@DickCappels

If you have a 100Khz PWM signal modulated near 100. %. Would the audio signal get through and drive the speaker? I think it would , but it feels wrong to me somehow.

 

If you mean the duty cycle of the PWM is modulated by an audio signal then, yes, it would. You would hear the audio from the speaker.

 

But if it is a constant near-100% duty cycle the average signal output is D.C. Which will be blocked by the capacitor, but it is not an audio signal. PWM of audio should get through.

 

Edit: Yes, that would work. At 100 kHz the 1 uf capacitor is less than an ohm. The mass of the diaphragm would provide the low-pass filtler.

 

The speaker itself should be a decent low pass filter at least that was my thought.If the signal arrives at the speaker with modulation intact I think the speaker would respond to the power carried by the PWM. I would love to do this experiment I am working on a simple protoboad layout which brought the idea to mind.

 

Make the capacitor much larger and then let the mass of the speaker provide the low pass filtering -seems to work on class D amplifiers.

I think the LC filter you describe is to reduce EMI not to condition the signal for the speaker. The SSM2375 uses spread spectrum to reduce EMI and uses no filter at the speaker output.

 

It is a good idea to have a filter set up to block the PWM "carrier" frequency from getting to the speaker otherwise you run the risk of burning the speaker out. Only the modulation component of the PWM needs to get to the load. A parallel L/C resonant at 100kHz in series with the speaker to block the carrier followed by a 100KHz series one across the speaker to "short" the remaining carrier out will help.

 

It is a good idea to have a filter set up to block the PWM "carrier" frequency from getting to the speaker otherwise you run the risk of burning the speaker out. Only the modulation component of the PWM needs to get to the load. A parallel L/C resonant at 100kHz in series with the speaker to block the carrier followed by a 100KHz series one across the speaker to "short" the remaining carrier out will help.

From the datasheet:
"In ultralow EMI emissions mode, the SSM2375
can pass FCC Class B radiated emission testing with 50 cm,
unshielded speaker cable without any external filtering."

 

At this point I am not too worried about EMI.. When I was a teen my Dad would yell at me to "turn'it off because my various electronics experiment trashed his TV reception. Some things never change.

I believe the speaker will be a high inductance to the AC, and not even be seen by the 100Khz, We are talking low voltages here, not line voltage.

 

From the datasheet:
"In ultralow EMI emissions mode, the SSM2375
can pass FCC Class B radiated emission testing with 50 cm,
unshielded speaker cable without any external filtering."
View attachment 139622

Probably true.
But I was not thinking of radiation of the 100KHz, but of damaging the speaker. Probably only a problem on higher power amps anyway.

 

If I used 12VDCI might get a 12VP-p waveform. RMS is around 8Vwhich works out to 1 1W approx. That is pretty loud. which makes it self limiting.