|
Speaker ProjectTesting the Subwoofer (1) |
The first step in testing a subwoofer driver is to find the impedance of the driver. This will tell you the resonant frequency (FS) and all of the T/S parameters (in particular QTS).
Before beginning, make sure that your jig (or series resistor) is working well by measuring a passive component or two. Do this by creating a generator then selecting Measure / Passive Component. This will also verify that your volume is set appropriately.
For testing the woofer, I set sample rate to 11KHz (as low as possible, because we only care about the low end here, and this maximizes the frequency resolution) and sample size to 128K (here we really care about frequency resolution).
The first impedance we find is the Impedance in Free Air. To find it,
1) Create a new driver and give it an appropriate name (mine is NHT1259).
2) hook the subwoofer up to the impedance jig (or series resistor).
a) The woofer frame should be as stable as possible while keeping air resistance to a minimum. The best solution is to have two tables - one on each side of the woofer and clamp the frame to the two tables (two clamps, therefore). This keeps the woofer in free air, clamped, and reduces any outside interference.
b) If that is too tough for you, have someone click the Impedance in Free Air menu option while you push down on the frame while the woofer is on a small table. Keep your arms away from the cone. Cheap and dirty and not ultra-precise.
When you see my first-pass impedance chart you will see shakiness in the low frequency region. Some of that may be from inadequate clamping. Some of that is normal.
3) Go to Options / Preferences and set the Sample Rate to 11000 and the sample size to 131072 (128K). You can use a smaller sample size if you have a slower CPU (I use a PIII-500) but keep the sample rate low so that the low frequency accuracy is maximized.
3) Open the driver resource (or click to it) and select the Measure / Impedance in free air command. This will run a short MLS signal (a 65000 sample signal at 11000 will take about 6 seconds) and then it will analyze the resultant signal.
After doing this, if I just open my measured signal (named NHT1259.Free Air) it looks like:
The problem here is lots of noise at the very high end so click the phase axis (precisely on the axis line) and you will get a rubber box around all of the axes with handles. Drag the right hand axis to the left - down to around 1K - and the chart will now look like this: [note: if you have trouble dragging the box around you can just go to Chart Properties via the right mouse button menu or the Edit menu or just click F4 and change the X axis min/max manually]. Another way to do this is just right click near the peak impedance and select Zoom In.
You can now see the impedance peak clearly. The resonance point is where the impedance peaks (and phase is 0). For the NHT that looks like about 20Hz.
To get a first estimate of the driver,
1) pull up the driver properties dialog (open the driver resource - which will be a blank display for now - right click in the window and select Properties, or just select the Edit / Properties command).
2) Go to the Data tab and enter a DC resistance and click the Use this DC Res button. I used a good quality voltmeter to measure the DC resistance, but most factory specs are pretty accurate. Speaker Workshop can not accurately estimate the DC resistance so this step is crucial for good results.
3) Click OK to close the driver properties dialog.
4) Select the Calculate / Estimate Parameters... command and use an appropriate frequency range. For me that was 10 Hz to 2500 Hz (clipping that high frequency noise out of the estimate). Make sure you make the upper cutoff high enough to show plenty of impedance or the estimates of the inductances could be pretty far off (which will not affect things much when it comes to designing a box).
5) When the process completes you can do a few things to check your estimate. They are enumerated on the next page.
last updated August 27, 2000