Another digitally-conrolled pre-amp

[currentflow]

I'm designing a digitally-controlled pre-amplifier based on the SSM2404P for input selection (non virtual-earth design), followed by a low-gain, non-inverting, high input impedance stage using the AD8599 and then a PGA2310. A two stage (overall non-inverting) baxandall tone control will follow the PGA2310 with the option to bypass. Sorry no schematic available yet! The microcontroller will be an ATMEGA32 which has plenty of program space and I have previous programming experience with this device.

There will be 8 stereo input channels, two uncontrolled buffered line outputs, a VFD/LCD, rotary optical encoder and infrared remote control. When the tone control is in the defeat position, the PGA2310 will directly drive the main amplifier which, in turn, has an input impedance of around 22K.

My first question is the acceptability of a +/- 12V supply for the whole amplifier. Although this is only limited by the SSM2404, I could use a more standard +/-15V for the remaining circuitry until the tone bypass (I also prefer to use the 2404 switch here), so it would seem perhaps pointless in using two supply levels. The intended pre-amp output level would be around 0.95V RMS maxmum. Would it be preferable to use a relay for tone defeat and restrict the 12V supply to just the input selector?

Second question - does anyone have any comments and/or criticisms with the above arrangement please?

Initially, this will drive my 50W+50W GainClone based on the LM4780.

Thank you,
Steve

[myhrrhleine]

Seems like a great plan to me
do keep us informed on your progress.

12 vs. 15v
How many dB headroom loss is that?
not much IMO

[janneman]

Sounds interesting! I agree that 12V vs 15V is moot.
Two comments:
- You probably would want to be able to have more than 0.95VRMS out. It may be enough for your current amp, but to be flexible in the future, a few volts would be better and should not be any problem;
- Why do you use that opamp to drive the PGA? I think you can just drive the PGA from the input selector. Less contamination of the signal path.

Jan Didden

[currentflow]

@Myhrrhleine
Assuming the AD8599 op amps' outputs reach to within 1.6V of the rails, with a 15V supply the maximum permissible peak voltage would be (15-1.6) = 13.4V. With a 12V supply (12-1.6) = 10.4V. (9.47V and 7.35V RMS respectively).

The PGA2310 data sheet quotes an output swing to within 1.5V of the rails.

The change in available headroom is 20.log(7.35V / 9.47V) = -2.2dB. I agree this isn't a dramatic loss!

@Jan
Good point. There appears to be more than sufficient headroom to achieve a 2V(RMS) output if necessary in the future. I would also wish to restrict the pre-amp gain according to the selected audio source in order to reduce the risk of clipping.

The reason I include a buffer between the input selector IC and the PGA2310 is twofold. Firstly, it presents a high-impedance load to the SSM2404 devices, which prefer a load of < 100Kohms in the interest of lower distortion and secondly, presents a low-impedance source for the volume control IC. The PGA2310 datasheet states:

"It is important to drive the PGA2310 with a low source impedance. If a source impedance of greater than 600 ohms is used, the distortion performance of the PGA2310 will begin to degrade".

Steve