This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Audio PWM generator

Other Parts Discussed in Thread: TPA3122D2, PCM1865, PCM1863, PCM1861, PCM9211

Here is what I am trying to do...  Convert an audio signal to a PWM signal,  invert and buffer this signal so it can be sent to a remote location using unshielded twisted pair cabling,  then at the remote location have a power stage to drive a loudspeaker.     I was thinking of using something like the tpa3122d2 to generate the PWM signal, but I'm not so sure it will work correctly for this application... as it will have no real load applied to it.   Does the tpa3122d2 require a load to correctly adjust the PWM output?   Is there something designed to provide a pwm output without the power stage that would be better suited to this purpose?  Any suggestions as to a power buffer I could use to drive the speakers? they will be 4ohm and looking at something in the 10 - 30 watt range

  • Hi James,

    For a PWM signal generation,  a low power class-D amplifier would probably be your best bet.

    However, usually low power means low voltage rails so the signal swing will not be that great. So even with a high power buffer at the speaker end, the maximum power can only be a big as the rail voltage of the input signal for this type of setup.

    Why not use a powerful class-D amplifier right at the speaker and send it a line level analog signal?

    Regards,

    Matt

  • Matthew Beardsworth said:

    Hi James,

    For a PWM signal generation,  a low power class-D amplifier would probably be your best bet.

    Do you have any suggestions on a low power class d amplifier that would be able to generate the correct pwm waveform without a load? 

    Matthew Beardsworth said:

    However, usually low power means low voltage rails so the signal swing will not be that great. So even with a high power buffer at the speaker end, the maximum power can only be a big as the rail voltage of the input signal for this type of setup.

    I was thinking I could perhaps use an optocoupler to change voltages at the speaker end.   since the signal is digital, either on or off, as long as I don't change the pulse width, I should be able to change voltages.. even without the optocoupler.. just a transistor driven into saturation by the signal would allow me to change rail voltages.

    Matthew Beardsworth said:

    Why not use a powerful class-D amplifier right at the speaker and send it a line level analog signal?

    that was my original idea, create a balanced line level audio signal and send that down the UTP cable, then just put each of the inputs into a class D amp in a BTL setup to drive the speaker.

    The distances involved are fairly large (700 - 800 feet)  and I'm thinking an analog line level signal would have some losses and would pick up noise along the way... even though the balance signal should reject a lot of noise.. it's still analog, and that's a pretty long distance to run.

    With a digital pwm signal any noise or signal loss could still be corrected for by just saturating a transistor with whats left of the signal

    James

  • Hi James,

    1. Most class D-amps should run fine unloaded. The output stage is just the output of a FET H-bridge so unloaded or very lightly loaded outputs should cause no problems. As for a specific part number, I would try the audio selection tool HERE for analog input class-D, or HERE for digital input class-D amps.
    2. I would also consider running SPDIF optical from the source to your speakers. Use an ADC on the front end and a SPDIF transceiver to convert to an optical interface. If you are using a stereo setup, you could use 2 ADC channels with 2 separate optical interfaces.

     

    The PCM1861, PCM1863 and PCM1865 EVM's are ADCs that use a PCM9211 transceiver for this purpose. However, they have one optical output so you could not run a separate line for stereo. They do encode the stereo into a single line. This part of the configuration is up to you.

    See the user's guide HERE for shcmatics.

    Matt

  • Matthew Beardsworth said:

    1. Most class D-amps should run fine unloaded. The output stage is just the output of a FET H-bridge so unloaded or very lightly loaded outputs should cause no problems. As for a specific part number, I would try the audio selection tool HERE for analog input class-D, or HERE for digital input class-D amps

    Thats good to know, Thanks.  The thing that had me concerned is while I was going through the selection tool, I noticed most of them stated they are 'closed loop'  I'm not really sure what that indicates, but my concern was that they would be sensing and compensating for the load.  

    I'm sure an optical system would have absolutely no problems with noise, however I think it would be cost prohibitive for this project.  This is part of a distributed audio system I am working on, there are going to be 3 channels, Left, Right, and a mono sub woffer channel.  I'm hoping to run a cat5 or cat 6 cable from the source using one twisted pair for each of the 3 channels.  There are going to be a total of 56 speakers on the system... 25 lefts, 25 rights, and 6 subs.    I'm trying to keep the costs down for what is located at each speaker, because 56 times anything multiplies to big numbers real fast.  I'm hoping to get it down to just a power supply, an H bridge power stage, and output filter circuit located at each speaker. 

    James