measuring t/s: voltage

[e-side]

Hello everyone,

currently I'm measuring some 15" woofers. The problem is that I get very different results when I use different source voltages.

The first time, the amplifier provided 0.75 V at 200 Hz. I got a Qes of 0.42 and a Qts of 0.37, so the driver seemed to be suitable for a vented box.

The second time, the amplifier provided 0.99 V at 100 Hz. The results were quite different from the first time, a qes of 0.71 and a qts of 0.63 . With these results, the driver would be suited for a closed enclosure rather than a vented one.

Can anyone tell me what voltage I should use when measuring the t/s parameters? I used the guidelines provided at http://www.diysubwoofers.org/measure.htm and http://sound.westhost.com/tsp.htm

best regards

Erwin

[Ron E]

There is no set voltage that you perform T/S tests at.

Most texts recommend a different procedure (constant current rather than constant voltage) from the links you specify. Do a search for 'Thiele Impedometer' and you will find an article on epanorama with the typical procedure.

T/S parameters vary with level, but they should not vary as much as you describe over such a small range of voltage - I would suspect the measurement method or perhaps you made an error.

[Vikash]

Just repeating really, but T/S do change with voltage but not to the degree your talking about over a relatively small voltage increase.

This is the kind of situation where the 'repeat count' functonality and consistency of Speaker Workshop come in to its own (as opposed to manual method)

[e-side]

Originally posted by Ron E
Do a search for 'Thiele Impedometer' and you will find an article on epanorama with the typical procedure.

Currently I'm experimenting with this way of measuring, and I just got another question: To obtain a certain voltage across the 'calibration resistor', the signal generator will have to produce a signal (of couse). How many Hz should this signal be to get reliable results?

For example, should i use a 100 hz signal to obtain a certain voltage across the resistor, or a higher/lower frequency?

regards

Erwin

[richie00boy]

I use a 50Hz signal as that's where most meters are most accurate. If you are using the voltage divider (pseudo current source) method, I use a 10 ohm resistor as the calibration resistor and develop 100mV across it.

It is also essential IMO to perform a calibration sweep across a few different frequencies to see how your meter reacts. Most do not stay the same.

As others have said you must have some other factor influencing your results because they should not change that much.

[e-side]

Originally posted by richie00boy
I use a 50Hz signal as that's where most meters are most accurate. If you are using the voltage divider (pseudo current source) method, I use a 10 ohm resistor as the calibration resistor and develop 100mV across it.

I used the same voltage as the Re of the calibration resistor (9.2 mV), at a 100 Hz signal. I'll try your settings the next time, thank you for the info.

Originally posted by richie00boy
[B]As others have said you must have some other factor influencing your results because they should not change that much.

I don't know what caused this difference in values. I always check everything while measuring, the only thing I changed from previous measurements was the voltage setting.

The point is, if the t/s parameters change with source voltage, what voltage would be the best to measure with...

best regards

Erwin

[AndrewT]

Hi,
it's a long time since I did the testing but I recall that the constant current method is done at very low voltages. T/S parameters are for small signal only. The speaker voltage (after the series resistor) at resonance was about 500mV and less than 100mV at medium frequencies.

The frequency to check your equipment should be in the long low impedance saddle above the resonant peak of the driver.

When you do a sweep it is easy to find the peak and also the low impedance above it. After a couple of octaves the impedance starts to rise again. This is then too high a frequency.

How have you solved reading the voltage at the very lowest frequencies? i.e. below resonance.

[e-side]

Originally posted by AndrewT

How have you solved reading the voltage at the very lowest frequencies? i.e. below resonance.

Actually nothing (if I understand the Q correctly), the multitester just gave the voltage level (scale at the multitester was set to 200 mV).

The constant current method (I just follow the guidelines from the epanorama article) gives accurate results in my case, so I'm able to get reliable t/s parameters now.

Thanks everyone for advice!

regards Erwin

[DanWiggins]

Personally I like to measure at a drive level where the average output of the driver in the range of interest will be ~85 dB SPL.

For example, for a subwoofer driver, I'll do a few FR measurements and adjust the drive level until the 30-100 Hz range averages 85 dB. Then I measure at that same drive level.

Likewise, for a midwoofer, I'll set the drive level so that the average SPL from 80 Hz to 3 kHz is 85 dB.

I believe this way you end up with T/S parameters that are consistent with typical medium-loud operation, and will give a different recommended box than what you would get from small, 10-50mV drive levels.

Dan Wiggins
Adire Audio®

[AndrewT]

Hi DanW,
I have never tested at high voltage.
Which parameters are most sensitive to the different test voltages?
By how much do they differ?

All,

Accurate low frequency voltage readings???
My multimeter cannot pick up consistent rms voltage readings at frequencies around 15Hz and below. F(lower) is often well below 20Hz. I tried using the scope set to no timebase to pick up the peak voltage and convert back to rms. This is not particularly accurate and leads to big changes in parameters when a spread of F(lower) are inserted into the spreadsheet.