OMAPL137 audioSample assertion failed, file Mcasp.c

Hi,

  I have done another test on my OMAPL137 EVM.

  I recompile them project  aic3106.pjt which is provided by spectrum inc. CD room with the EVM. the follow is the result:

  It seems that all the tests passed, however i hear nothing from HP. I confirm that my headphone is OK.

  If AIC3106 is broken, would these test pass? I just don't known what's going on with my EVM board.

Would someone help me to figure out the problem? your help will be appreciated.

Did you make any changes to the code or the board before recompiling? Also make sure you have the 'evmomapl137_dsp' gel file loaded before connecting in CCS.

Jeff

Hi,

Usually this assertion occurs when the AIC31 is not supplying the clock. How much time does your sample run before this problem is observed.

Can you check the following

1. Place a break point at this assertion.

2. let the program execute and when this assertion occurs check the XSTAT and RSTAT registers if any errors have occured.

3. Check if the clocks are coming in from the AIC31 codec.

regards,

imtiaz

Hi, Jeff

I have tested the original projects many times, and also evm6747_dsp.gel (I run under dsp) was loaded before connecting.

I'll go on check the other suggestions, thank you all the same.

Bo Li

Imtiaz said:

 You are using a very old version of the PSP drivers. please use the version 01.30.01 from link given below

http://software-dl.ti.com/dsps/dsps_public_sw/psp/BIOSPSP/01_30_01/index_FDS.html

Hi imtiaz,

I check both under psp 1.20 and 1.30.01. In 1.20, there is "Assertion failed, 0!=timeout), file Mcasp.c, line 3683", while in psp 1.30.01 nothing in stdout, the program just running, with no sound output.

Imtiaz SMA said:

Usually this assertion occurs when the AIC31 is not supplying the clock. How much time does your sample run before this problem is observed.

 

The sample run about 1 or 2 seconds, then the assertion occurs. I cannot hear anything.

Imtiaz SMA said:

 Can you check the following

1. Place a break point at this assertion.

2. let the program execute and when this assertion occurs check the XSTAT and RSTAT registers if any errors have occured.

3. Check if the clocks are coming in from the AIC31 codec.

XSTAT is 0x10c at first, then it become 0x008 which means "Transmit clock failure". RSTAT is 0x104 which means "Receive clock failure".

GBLCTL is 0x200 at first, then it becomes to 0x300 and remains this value and this value causes the "timeout" decrease to zero and assertion error occur .

0x300 in GBLCTL means that Transmit and Receive state machine is held in reset.

I meseaured the signal of pin 38(BCLK) and 39(WCLK) of Aic3106. They approximate to square wave with 100mv amplitude and 400ns period. I think this cannot satisfy the serial clock requirement.

I do not known how to fix this. Do i have to replace the aic3106 with a new one? Can I set some parameters and let the McASP generate the serial clock? What parameters should I change?  "Mcasp_HwSetupData mcaspRcvSetup mcaspXmtSetup" ?

Your advice will be highly appreciated!

Bo Li

Did you come to any conclusion regarding why you dont have any audio from AIC3106.

Any usefule insight.

Regards