Hi Rowdy,

Actually there's no "impedance match" problem for class D audio amplifier at all. The internal mosfet is working in on-off mode. When it's turned on the equivlant resistance is only about 0.1Ohm. The max power of class D amp is depend on the maximum current it can drive. When the current is fixed the bigger the load impandence, the bigger the power. That's why the output power is larger for 16Ohm than 8Ohm. Of course the max power is limited by some other factors such as supply voltage.

Ok, when switched on , the resistance of each MOSFET probably is about .1 Ohm. BUT, impedance is more than just resistance. The current limiting effect of the class D operation IS a type of impedance. ANYTHING that impedes current is impedance and can be measured in Ohms. Also, keep in mind that it takes time to switch from all the way off to all the way on. If you find the average resistance during this transition and multiply that by the switching frequency, you have that much more resistance to add to the .1 Ohm of on resistance. According to my calculations based on P=I^2 * R and what the data sheets shows, you have about 1.18 A of current with either 4 ohm or 8 ohm speakers, but only .968 mA of current with 16 Ohm. If it were as simple as multiplying some constant current times the speaker impedance, you would have 1.18 A ^ 2 * 16 Ohm to get about 22 W with a 16 Ohm speaker. You could also get 118W with a 100 Ohm speaker if this were the case.

Also, Class D operation is simply applying pulse width modulation to the signal and using the pulse width modulated signal to switch a CMOS or H-Bridge (or something similar) output stage. If there is some current sensing feedback to adjust the pulse width to achieve a desired output current, then this is a feature that is an addition to class D operation and not necessarily a part of class d operation any more than current or voltage feedback is a part of class a or class ab operation rather than an additional feature. Perhaps this ic is doing some sort of current sensing feedback. This does not change the fact that it has a significant output impedance. That impedance might not be absolutely fixed, but it remains higher than 8 Ohms.

The maximum output voltage at 16 Ohms would be .968 mA * 16 Ohms, or 15.5 V (Yeah, I know it wouldn't be precisely .968 mA). The output voltage at 8 ohms would be 1.18 A * 8 Ohms, or about 9.4 V. The output voltage for 4 Ohms would be 1.18 A * 4 ohms, or about 4.7 V. Regardless of class of operation or types of feedback or types of impedance, this tells me the output stage is getting loaded down more as you decrease speaker impedance, and this is not unlike what you would have with a Class A or Class AB amplifier with an output impedance significantly greater than 8 Ohm. What's great about Class D operation is that much of its output impedance is reactive rather than resistive, meaning less power wasted as dissipated heat.