Hi!
I'm making a simple active 4-channel summer on TL074 opamps. Power rails are +8V/0/-8V. Inputs are 0dbV, all are buffered. I'm interested in choosing the best sounding opamp operation mode for buffers and summing. Does it matter soundwise if I bias summing opamp to +4V and the signals are mixed positive, not crossing 0? I read that biasing opamps is a class A circuit as with transistors and tubes, but a majority of mixer circuits I found, even high quality, does not use biasing. If it really matters when it is not a problem for me to extend power rails to +-12V for example.
Thank you.

 

TL071/2/4 max power supply is ±18V.
You do not need to bias the inputs.

For best performance, use TL071 since you can better filter the power supply rails and avoid crosstalk from adjacent channels.

 

TL071/2/4 max power supply is ±18V.
You do not need to bias the inputs.

For best performance, use TL071 since you can better filter the power supply rails and avoid crosstalk from adjacent channels.

Thanks for your reply!
I have a +-15V supply transformer and use 3 regulators connected in series to reduce PSRR: 15V->7x12->7x10->7x08->8V. It gives enough headroom for mixing 4 0dbv signals. Are there any other advantages in using full swing power rails except headroom?

 

Although your signal might be a nominal 0.775Vrms, the nature of audio signals is such that even at this voltage, peak voltages can easily be of an order of magnitude greater.

Therefore to avoid clipping within the pre-amp section I’d recommend supply rails at least +-15V and that you minimise any biasing offset, which will otherwise reduce the headroom to one of the rails.

 

Although your signal might be a nominal 0.775Vrms, the nature of audio signals is such that even at this voltage, peak voltages can easily be of an order of magnitude greater.

Therefore to avoid clipping within the pre-amp section I’d recommend supply rails at least +-15V and that you minimise any biasing offset, which will otherwise reduce the headroom to one of the rails.

Thanks and yes, that makes sense if the inputs are fed with some random voltages around 0dbV, but I forgot to mention that all 4 inputs are fed from DACs with 2Vpp max outputs. So even 4 sines in sync will give 8Vpp max, no spikes, right?

 

If your DACs are output an audio signal at 0.775Vrms with a 2V peak to peak maximum – you will have clipping of the signal out of the DACs (unless there is some soft clipping within the conversion response).

But on the flip side, the summed output from the op-amps will be limited to 8V (as you say) and that only when all inputs are at max, so you could probably get away with +-8V rails.

 

Thanks for your help

 

The TL07x opamp has a very low distortion of 0.003% and you want to reduce it by operating its output in class-A? A lousy old LM324 and LM358 has the output transistors with no bias current producing severe distortion and their distortion is reduced when an output transistor operates in class-A.

The TL07x and TL08x opamps and others have a problem called Opamp Phase Inversion that causes an output to suddenly go as high as it can when an input gets within 4 volts (worst case) from the negative supply. It causes the worst distortion you ever heard. Look here where it messes up a sinewave: