DM6446, boot problem at low temperature.

Hi,

The operating condition for DM6446 is 0-85 unless you get the industrial temp parts.

Have you considered using those parts?

See it here.

http://focus.ti.com/docs/prod/folders/print/tms320dm6446.html

Hi,

The part number is TMS320DM6446AZWTA, according to the datasheet: operating temperature : -40 to 105 C

The NOR FLASH is PC28F128P30T85, operating temperature : -40 to 85C

Do you have a soldering profile for the DM6446?

Best Regards

Hi, is Figure 16 of SPRAA99 what you are looking for?

Jean-Francois Landry said:

We have a problem to boot DM6446 at low temperature (-20 Celsius).

Normally the boot is made on NOR Flash, but low temperature units rather start on UART0.

Others refuse to boot. If I boot on UART0 it's not possible to access the NOR flash.

At -10 Celsius, all units are working properly. Is someone already had this kind of problem?

Since you mention you are using the extended temperature parts Paul pointed out, this does seem like a bit of a concern, however in general thermal related failures tend to be one of the surrounding peripheral devices failing which causes problems with the processor, as opposed to the processor itself failing. In this case since the device is still functioning it should be easier to try to debug the problem. I see two most likely points of failure for this situation, either the memory itself is locked up or one of the interface lines is getting corrupted leading to general failure to read, or the device is inadvertently getting put into the wrong boot mode. I am suspecting the former since you say you still cannot access NOR after the device has booted via UART0.

    The first place I would look would be the NOR flash you are trying to boot off of and any discrete devices between the DM6446 and the flash that could be causing problems. If you can probe some of the signals on the EMIF to see if there is any reaction from the processor (you should see the CS line assert as well as the OE signal and address bus), and than from the NOR flash (which should assert the data bus), it should help in better determining where the failure lies.

    I would also check the boot pins on the DM6446 to ensure they are still being driven in the boot mode you expect, particularly if they are managed dynamically by a FPGA or CPLD. Additionally it may be worth verifying the various voltage rails, and clocks on the board, typically in thermal failures I have seen it is something foundational like clocks or voltages that goes out of spec at temperature extremes.