That one depends on the design goal. For a subwoofer less carrier supression is needed than for driving a full-range speaker and the stetement above has indeed some truth. But take into account that speaker voice-coils can seldom be regarded as lossless inductors at high switching frequencies.
OTOH it is easier to take feedback from the filter output if it's cutoff frequency isn't too low.
Regards
Charles
NO CHANCE..2-5khZ is alright from shanon theory...
BUT you will hear the 2-5kHz signal REALLY loud!!!!!! And it might also burn your voice coil, as at 5 kHz the speakers impedance typically will still allow a lot of current....
I would propose to shift the switching frequency above what you can hear.
2..5kHz you will always hear, even with a filter.
BUT you will hear the 2-5kHz signal REALLY loud!!!!!! And it might also burn your voice coil, as at 5 kHz the speakers impedance typically will still allow a lot of current....
I would propose to shift the switching frequency above what you can hear.
2..5kHz you will always hear, even with a filter.
Filterless Class D
Filterless class D technique usually refers to circuitry where the PWM carrier cancels at zero net output level, which should also occur when the net input signal is zero. This approach has not been not very practical for DIY circuits because the amount of needed circuitry is approximately doubled, and the amplitude of the output switching components increase in proportion to the audio signal carried by the PWM. The second drawback could result in an AM transmitter if the outgoing and return audio paths to the speaker are not kept very close together, as in a twisted pair. As the switching frequency increases, a shorter non-canceling output loop to a speaker will tend to radiate more. But since the PWM output of a filterless amp would be a square wave, the EMI never really approaches zero, especially if the square waves have quick rise and fall times rich in high frequency harmonics.
amp_man_1, your proposed circuit does not produce sharp-edged square waves, and with the filtering you have suggested, EMI could be quite low. Your low switching frequency of 5khz reduces EMI 20dB lower than if the frequency were 500khz. Hope I did not mistake anything.
Filterless class D technique usually refers to circuitry where the PWM carrier cancels at zero net output level, which should also occur when the net input signal is zero. This approach has not been not very practical for DIY circuits because the amount of needed circuitry is approximately doubled, and the amplitude of the output switching components increase in proportion to the audio signal carried by the PWM. The second drawback could result in an AM transmitter if the outgoing and return audio paths to the speaker are not kept very close together, as in a twisted pair. As the switching frequency increases, a shorter non-canceling output loop to a speaker will tend to radiate more. But since the PWM output of a filterless amp would be a square wave, the EMI never really approaches zero, especially if the square waves have quick rise and fall times rich in high frequency harmonics.
amp_man_1, your proposed circuit does not produce sharp-edged square waves, and with the filtering you have suggested, EMI could be quite low. Your low switching frequency of 5khz reduces EMI 20dB lower than if the frequency were 500khz. Hope I did not mistake anything.
Much thanks for your thoughts, Mr. Pass. That view is very good as even on my self-oscillating design which takes all the feedback after the filter, the resulting variation of switching frequency is from about 30khz at full output amplitude to over 200khz at zero output.
The thought you gave about hearing something audible with low frequency PWM sources fits my experience. My electric bicycle makes a 200hz whine, while the old 12 volt power supply switching at around 7khz is so annoying that I had to place it within concrete building blocks to muffle the squeal.
The thought you gave about hearing something audible with low frequency PWM sources fits my experience. My electric bicycle makes a 200hz whine, while the old 12 volt power supply switching at around 7khz is so annoying that I had to place it within concrete building blocks to muffle the squeal.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.