The Final Outcome

Voicing a speaker is much more complicated than what was described in the previous chapter. For presentation purposes, the tweeter and woofer were separated, but in reality when one is voicing a speaker, both woofer and tweeter are connected. This is necessary as we need to listen to the full spectrum of frequencies in order to make any sense of what we are hearing. It's simply impossible to have music through the tweeters alone and try to identify where the problem areas are.
So it is more like one minute tweeter and the next woofer, and it ping pongs this way to and fro. And after days of tuning, when you think you have finally gotten it, a single adjustment throws everything off and it's back to square one.


Figure 36 shows the individual acoustic rolloff of the woofer and tweeter with their networks. But this does not show us whether there are any problems with the crossover.

To see whether our 4th Order Filters are crossing correctly, a 300 Points Precision Gated SPL Sweep was done with a 2.3 mSec Meter On Time with both networks connected. Microphone was at 1 Meter, on axis with the tweeter with 2.83V input signal.


(Figure 36) Final High Pass & Low Pass Network


Figure 37 is the result of the sweep. Note that 6dB above the crossover point (1,600Hz), the woofer and tweeter sum at unity, that is there is no peaking and no depression.

In Figure 38, the graph was smoothed by 0.1dB to remove the gating artifacts. For all intents and purposes, this would be an accurate representation of the speaker's frequency response.




With the response now finalized, an impedance sweep is done to to determine the load presented to the amplifier.

In Figure 39, we can see that the port is well centered at 40Hz.

Nothing out of the ordinary along the way till the crossover at 1,600Hz. At this point, the transition from woofer to tweeter caused a bump after which it nose dived to a low of 4 ohms at about 2,700Hz before leveling off at 5 ohms. At 8kHz onwards, the tweeter impedance started climbing due to the inductive nature of the voice coil.

These speakers will not present any problems to power amplifiers. Most modern day amplifiers are designed to operate down to 4 ohms after all. On hindsight, a series resistor could have been used instead of a L-Pad. This would raise the mean impedance of the tweeter to be more in line with the woofer but being a purist at heart, I have never been comfortable with any series resistors in my networks. I see them as current limiting devices.

(Figure 37) Gated SPL Sweep of Speaker

(Figure 38) Smoothed 0.1dB

(Figure 39) Speaker's Impedance


For completeness, the Transfer Functions of the networks were generated (Figure 40). This shows the actual electrical responses when the networks are connected to reactive loads as opposed to resistors.

Note that the woofer rolloff rate is much steeper than 24dB/oct. In fact, it was measured at about 48dB/oct. Tweeter's rolloff is about 16dB/oct.

(Figure 40) PL18 & PL27 Network Transfer Functions


60 Downes Street | Calais | ME 04619 | USA