Zaph|Audio - SB Acoustics 2-Way : Quasi-Anechoic Measurements

Measurements: Large Baffle (Quasi-Anechoic)

All drivers were first measured on my 2m x 2m baffle to get a quasi-anechoic frequency response before any break-in or modifications. This provides data down to about 250Hz as shown in the graphs below. I use LAUD to measure and the free version of Praxis (successor to LAUD) for the displays, both from Liberty Instruments.

Let's start with the woofers.

Something doesn't look right, does it? Those sets of graphs really are for those drivers and all are from an impulse response made on a large baffle. In fact, all four are from the same two impulse measurements. So what's the difference? We'll see that later. Consider the implications.

Now let's go to the tweeters.

Both drivers have their issues, but the only one I treated (for now) was the tweeter. I modified the pole-piece vent treatment with my own (more on that later). The results have several characteristic changes of significance, an altered Fs (a bit higher with higher Q), a smoother response across almost the entire bandpass and improved consistency between the two units, shown below.

Had it been easy to replace the very small chamber, the bump would not have been so prominent. There's too much risk of damage. This may be considered good, in that it's built solidly. We'll deal with the bump in the response due to the Fs increase later. As you'll see, the baffle layout was chosen with this taken into account.

The change in tweeter Fs does have an impact on the crossover, but that's the price to pay (in this case) for a smoother response. As we'll see later, this won't be much of an issue.

Measurements: Woofer T/S Parameters

The Thiel/Small (T/S) parameters are always a bit of a contentious issue. I can't recall any two measured cases from different tests having much consistency. The problem is partly due to the methods (added-mass or delta-compliance) and partly due to test conditions, such as test power levels, software (curve fitting and other variables) and break-in (none, length, etc.). To that end I started by using the supplied parameters because I lean to towards the manufacturer having experience with a much larger sample range than the two that I have on-hand. This may need to be addressed again when the project is complete and I can devote time to focus on this aspect closely to investigate any disparities in modeled vs. measured results.

With that in mind, the first consideration for the design is the box type and tuning, taking the T/S parameters and these measurements into account.

David L. Ralph © 2009