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.

TMS320C5517: McBsp Fsx framing.

Part Number: TMS320C5517

Hello everybody,

Our TMS320C5517 application, running @ 100 MHz, uses McBsp for trasmissions.

McBsp is programmed with internal master clock, internally generated Fsx and Clk (25 MHz),
one bit clock Fsx delay.

The following feature is requested:

Transmit through DMA 10 32 bit words at a random intervals: the 10*32 data bits must be contiguous.
Once the 10 words are sent, Fsx MUST NOT be pulsed again, until a new transfer is started.

1)Is such feature possible ?

I notice that using internal Sample Rate Generator (SGR), and writing continuosly words to DXR through CPU,
such words appear correctly in the McBsp Tx frame: however frames are continuosly repeated.

If a 1 s delay is inserted after the completion of a 10 words write cycle, Tx frames repeat themselves at
the same rate than before: the frame is however composed only by the repetition of the last words.
This occurs even in the DSP is halted.
This is compliant with with beaviour of SGR I found on TMS320C5517 Technical Reference.

In other words, is it possible to generate a 10 words frame as long 10 DXR write are completed,
and then preventing additional Fsx frames ?

TMS320C5517 Technical Reference actually grants a framing paced by DXR-XSR copy: in such case
single words are framed, and frames actually stop as soon DXR-XSR copy stop: this option is however
NOT suited to our application, as word data bits are not consective.
Every FSX clocks a spurious data bit: in other words, Fsx interval is 33 CLKX instead of 32.
If I set zero bit delay, word data bits are indeed contiguous, and Fsx interval is 32 CLKX,
but the 1 bit delay is required by the receiver. 

Please note that need to use DMA : I would like to skip the use of DMA IRQ.
but whatever strategy for solving such problem is welcome.

Thank you for your attention.

Misha.

  • Hello Misha,

    I would like you to know that our key person for this issue is currently not available but we’ll reply as quick as we can.

    best regards,
    David Zhou
  • Hello E2E support,
    I could find myself an acceptable solution to the problem posted on Jan 16,
    at the cost however to change the inputs specification.
    N.B. Too bad that C5517 DMA allows sending only 4*(2**N) bytes.

    The true problem however is that many E2E posts are left pending mostly forever.
    Please note that these posts are mostly related to designs on working sites,
    and require a fast reply.
    Moreover recently TI dropped the direct local support of field engineers,
    so only E2E remains.
  • Hi Michele,

    We apologize for not responding to your question sooner.

    What solution did you arrive at? Does the McBSP CLKSTP mode suit your requirements? Or maybe a combination of CLKSTP and TDM modes? Or do you still get unwanted FS pulses between consecutive words?

    Thank you for following up and providing constructive criticism. We acknowledge the problem and aim to improve our quality of support with reduced response times.

    Is there anything else you need help with?

    Regards,
    Mark