Please Explain PWM Amplification

[rueffy]

Hi All,

Ive built and tested Ron Jones's LASER pointer communication system and I must say it works quite well, the audio quality at the output is very good.

However I was hoping someone could explain how the audio amplifier within the receiver actually works? The signal being fed into the amplifier is PWM. I was under the impression the PWM needed to be converted back into audio / demodulated before it could go into the LM386 for amplification.

If anyone could explain why this is not the case that would be good.

Please find a link to the circuit below.

Thanks,
Rueffy.

 

[Sigurthr]

Without pouring over the schematic; the general scheme for audio over a binary/TTL based system is related to that of a Class-D amplifier. An on/off carrier wave is frequency modulated by an incoming analog audio signal creating a PWM modulated signal. This signal consists of the audio input signal + a ton of upper harmonics and the much higher frequency carrier wave. The signal is then amplified (by any type of amplifier, but most often a class-D one in power settings, - not the case here where a class a-b is used) and then fed through a low pass filter that removes the pwm carrier frequency and all the harmonics. What is left is the recovered audio signal, amplified by the gain factor of the amplifier used. In the circuit you posted it relies on the inductance and self capacitance of the external speaker to at the very least dampen, if not fully filter out the carrier signal. If you added a high value capacitor across the speaker leads you might get better audio quality out as this forms a low pass filter.

The whole reason for this is that any device which can amplify is the most efficient and can handle the most power when it is fully switched on. When it is not fully switched on, but only partially on trying to reproduce a varying signal it wastes a ton of energy in the form of heat. So by changing the signal from a constantly varying one to one that is always fully on or fully off allows the device that is going to do the amplification to be the most efficient and heat up as little as possible.

 

[djQUAN]

The signal being fed into the amplifier is PWM. I was under the impression the PWM needed to be converted back into audio / demodulated before it could go into the LM386 for amplification.

Ideally, you'd simply just need a low pass filter to filter out the PWM signal.

But since the design is such low power and the speaker has significant voice coil inductance at high frequencies, The speaker cannot "follow" the high speed PWM signal basically acting as a mechanical low pass filter demodulating the signal from the PWM carrier.

 

[rueffy]

Excellent thanks for the info. Very simple demodulation process, not sure why some of the circuits online are so complicated if its just a matter of integrating a low pass filter. I'm sure there's more to it though.

 

[Bionic-Badger]

Well your speakers will probably do some filtering for you too, so adding a low-pass filter may not matter too much.

 

[rueffy]

Thank you to all those that helped contribute to the development of my circuit. The project is now fully functional and I've made a video as part of my presentation to a class, the video shows the system in operation with a full description of signal flow etc.

http://www.youtube.com/watch?v=CCAkpJ-jG6A

rueffy.

 

[Sigurthr]

Looks good! Reminds me of the one I built a few years back for voice comm. You're right (in the other thread) about aiming difficulty. Best I could get distance wise was about 100meters. A moderate divergence seemed to aid in aiming, as the photodiode doesn't need such high intensity as a well collimated beam.

 

[rueffy]

Looks good! Reminds me of the one I built a few years back for voice comm. You're right (in the other thread) about aiming difficulty. Best I could get distance wise was about 100meters. A moderate divergence seemed to aid in aiming, as the photodiode doesn't need such high intensity as a well collimated beam.

That's a good point, thankfully my diode has a rotating lens for beam dispersion control, not sure how common that is? I've tried about 3 different diodes all up, one of them came separately from the lens, which was a real pain as it involved having to position the lens exactly the correct distance away from the diode, I gave up on that and just stuck with the dot module.

I also tried purchasing a small parabolic reflector (actually a solar cigarette lighter) to increase the target area to just smaller then my hand, but it did a terrible job of returning the light to the correct position, not sure why this was, it seems to work well for incandescent sunlight...must be something behind the science.