Electrical vs Actual Acoustic Response

One thing we should bear in mind with Filter Theory, though a complex science on its own, is that they are based on electrical responses. In practice, a passive crossover is not connected to a resistor like in the previous simulations. A driver is an electro-mechanical device, hence it stands to reason its behavior is reactive in nature. We will now examine the deviation between theoretical electrical simulations with actual acoustic responses.


In Figure 28, the upper curve is the PL18 without any crossover. The lower graph is a Gated SPL Sweep of the PL18 with the above crossover.

The difference in SPL is due to some additional equalization circuit which we shall ignore for the time being. What is of interest is the rate of rolloff.

In simulation, it's supposed to be 24dB/oct. Acoustically, it was measured at 17dB/oct.

(Figure 28) 4th Order Linkwitz Riley Low Pass


Referring to Figure 29, the top graph is of the PL27 without crossover.

The lower curve is a Gated SPL Sweep of the PL27 with crossover together with a L-Pad and a Notch Filter to neutralize the free air resonant impedance peak.

Acoustic Rolloff is measured at 18dB/oct.

(Figure 29) 4th Order Linkwitz Riley High Pass


We can conclude that our earlier Spice modeling of the networks is not an accurate reflection of the true acoustic response (Figure 30).

Electrically, into a resistive load, it was at 24dB/oct. When the same crossover is connected to a driver, acoustically it rolled off at about 18dB/oct.

We should therefore, be careful not to make the assumption that a crossover modeled to rolloff at 24dB/oct electrically, will rolloff the same acoustically.

(Figure 30) 4th Order Linkwitz Riley Low Pass with High Pass

60 Downes Street | Calais | ME 04619 | USA