Headphone Amp Volume Control

[imloggedin]

I am testing a INA1620 evaluation board and I am curious what the proper way to implement a volume control is while keeping a balanced input signal? The feedback resistors are contained in the chip. Most headphones amp schematics are using a unbalanced input and a inverting/non-inventing architecture. This appears to be a differential amplifier with balanced inputs (or is it?). Thanks for any advice!

Here is a link to the data sheet: https://www.ti.com/lit/ug/sbou205/sbou205.pdf?ts=1624905466258

 

[JohnRoberts]

The "proper" way is probably to add another active input stage to maintain the balance, then use a simple attenuator between the two active stages.

Last century I did one amplifier with a passive shunt resistive pad between two precision resistors feeding the + and - inputs. This is cheap but does not provide very deep kill when turned fully off. When the two inputs are shorted together the level kill depends on the quality of the amps CMRR. Good for simple gain trim, not great for turning input signal completely off.

JR

 

[imloggedin]

Thanks John. Are there any other drawbacks to putting a resistor between + and -? Does this effect the source or the input of the opamps in a negative way?

 

[imloggedin]

Also, I assume you mean something like this? (see attached).

 

[JohnRoberts]

The main trade off is that you don't get deep volume kill, but fine for a level trim. Perhaps use 1% resistors.

JR

 

[imloggedin]

It appears that some pins allow me to put resistors in parallel with the 1k feedback resistors to form a gain control. That being said, why is it so common to form a gain control from the feedback of the opamp instead of attenuating the input signal itself?

 

[JohnRoberts]

I don't know that any one way is most popular but there are consequences to doing it that way.

An amplifier stage will also amplify it's own input noise. Lower gain results in lower output noise, as compared to just padding down the input that keeps the full amplified noise.

JR

 

[Newmarket]

It appears that some pins allow me to put resistors in parallel with the 1k feedback resistors to form a gain control.

Yes - but you'll lose at least some of the performance (CMRR) advantage of having the on chip precision matched resistors (on same die and laser trimming - gives excellent absolute and ratiometric tolerance with temperature). Whether that matters kind of depends on your motivation to keep things balanced / differential.

 

[imloggedin]

Yes - but you'll lose at least some of the performance (CMRR) advantage of having the on chip precision matched resistors (on same die and laser trimming - gives excellent absolute and ratiometric tolerance with temperature). Whether that matters kind of depends on your motivation to keep things balanced / differential.

Will this be any more performance loss than adding a feedback capacitor for bandwidth limiting?

 

[JohnRoberts]

To maintain balance it is good practice use similar value caps in both networks (i.e. 100pF feedback, and 100pF to ground).

JR