DM642 Reads without asserting CEn

I have run into an issue on one of my boards (custom) where the DM642 will assert multiple reads without dropping any of the CEn signals.

This happens when the DSP supposedly reads from CE3. I have tried both a mem-copy and EDMA with the same results on this particular board. Other boards work fine, but this one does not seem to drop the CEn signal. It does on test code, but not my "full up" code. The CE3 chip select works, as I have probed it with an O-scope.

I was wondering if any of you have seen behavior like this, and what could be the cause of it?

Thanks in advance,
-Jim
Check Out Industry's First Single-Chip, Multi-Format, Real-Time HD Video Transcoding Solution for Commercial & Consumer End Equipment: www.ti.com/dm6467 Start a New Thread

Jim,
>
>On Jan 14, 2008 8:16 PM, wrote:
>>
>> I have run into an issue on one of my boards (custom) where the DM642 will
>> assert multiple reads without dropping any of the CEn signals.
>>
>> This happens when the DSP supposedly reads from CE3. I have tried both a
>> mem-copy and EDMA with the same results on this particular board. Other
>> boards work fine, but this one does not seem to drop the CEn signal. It does
>> on test code, but not my "full up" code. The CE3 chip select works, as I
>> have probed it with an O-scope.
>>
>> I was wondering if any of you have seen behavior like this, and what could
>> be the cause of it?
>
>I have seen similar symptoms with external memory problems. Simple
>test cases [run from internal memory] would work fine, but larger
>programs [run from external memory] behaved badly when accessing CE2.
>
>I suggest running simple programs from internal memory to attempt to
>characterize the behavior. Keep in mind that the symptom that you
>are seeing [CEn not working] may not be directly related to the 'real
>problem'.
>
>Let us know what you find.
>
>mikedunn
>>
>> Thanks in advance,
>> -Jim
>>
>
Mike,

First and foremost, thank you for responding to me. Indeed, my test code works just fine. However, in our case the DSP does not run out of external memory.

Some of my findings:
1) Turns out that the "spurious" reads with no CEn assertion starts whenever I have a data transfer (interrupt driven EDMA) from an external SRAM.
2) Getting rid of the EDMA and moving to a datcopy fixed the issue.

The datcopy is just a temporary solution. Any insights?

-Jim
Check Out Industry's First Single-Chip, Multi-Format, Real-Time HD Video Transcoding Solution for Commercial & Consumer End Equipment: www.ti.com/dm6467 Start a New Thread