• State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 31 subscribers
  • Views 279 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

TAS5508-5124K7EVM disturbance at audio output

Alessandro Crespi
Prodigy 30 points
Hello,
 
We purchased a TAS5508-5124K7EVM evaluation board that outputs an annoying disturbance on the audio output.
 
Our system consists in a pair of tweeters and a woofer connected to J101, J103 and J105 respectively. We filter the signals through the DSP DAP channels 1, 2 and 3. Our filter design is based on 96kHz of sampling frequency.
The two files attached are two configs: REV_A (built based on the original init.cfg) works well concerning the filtering, but the volume does not work; REV_B (built from scratch) fixed the volume issue but it is also altering the final audio frequency response.
 
The disturbance is a sort of harmonic signal, sometimes really distorted (chirp), which depends on the sampling frequency (manually set on the USB eval board): starting from 96kHz the disturbance is higher, then switching to 48kHz simply modulates the noise.
We shot a little video of our setup and the disturbance's behavior, here it is the link: 

https://goo.gl/photos/FT8ETM7o7s2uGJVHA
 
Moreover, the disturbance does not have a predictable behavior. Sometimes it acts like the video above, other times it produces a frequency sweep (firstly we thought to a parasitic capacitance). Other times, on the other hand, the system works fine for just a dozen of seconds.
 
We tried all the analog and optical input stages, without any changes. We tried also different PSUs (both switching and linear) either connecting the board to GND nor without any other connection.
 
Finally we purchased another eval board, thinking that the previous one was bugged. The problem still remains.
 
We would like to go deeper in the details of this issue as soon as possible. Thank you for your consideration.

digital_habits_revA.cfgRev A

digital_habits_revB.cfgRev B
  • 0
    TI__Mastermind 27870 points
    Hi Alessandro,

    Are these same artifacts happening with a default setup of the EVM?

    You stated that the rev A cfg stopped the volume from working, what do you mean by this? That the volume control in the GUI was not working, or were you performing I2C writes to the volume registers and there was no effect?

    An altering in the audio frequency usually means that there is a filter in use that has coefficients calculated for a different sampling rate. This might be due to the bank switching control. When you change sampling rate as you did in the video you need to make sure you load the right coefficients for each sampling rate, this can be done by using the bank switching.

    Are you able to get a capture of all three clocks and the data on a scope with at least one complete frame? This would be to verify if the clocks are synchronous and the data is in the right position.

    Justin
  • 0 in reply to Justin Bohr
    Prodigy 10 points
    Hello Justin,

    First of all let me thank you for your support.

    In the rev A cfg file, the GUI works fine and the I2C led on board blinks as we modify any parameter. Still there was no effect on the audio output volume. Aiming to understand what was going on, we wrote manually a new cfg file, rev B, omitting some register initializations.

    Concerning the audio sampling rate, you're right; different rates need different filter designs. In fact, switching from 96kHz to 48kHz actually does alter the final frequency response; we did this experiment just for verifying if there is any relation between Fs and the disturbance. We could conclude that the noise is not related to our filter design nor to the chosen data format (I2S 24 bits).

    We are investigating for any clocks issue, as you suggest; we will share with you a capture of the clocks and data as soon as possible. Which is the relationship between the synchronization and the disturbance, in your opinion? Should we focus on some special features?

    I hope to hear from you soon,

    Alessandro