900% Vas tolerance?

[Otter]

this one is something to think about...

I mounted an accelerometer on a woofer (Adire DPL12, sealed enclosure), and measured the input voltage on the voice coil and the output voltage from the accelerometer. If I calculate 20*log(output/input), I get the dB SPL curve with an offset. This should be theoretically correct I guess.

Now here comes the problem. I've measured the response in 200 different frequencies, and made a plot in Excel. I have a very nice 'peak' between 40 and 100Hz, as I expected (box too small).

To be more precise: Qtc = 0.35 (0.4 according to datasheet, measured 0.35). Box = 50 liters. 'Peak' = 5.5dB, which mathematically results in a Qtc = 1.81. Since I know Qtc, Qts, Vb, I can calculate Vas. Turns out Vas is 1420 liter, according to the datasheet this should be 149 liter.

What is wrong? It is not a measurement error (200 frequencies, very smooth curve), it is a method error. I just can't figure out what it is. Voice coil series/parallel connection doesn't matter, I already tried that. I also used a DMM instead of a scope, and vice versa. It always turns out the same.

Funny thing is I've used the same method on a different woofer in the past, which resulted in a Vas that was close to the datasheet value (less that 20% deviation).

Any suggestions are welcome.

[BassAwdyO]

maybe you got an extra zero somewhere.... 142 wouldnt be a bad measure. What is the equation you are using?

what are your values for Qts and Qtc?