be reduced to a minimum of 20 L if there is sufficient headroom dimension for a 4” x dia 7” port.

I modeled your suggested VifaBC18SG49-08 in a minimal box that happens to be a Damped Box Damped –Vent System (Aperiodic) in order to avoid a too large box and a peaking low end.

The Stuffing density in the totally filled port is 32 kg/m^3 about 0.054 kg and the box should be filled with12.8 kg/m^3 = 0.125 kg stuffing.

A box with the outer dimension of 15.5”x11”x10.75” with the internal dimension of 14”x9.5”x9.25” will do and must be doubled for a WMTWW box i.e. to 31” + 5”(width of MT enclosure) = 36” in total.
See the simulation 2(2).

Since I’ve got no response of my effort to help you with the Vifa BC18SG49-08 drivers and not even a comment of my exemplified findings how to use these in a ported or closed box I will not speculate trying to ‘extend the bass output’ as the gross volumes then needed for the W drivers will be ridiculous for a center speaker.

In my examples shown, the net volumes for two drivers will be around 45 L (closed) to 89 L (small ported aperiodic box) and I think this is large enough for a center speaker recommendation.


b

1(1)