ADAU1701 I2S Input
I am having trouble connecting I2S data from an AD1871 to an ADAU1701 evaluation board and am having two seemingly related problems: 1. I cannot get digital audio into the DSP, and 2. When trying to do so, my analog inputs and cut out.
I have a simple schematic just to test the connection: I am sending the same analog audio to both channels of the AD1871 as well as the L analog input of the ADAU1701. The AD1871 is operating as the master over I2S, but when I connect the L/R, B, and Data pins to the headers on the eval board, and set all the corresponding registers for the ADAU1701, I get no digital audio and the analog audio on channel 1 cuts out. When either the L/R or B registers are reset, or the physical wire is detached or grounded, the analog audio restores.
The AD1871 is connected to the ADAU1701 with single wires, all ~6 inches long. I have tied their ground together, and both are operating off the same clock from a function generator.
I have attached my register settings, schematic, clock signals (top to bottom: MCLK, LRCLK, BCLK, DATA), as well as oscilloscope captures of the I2S clock signals at the header pins on the eval board (top) and at the DSP itself (probing the pins on the DSP directly) as I am concerned that they might be too dirty by the time they are reaching the DSP.
Thanks for your help,
Tom
Audio Signal Processors
Integrated/Special Purpose A/D Converters
SigmaDSP Audio Processors
ad1871
Audio A/D Converters
adau1701
Show More
Just an update - I solved the problem getting the audio into the board, as well as with the analog audio cutting out. The data was too dirty by the time it got from the header pins to the DSP for it to even work.
I've shielded all the I2S lines and got them to ~6 inches long, about as short as this setup will allow. Now I can hear the audio on both channels, but it sounds a little distorted and is dominated by static and random noise. I'm not too familiar with what dirty digital audio would sound like. Do you think this is a symptom of the data or perhaps something else.
The audio source is a line level signal from a laptop going through a 10 uF DC blocking cap.
Hi,
Sorry for the late response. For the past few weeks I have been unable to devote enough time to following the forum. Starting next week I should be able to get more active in the forum again and help people out with their questions and issues.
The original scope plots you sent showed some pretty bad signal integrity, but not terrible signal integrity. I would imagine that those signals would still be at least somewhat recoverable on the receiver side. However, it's important to note that the clock signals associated with digital audio are very high frequency, and they do not lend themselves well to running over fly wires.
First of all, the BCLK and MCLK signals are in the MHz range, so shielding, short wires, and local ground connections are necessary when running these signals between boards. Try to take one ground wire per signal wire and run them in a twisted pair between the boards.
Since you probably do not have the appropriate termination, there will be reflections from the receiver that go back onto the wires. One way to help reduce this problem would be to add series resistors on the source side that would help to smooth out the edges of the source signals, resulting in less overshoot and reflection.
Usually, when there are signal integrity issues on digital audio buses, the result is all-or-none operation. That is to say, either the integrity is so bad that no audio gets across, or it meets some minimum threshold where all of the audio gets across.
If you're hearing noise and distortion, that sounds like it would be either bit-level failures or alignment issues. Bit-level failures would generally be caused by large overshoot or reflection, which could potentially cause the bit to be in the wrong state when it is read into the receiver port. Alignment issues would be caused by the register settings for bit polarity and MSB position.
Hopefully this can all be solved by improving the signal integrity between the boards by using twisted pair (signal-ground) for each signal and maybe adding a series resistor (try 20 ohms to 100 ohms) for the signals on the source side.
This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.
Thank you,
EZ Admin
Please log in to reply.