deriving T/S parameters from common specs

[hitsware]

The loudspeaker shall be a 6.5” extended range coaxial transducer, and shall meet the following
physical specifications:
Diameter: 6.5”
Mounting Diameter: 6-1/8”
Depth: 3”
Weight: 2-1/2 lbs.
L.F. Cone Construction: Rubber roll suspension

Magnet Circuit Type: L.F. SEALER-LINEAR, 12 oz. Ferrite H.F. 1 oz Ferrite. Flux
Density: L.F. 11,000 Gauss. L.F. Voice Coil Diameter: 1”. Rated Impedance: 8 ohms. Rated
Program Material Power: 50 watts. Rated Power (RMS): 25 watts. The loudspeaker shall meet the
following acoustical specifications: Sensitivity: 90 dB above .0002 Dynes/CM2. Resonance: 45 Hz.
Frequency Response: 30 to 20,000 Hz (+/-6 DB). Crossover Frequency: 3,000 Hz.

anyone dare a formula ? .......

[rjbond3rd]

Howdy hitsware, an interesting puzzle. I'll go out on a limb and say it can't be done but I'm speaking as a newb.

At a minimum, we need Fs, Qts and Vas, yes? Fs=45Hz. Maybe there is a way to determine or guesstimate Qes but how would we ever derive Qms and Vas from the data provided? how would we ever know how stiff the suspension is?

Maybe if we were to take the frequency response as gospel, there might be a way to "back into" certain minima or ranges for Vas and Qts. I guess that's the whole point but believing the frequency response would be a great leap. Still, an interesting challenge. Happy New Year!

[HK26147]

Drivers are not made with ultra-tight tolerances. Because it is not uncommon to have T/S parms 15% off of manufacturer's published specs; I measure the T/S parameters.

Syd

[rjbond3rd]

Okay, actually I believe there is a way.

Take a frequency response graph (on paper), and draw the quoted response. (Yeah, this is all ridiculous but let's go with it.) So now, for each frequency (i.e., point along the curve), you have the decibels of SPL. Let's ignore whether this is in an enclosure or an IEC baffle for the moment.

We have Fs as 45Hz. Now we want to find possible values for Vas and Qts. We'll use the formula:

G(Fr)=20*log10(Fr^4/Sqr((Fr^4-Fr^2*A2+A0)^2+(Fr*A1-Fr^3*A3)^2))

Where:

* A0=(Fb*Fs)^2/Fs^4
* A1=(Fb*Fs^2/Qb+Fs^2*Fs/Qts)/Fs^3
* A2=(VAS*Fs^2/Vb+Fs^2+Fb^2+Fb*Fs/Qb/Qts)/Fs^2
* A3=(Fs/Qts+Fb/Qb)/Fs
* Let's write Fr=F/Fs where F is the frequency of interrest
* G(Fr) is the SPL in dB

For the formulas and more info: http://homeswinghome.free.fr/audiore...bassreflex.htm

Now you just solve for Qts and Vas. Probably it's easiest to plug in a guess for Qts and solve for Vas. Whaddya think?

[hitsware]

Awkward. I'll have to digest that some.

[rjbond3rd]

Awkward to crunch by hand, but just imagine -- you would be deducing the missing info, specifically the Vas (with a guess for Qts). I think that's cool.

But obviously just an exercise given the (outrageous) assumptions required. I guess the conclusion is that given an incomplete (and almost certainly unreliable) set of specs, yes, you can theoretically /derive/ some missing values, you just could never /rely/ upon them.

But maybe someone else has another way -- there's always another way.

[hitsware]

These are ceiling drivers.
Simular to car drivers
in the sense that they are
usable to many sans the
'specs' that we like to
dwell on.

[mwmkravchenko]

Here is a link to a great program to help you figure things out

http://home1.stofanet.dk/cfuttrup/dpc.htm

It's a bit lacking in the interface but it will do the trick.

Mark

[hitsware]

https://www.wmballen.com/ProductDeta...product_id=253

What happens @ 15 Watts ?
What does the 'frequency response' imply ?