Output coupling capacitor size

Hi all,

for my phone project i am now designing the audio circuits.
I have a SIM module with a mono output and amplify it with a PAM8302A amplifier to a 8 Ohm speaker.

Now, there are tons of tutorials/post about chosing the right capacitor in terms of capacitance value.
My questions is: What about the power rating? I want my PCB to be as small as possible, and i'm afraid a too small SMD capacitor will pop on my PCB. After all, the whole power for this 1,5W speaker gets transferred through the capacitors. I haven't found a good answer yet.

Can you help with that?
(For reference, i will be using a 220 Ohm resistor in series with the 8 ohm speaker and a 47uF output coupling capacitor, to bring my cutoff frequency down to around 15Hz. I tried it and it sounds great).

Thanks and have a good day!
-Maik
 
You realise that putting a 220R resistor in series with an 8R speaker will reduce the power delivered to the speaker by a factor of around 1/1000th, right? Also, the PAM8302A datasheet shows a btl output that doesn't require any DC blocking caps. Sorry, I'm not sure I understand what you're trying to do...
 
Sorry, i should have specified it a bit more.
I'm using one speaker for both ringing sound and also when speaking.
So i will have a controlled switch to either put a 220 Ohm resistor in series (when talking) or to directly connect it to the speaker (so 8 Ohm only).
I know the datasheet doesn't specifiy an output capacitor, but wouldn't it still be good to filter out any DC parts?

@gentlevoice: Thank you for the tip! I will look into it..
 
Ok, I understand now. No capacitor needed, any DC offset will appear at both outputs so the speaker won't see any. The only thing I can think of is if one of the two internal power stages failed taking the output to rail or ground. If you want a cap for peace of mind, since we're talking about a phone and, I assume, a tiny speaker, a cut-off of 15 Hz is totally unnecessary, 150 Hz would probably be more than enough.
 
  • Like
Reactions: ma__wu
To give you an idea, the frequency range used in old analogue phone networks was 300 Hz - 3.4 kHz. In mixing modern music (I have some experience there) it's common practice to high-pass vocals anywhere from 100 to 250 Hz, depending on style, male vs. female, rest of the mix, etc. A deep voice can indeed go lower than that, but those frequencies do little for intelligibility, in fact it often improves when you cut them, and the brain has no problem reconstructing any missing fundamentals from the harmonics. Note also that a simple, 1st order high-pass filter like you will use doesn't really "cut" all frequencies below Fc but rolls off gently, so e.g. with an Fc = 150 Hz, 80 Hz will only be down by about 6.5 dB.
 
  • Like
Reactions: ma__wu