Real-Life Measurements Page 2

However, the rising impedance above 5kHz, coupled with a positive phase angle---intended, I assume, to be equivalent to the tweeter voice-coil inductance---is almost never found in a real speaker. I modified the Kantor circuit by adding a Zobel network---a resistor and capacitor in series---across the input terminals (fig.3). The impedance magnitude and phase of the modified simulator are shown in fig.4---the load is now moderately demanding but not untypical. All Stereophile amplifier reviews will now include measurements made with this load. We shall see if significant differences emerge.

Fig.3 Circuit of Ken Kantor loudspeaker simulator, modified to include Zobel impedance compensation in the treble.

Fig.4 Modified Kantor speaker simulator, electrical impedance (solid) and phase (dashed) (2 ohms/vertical div.).

Letters in response:
Why did you add the Zobel network? It has a cut-off of about 10kHz. Did you want to provide an ever-increasing capacitive load to the amp at high frequencies or just limit the spurious responses of all the other Ls and Cs?

If you look at the impedance plot of the Kantor load, which is intended to resemble a small two-way sealed-box speaker, it is much kinder to amplifiers at high frequencies than any speaker I've measured. All you have above 5kHz is the rising impedance above 8 ohms due to the simulation of the tweeter's voice-coil inductance. I added the network purely on an empirical basis to make the impedance magnitude and phase look more like that of the speakers I have measured. It is not quite what is needed, but I only had a small selection of suitable component values in my surplus parts box. It is close enough, however.---JA