• State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 98 subscribers
  • Views 328 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Low level board testing

Rob Beck
Intellectual 970 points

Seeking some advice upon this one ...

We have a board we have made using a C6457 that has a DDR2 fault. I wrote some diagnostic software to test the DDR2 but the software locks on a CSL call that initialises the DDR2 - seems like a read line is not connected from the DDR2 module to the DSP according to a CCS message that comes up.

So, we need an even lower level of board testing, effectively to test the individual connections between the C6457 and the DDR2 module for opens, shorts, crossed etc.

The DDR2 controller in the DSP hides the DSP pins that talk to the DDR2 module and only allow mode and timing set-up. The CSL appears to drive these DDR2 controller registers.

It appears that the only recourse would be to use JTAG boundary scanning.

We could obtain an SDK (if one exists) for our Spectrum Digital XDS510 USB emulator and write some software that drives it. I guess we would need a file from TI that defines the scan chain?

Or, we could buy Spectrum Digital 'DiaTem' Boundary Scan Test Software which I assume would allow us to avoid writing a lot of software but we would need to set up some tests.

Has anybody performed this low level type of testing before and has an advice of best approach?

  • 0
    TI__Guru*** 125430 points

    Does every board exhibit this behavior or is the issue only with a single board, i.e. manufacturing defect?  If the issue is with all the boards, i.e. you think you have hooked up the DDR2 incorrectly, then perhaps posting some schematic snippets is the best thing to do.  If the issue is with only a single board then it will be a bit tougher.  In that case perhaps you need to reflow the board or perform an x-ray.  I think the benefit of boundary scan will be very limited assuming that you have no way of performing boundary scan on the DDR2 at the same time.  Boundary scan is most effective when you can scan both devices that are connected together, i.e. to drive a pin on one device and read it on the other.

  • 0 in reply to Brad Griffis
    Intellectual 970 points

    The DDR2 issue is with one board only. Some others have PHY issues - we cannot get Ethernet - but some boards work correctly and are reliable.

    I already have diagnostics for DSP, DDR2, flash, PHY, switches and LEDs that test if everything is working and some connections - DDR2 and flash address and data lines.

    I've been asked to provide some lower level tests that can identify things like - is the DDR2 read line connected OK to the DSP - because my diagnostics cannot do all of this: The diagnostics can only indicate areas such as DDR2 OK, flash OK, DSP OK etc.

    The only device connected to the scan chain JTAG connector is the C6457 DSP. I thought maybe get access to the DDR2 controller pins and perform dummy read or writes. I understand that the DiaTem Boundary Scan Test Software might actually do this - test memory that is not in the scan chain?

    I'm just looking at this level to cover the case where diagnostics software will not operate and the exact cause needs to be established - i.e. diagnostics will tell us DDR2 does not work because the CSL hangs when initialising DDR2 but we do not know the exact cause - i.e. a read line from the DDR2 might not be connected to the DSP.