I need help to better understand differential amplifiers, when interfaced with an decoder IC. For the circuit below, which is from the reference design here:

enter image description here

It shows a mono differential amplifier (TPA751), driving a speaker on the same PCB. Not shown is the MP3 decoder IC (VS1003) driving the amplifier input. The amplifier has gain of -2Rf/Ri=2*47k/10k= 9.4. So the differential output is 9.4x the differential input.

As per the VS1003 datasheet, the output voltage from the CODEC (and input to the amplifier) is 1.7 Vpp (peak to peak) with a 1.2V DC offset. The outputs from CODEC and into the amplifier IN- and IN+ terminals are opposite of each other (ie. IN+=--IN-). I’m trying to understand how this circuit would work and would appreciate clarifications on the following points:

  1. 1.7 V peak to peak for the CODEC output means an amplitude of 0.85V. Therefore, with the 9.4 gain of the amplifier, the output to the speaker will be 0.85*9.4 = 8.0V peak to peak (OUT+ relative to OUT-). Is this correct?

  2. If we measure OUT+ (pin 5 of the amplifier) with respect to ground and not to OUT- (pin 6 of the amplifier), the maximum amplitude will be 4V (ie. 4.7*0.8). So the signal will be from 0V to 4V. Is this correct?

  3. Assuming the amplifier can produce a maximum voltage swing of VDD-1V, this means to avoid clipping, the power supply of the amplifier needs to be at least 5V. Is this correct?

  4. My understanding is that speakers can’t handle DC offset. They must be driven by an AC signal. For our case above, based on the datasheet, there is no DC offset at the output. However, the output signal can be though of an AC signal with amplitude +/- 2V and a DC offset of 2V. Is this considered bad for the speaker? Or is this not a real DC offset?

Thanks! (This is my first ever post in this forum)

  • \$\begingroup\$ Correct, most speakers do not like a DC offset. Note that the speaker is not referenced to ground. This amplifier is designed so that each output will have the same DC offset, and therefore will not produce a significant differential DC voltage to the speaker (assuming that the input is AC coupled). Your drawing above omits the input DC blocking caps which are very important. \$\endgroup\$
    – Mattman944
    Commented May 8, 2019 at 9:00

1 Answer 1


1) No. The amplifier has 9.4 gain so 1.7 Vpp at input is 15.98 Vpp at output. For some reason you halved the input Vpp voltage to peak amplitude. Obviously the decoder is not set to full volume.

2) Yes, each output pin has maximum output of 4 Vpp when used with 5V supply. Input voltage and gain must be chosen to be in useful range so it does not clip.

3) Yes, sure it needs 5V supply to provide 4Vpp single ended outputs and this is 8Vpp differential to the speaker.

4) As there is 2VDC bias on both outputs, and speaker is connected between both outputs, there is no DC bias over the speaker. 2V minus 2V is 0V.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.