Hello,
I think my second Netra evaluation board failed due to the GPMC port.
Using the first Netra board, I attached scopes to the CS0, WEN, OEN, Data and Address pins. I configured CS0 using the code in norflash.c, supplied by Spectrum Digital as part of the BSP. I was able to see the correct signals on all the lines except for WEN and OEN. I tried tweaking the code but I cannot get those two pins to assert low. I finally ohmed out the signals and discovered that the WEN and OEN pins had only ~100 ohms or so to ground as compared to the other lines which are in the Mohms.
My very very cooperative TI rep, provided me a second Netra board. I proceeded to check out the same pins and all of them works. I developed an interface board with a NOR flash and connectors to J18 and J20 of the Netra board. I used the 3.3V from J18 to power the flash. I was able to query the flash using the CFI capability and read the flash info. So everything appears to be fine.
However, that was short lived. I suddenly was unable to read or write to the flash. I proceeded to check the pins. The CS0, WEN, and OEN pins were fine. A1 was fine. However, D0 and D2 was pulled high and I was unable to drive it low. D5 was OK. So, I ohmed out the pins and sure enough, D0 and D2 have ~100 ohms to ground as compared to the other pins with Mohms to ground.
So, what's going on? I have worked with at least four different CPU evaluation boards in my career with never a failure. Now I think I got two CPUs with failures on the same peripheral.
Does anyone have the same issues with reliability with this GPMC port? It appears that the rest of the CPU is fine. My test code that uses DDR2 and UART2 is still running fine.
Thanks,
Victor