TLV320AIC3100:

Hi Suzuki-san,

Power dissipation is usually defined as our max spec that correlates to how much power a device can dissipate continuously without causing damage. Where are you getting greater than 750mW from? 

As for the calculations, you're close! The power dissipation will be a combination of the power lost through the supply and power through the switch itself. P = IV or P = I^2R as you mentioned above. So if we know the supply voltage and the current through the switch we can determine this value. The equation is as follows : 

So the supply voltage and supply current are your Pd for the supplies, and the current and RON for the switch part. But note that the amount of switches that are active will also make a difference. Since you're looking for a 2channel devices, we'll call that as 2 switches on.

I've pulled some of this from this E2E FAQ
https://e2e.ti.com/support/switches-multiplexers-group/switches-multiplexers/f/switches-multiplexers-forum/1063078/faq-how-to-calculate-the-max-junction-temperature-of-an-analog-switch-multiplexer

There is an excel at the bottom that allows you to input your specifications and it will automatically solve for the max power dissipation. These devices are simple passive switches though, so as long as you adhere to the specs of the datasheet, you should be fine. 

I see Nir recommended the TS3A24157 in the previous post? Pd was 180mW and has an RON <1ohm. Was there an issue with this suggestion? The device can support up to 3.6V and 300mA current through the signal path which may be pushing the limits here. However the peak is a bit higher; 500mA with 1ms <10% duty cycle pulses. 

I think we should re-evaluate what the actual current and voltage requirements are.

As you mentioned earlier P=I^2*R, so I = sqrt(P/R) = sqrt (1W/8ohm) = 353mA (So maybe TS3A24157 won't work if the signals are longer than 1ms <10% duty cycle). 

For voltage, we can start with the same P = I^2*R and adjust for voltage which gives us P = V^2/R. From the previous E2E post, I see that your speakers are 8ohms. So V = sqrt(P * V). So V = sqrt(1W * 8ohm) = 2.83V. 

So I think we would need a device that can support 353mA and 2.83V, and we'll keep RON <1ohm.
TS5A26542 may be a good choice. It can support up to 450mA continuous. 

If you were to use the device under the conditions listed above, at max temperature both switches on and we were to have the max supply current and voltages, the math indicates here that it would only dissipate worst case 200mW. 

Now for the mux specifically in this use case though, the RON will be 0.75ohm. P = I^2*R = 0.353^2 * 0.75 = 93mW. The package thermal impedance is 102C per Watt. So 102C * 0.93mW = 9.5C. So the device will heat up 9.5 degrees C under the current conditions. That should be fine I would think.

Thanks,
Rami