Measure T/S.. Insanity check not equal to Fs HELP

[gjeff80]

Ok, can you guys help me. I know I'm doing the steps correctly because I measured two seperate woofers (using steps here http://www.diysubwoofers.org/measure.htm) and one comes out fine, but the other I am having a problem getting the Fl value to check. What should I do for this one woofer? Here is what I end up measuring. I determine Fs to be 37Hz, and then when I find the upper frequency that has the same -3dB voltage, it is 63Hz, and the I find the lower frequency that has the same voltage and I get 15Hz, BUT when I use those two numbers, the Fs (sanity) Check value comes out as 30Hz, what should I end up doing here? I know I'm doing the procedures correctly, because I used another woofer, and the FsCheck value came out withing 1Hz of what I measured. Should I just ignore the Fs insanity check value in this case??

Does anyone know if these values would sound about right for a Cerwin Vega 189 E - 18 Inch woofer?

Qts- .28
Vas- 165.0l
Fs- 37.00
Re 4.56
Qms- 1.25
Qes- .37


Thanks
Glenn

[Ron E]

When the sanity check does not work, use Fs=sqrt(Fh*Fl) in the formulas.

Vas seems a bit low, show us your calculations.

[gjeff80]

Ok, I ended up running through the entire process again, using this method:

http://www.epanorama.net/documents/a...arameters.html

Here is a snapshot from the excel spread sheet I whipped together to calulate the formulas using this method. The specs are a little different, but kind of close, How do these look?

Note, I used liters for the box volume, and the voltage is in miliVolts.

[Ron E]

You need to replace Fs with sqrt(Fl*Fh) in the Qms formula when Fs<>sqrt(Fl*Fh), otherwise errors in Qms can get rather large - in your case ~10%. The differences in peak symmetry are caused by voice coil inductance.

Both methods have issues with frequency responses of meter and amplifier used. Most meters under $100 roll off below ~40Hz. You can correct, but it is tedious, you need to recalibrate the impedometer at every frequency.... One way to doublecheck that this is not an issue is to use a CD player and Test CD with low frequency tones. Most (component, not PC) CD players are dead flat to at least 10Hz.

The proper way to calculate Vas in a sealed box is to use:
Vas = Vb*(Fc*Qec/(Qes*Fs) - 1), where Qec and Fc are the measured "Qes and Fs" in the test enclosure.

[gjeff80]

Hmm, I thought the resonance frequency was the frequency at which the voltage was highest across the terminals of the woofer. When I used the other method to measure the T/S specs where you measure the voltage across a resistor and you look for the lowest voltage I got 37Hz. But when I use sqrt(Fl*Fh) it gives me about 30Hz, but when I check the voltage at 30Hz, it isn't the highest reading voltage (or lowest using the other method).

Glenn

[Ron E]

Yes, the resonance is very close to the frequency where impedance is highest, depending on a number of factors like box losses, loading conditions, etc.

I am not saying that your T/S resonance frequency is sqrt(Fh*Fl). I am saying to properly calculate Qms, you need to make that substitution under the circumstances you are experiencing - assuming you don't have the frequency response issues with your test setup that I mentioned.

Said another way - when Fs is not equal to sqrt(Fh*Fl) and there are no issues with the test setup that can cause this, it is due to voice coil inductance and to properly calculate Qms you must use sqrt(Fl*Fh) instead of Fs in the Qms formula to get a more accurate value for Qms.

My reference is J.E. Benson from 1969 from the book "Theory and Design of Loudspeaker Enclosures". The reason you don't see this correction very often is that few hobbyists have read Benson.....

[gjeff80]

Thanks Ron,

So what would actually be the Fs of the woofer? In the excel photo I showed, would I use the value of FS, or use the value of 30 (which I got when I did sqrt(Fh*Fl)), in this excel sheet, that is the formula that I used when it calulates Qms

[Ron E]

As far as T/S parameters go, Fs is the peak frequency.

Fs=Fs

Using the notation from the epanorama site:
If Fs is not equal to sqrt(F1*F2)
Qms=sqrt(F1*F2)*sqrt(Rc)/(F2 - F1)

If Fs=sqrt(F1*F2) (or is within a small fudge factor)
Qms=Fs*sqrt(Rc)/(F2 - F1)

Qes=Qms/(Rc-1)

Vas calculations are the same....