ROB Splitter Blend + digital potentiometer

[wcraffonara]

Hi,
I've build the ROB Splitter Blend circuit and it works perfectly!
Now I've add a digital log potentiometer CAT5116 to one of the send channels to mute or slowly increasethe volume to this send channel. This works also perfectly but there is a problem when the input signal becomes very high! Even if the wiper of the digital potentiometer is set to RL there is a clipping coming through the wiper output of the potentiometer.
I think this clipping is due to an input voltage at RH greater than 5V ! As you can see from the schematic the op amp is powered by 9V and the bias voltage is set to 5V.
What is the best solution to prevent clipping/distortion from digital potentiometer? Should I set the bias voltage of the op amps to 2.5V? In this case I cannot use the TL072 because this is not rail-to-rail !

Thank you so much!!!
Here is the schematic:

https://www.dropbox.com/s/z8gm1dteq976mby/buffer.png

[PRR]

> due to an input voltage at RH greater than 5V

Maybe; but I am sure it is going below zero. The digi-pot has internal diodes which turn-on when any pin goes below the pot's "GND" pin more than 0.5V. Which means that soft playing may be fine, strong playing will be damaged.

There's no real good total fix.

You *must* bias-up the pot so signal swings don't go outside the supply rails.

A 5V opamp helps; your TL072 isn't happy much below 7V, so we may as well leave it at 9V.

The 3V bias I propose below is a compromise between the 2.5V which centers the pot in its 5V and zero supply rails, and the 4.5V which centers the 9V opamp. The TL072 will swing down to within 1V of the negative rail, so you have 2V peak downswing, enough for most un-boosted guitar. From 3V bias we can swing up 2V before hitting the pot's 5V rail.

[wcraffonara]

Hi PRR!
thank you very much! This looks great!
I found two rail-to-rail op amps from TI that should be better for this application: OPA2141 and OPA1602. What do you think?