Hi,
Can someone confirm that the TRST pin should have an external pull-down resistor and what value it should be?
The datasheet for the OMAP-L138 (SPRS586D) says in section 5.34.1 - JTAG Port Description:
"TRST holds the debug and boundary scan logic in reset (normal DSP operation) when pulled low (its default state). Since TRST has an internal pull-down resistor, this ensures that at power up the device functions in its normal (non-test) operation mode if TRST is not connected. Otherwise, TRST should be driven inactive by the emulator or boundary scan controller. Boundary scan test cannot be performed while the TRST pin is pulled low."
This paragraph from the datasheet seems to imply that it's OK to leave TRST unconnected. And there is no other mention in the datasheet about the electrical characteristics or requirements for the TRST pin.
But, I found the TI wiki page OMAP-L138 Bootloader which says:
"It is important for the device to experience a POR when initially powered up. This will clear the emulation and PLL logic, as well as latch the boot pins correctly. Therefore make sure TRST is externally pulled down on your board. Not doing so can be the cause of many boot related problems.
And, there are several posts on this forum from TI employees regarding TRST:
- nTRST on OMAP-L138 / AM1808 - tlee says TRST uses an internal weak transistor, not an internal pull down.
- OMAP L-137 putting SDRAM in self refresh mode in DSP bootloader - jc-ti says TRST needs an external pull down.
- AM1705 UART2 boot issue - jc-ti says TRST needs to be pulled down.
Lastly, we have the LogicPD board which utilizes a 10K pull down, which seems to be more evidence that the external pull down is required.
So, from all of this, I'm inferring that there should be an external 10K pull down on the TRST line. Can someone confirm? And if so, can a request be made to make this clear in the datasheet?
We missed this in the design of our board and have TRST unconnected. We have been using it with an XDS100 emulator, a J-Link-ARM emulator, and booting from SPI FLASH with no emulator attached. We have experienced some failures to boot, and sometimes the ARM will randomly stop executing code for no apparent reason. We haven't yet been able to identify why the ARM stops sometimes and I'm wondering if it could be the result of the chip not getting a POR (power on reset).
Thanks,
Arthur