Input gain and modulator gain (Excel sheet)

[thomaseliot]

As stated by Jan, exact settings for input gain and modulator gain, are, sonically, the most important improvement.

As I'm not sure to have the best values, I made some calculations on Excel, based on Tripath's Ta2022 datasheet. Here is a link to download:

http://digilander.libero.it/epiworkr/Gain.xls

It may be useful to somebody and, over all, somebody can check the formulas .

Thomas

[halx00]

Just checking my own amp input setting and noticed that D7 INPUT Gain sum D5/D6X-1 is the wrong way round.
D6/D5X-1 is correct.
Rf/Ri as stated in datasheet.
Other than that, good work
thanks

[thomaseliot]

quote:Originally posted by halx00
[br]Rf/Ri as stated in datasheet.

Error corrected. Now the file to download is updated.

Thanks

[Jan]

Thanks for the Excel sheet!

Some comments;

I tend to use slightly higher gain than Tripath, to have more margin in the input stage, so you hear the distortion before you fry the chip input.....

The feedback paths are from the switching outputs, before the output filters. Therefore they are quite noisy buy nature, and lead back to the analogue section of the chip, where you do not want noisy signals. The inputs of the chip are high impedance and thereby sensitive to noise on any long leads, so you want the components close. The solution was to split the resistance into two resistors in series. The larger value resistor should be away from the chip, the smaller value, close to the chip, so do not interchange these positions, even if that would provide the same resistance. If you replace these resistors for other values, I would recomend values between 1K and 3.6K for the lower value and the rest for the high value resistor. The two series resistors also have the advantage of allowing you to pick quite an exact value for the feedback, even if only E12 series resistors are available.

I also posted some tips on measuring the signal levels, under the new subject Audio Measurement Techniques.

[thomaseliot]

quote:Originally posted by jan

The feedback paths .... If you replace these resistors for other values, I would recomend values between 1K and 3.6K for the lower value and the rest for the high value resistor.

Thanks Jan. I updated again the gain.xls with your input. Other corrections?

Of course, anyone can use and modify the file.

Thomas

[FFF]

I expect the TA2022 (amp-1x, -5, -10x) Gain-spreadsheet is still useful. Then: THX.
(this thread is now linked to in TA2022-amp AI's)