• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 30 subscribers
  • Views 437 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AIC34 AGC target level seems to be sampling rate dependent

Roland Alber
Intellectual 305 points


Hi,

we have a strange phenomenon with the AGC of the AIC24 Codec: The target level seems to be dependent on the sampling rate. Our measurements with 1kHz input signal: 48 kHz sampling rate result in ~71% level, 44.1 kHz in ~68% and 22kHz in ~33%.

We cannot find appropriate infos or formulas in the data sheet, so can anybody please explain us the dependency of the AGC target level on the sampling rate and how to calculate the target level correctly?

Thanks and best regards

Roland

  • 0
    TI__Guru**** 189401 points

    Hi, Roland,

    Can you post the complete part number of the device in question?

    -d2

  • 0 in reply to Don Dapkus
    Intellectual 305 points

    Hi Don,

    the precise part number is TLV320AIC34IZAS.

    Addendum to my request: I used "target level" for the level the codec AGC produces, so it is not the target level set, but the level which is achieved after the AGC, i.e. at the ADC input. We also noted that a test sine wave of 1 kHz @ 48KHz sample rate has a little amplitude modulation at a period of 2-3 sine waves, i.e. 2-3 ms. The modulation seems to be roughly 0.5 dB which is also the AGC / PGA step size, i.e. the AGC seems to toggle around the last bit. What is not clear is the AGC control loop frequency, i.e. the PGA gain update frequency. Can you give us a formula for this also? Thanks.

  • 0 in reply to Roland Alber
    Intellectual 305 points

    Hi Dan,

    here are more details to our question regarding TLV320AIC34IZAS AGC "output" level:

    AGC-Params:

    AGC Attack Time 8ms
    AGC Decay Time 100ms
    AGC Target Level -10dB
    AGC Max PGA Gain 40dB

    The achieved level either measured at the output of the codec (HPLOUT A) or measured by looking at the values of the samples produced by the ADC are according to our measurements (see table below) strongly dependent on the test tone input sinewave frequency and on the ADC sampling rate. Also we were not able to achive -10 dB FS using a sine wave test tone. Although when lowering the AGC target level value, the AGC output level seems to be lower by the amount the theoretical calculation would indicate, i.e. the relative AGC output level behaviour seems to be ok and the AGC works.

      Measured Output Level mVpp at HPLOUT A
    Test-Frequency at LINE1LP/M_A Fs = 22kHz Fs = 44kHz Fs = 48 kHz
    500 260 365 360
    1000 205 269 270
    2000 184 207 208
    3000 179 192 193
    4000 175 178 179
    5000 162 172 174
    6000 158 167 165
    7000 152 160 162
    8000 147 154 148
    9000 144 144 144

    Can you please explain why the level produced by the AGC is sampling rate and test tone frequency dependent and give us a formula on how to calculate the AGC output level and how to interpret the set FS target level which we do not achive using the AGC (i.e. why can't we get the e.g. -10 dB FS as decribed above)?

    Also some notes on the AGC measurement value integration/averaging time (is the the attack time?) and the used algorithm could be very helpful for us to correctly program the AGC for our application which might use different sampling rates.

    Thanks.

    BR Roland