Once the prototype is built and populated (drivers installed, fiberfill inserted in the enclosure, ports cut to size and installed) the final testing proceeds. There are two sets of things being done.

1) Verify the port length calculations. 

2) Do driver frequency response tests designed to let us prototype the crossover.

Port Length Verification

To verify the port length we simply test the driver's impedance with the port installed. The tested impedance was

This is for two drivers in series. If the two drivers are identical we should see 2x single driver results. Note the typical two-hump ported impedance curve.

Next we do a near field frequency response curves for the driver and for the port. To test a woofer nearfield the microphone is placed as close as possible to the dust cap. The port nearfield is tested by putting the microphone in the center of the port output tube flush with the enclosure wall. Don't change volumes between the two tests (set the volume so both tests are within range then do both tests). 

This produces results that are frequency response accurate at the low end up to 200-500 Hz for standard woofer. The near field results for the PL18 are

The red curve is the port response, the black curve is the (amazingly flat) driver response. The large glitch at 60Hz is obviously power supply ripple. 

 To see what the final results are we do a Merge Port Response command (Driver / Merge Port). The merged results (compared with prediction) are

Note that our bass response is not as predicted. This is due to the port length being off a bit. We'll tune the port more carefully for the final product when the crossover is final and the series impedance is known. Just FYI the sealed box is down 8.5dB at 45Hz so even mistuned the port is helping.

Next we do on-axis frequency measurements. To do these bring up the driver and do a Pulse response test. It looks like this

The first marker should be set to just before the pulse and the second marker can be set anywhere, usually before the first reflection. Here the first reflection is very early (using the distance calculator we see that is about 9 feet, which seems right for a stand mounted speaker with floor bounce). In general I use a much wider marker setting and ignore the low-end results and visually integrate the noise.

The on-axis results for the woofer and tweeter are done using the same volume levels (so they can be merged) and produce:

Note that the tweeter efficiency is greater than the dual woofer efficiency (not surprising) and that needs to be included in the crossover design.