Jtag chain for IBM PPC 405EP and TI DSP TMS320C6711C

Hello,

I am designing a board with IBM PPC 405EP and TI DSP TMS320C6711C. I want put those devices in a single JTAG chain so that i can carry out boundary scan as well do the emulation by having jumpers appropriately.

But i have one problem with defining a initial state for nTRST signal. In DSP it is pulled low internally and datasheet specifies not to put a pull-up on this line. But PPC datasheet says to put a pull-up on this line. Can anybody help me sort out this issue.

I had copied the references made in both datasheets below. Pls help...

TI Dsp "The TMS320C6711/11B/11C/11D DSP requires that both nTRST and nRESET resets be asserted upon power up to be properly initialized. While nRESET initializes the DSP core, nTRST initializes the DSP's emulation logic.Both resets are required for proper operation. While both nTRST and nRESET need to be asserted upon power up, only nRESET needs to be released for the DSP to boot properly. TRST may be asserted indefinitely for normal operation, keeping the JTAG port interface and DSP's emulation logic in the reset state. nTRST only needs to be released when it is necessary to use a JTAG controller to debug the DSP or exercise the DSP's boundary scan functionality. For maximum reliability, the TMS320C6711/11B/11C/11D DSP includes an internal pulldown (IPD) on the nTRST pin to ensure that TRST will always be asserted upon power up and the DSP's internal emulation logic will always be properly initialized. JTAG controllers from Texas Instruments actively drive nTRST high. However, some third-party JTAG controllers may not drive nTRST high but expect the use of an external pullup resistor on nTRST. When using this type of JTAG controller, assert nTRST to initialize the DSP after powerup and externally drive nTRST high before attempting any emulation or boundary scan operations. Following the release of nRESET, the low-to-high transition of TRST must be "seen" to latch the state of EMU1 and EMU0. The EMU[1:0] pins configure the device for either Boundary Scan mode or Emulation mode."

IBM PPC 405EP " nTRST on the PPC405 based processors shouldbe pulled up with a 10K resistor. Also note that nTRST must be asserted low in response to a power-on or system reset or else the processor may not boot reliably. It is recommended that nTRST from the JTAG connector be logically ORed with power-on reset of the board before being connected to nTRST on the processor"

Pls suggest some solutions....

Reply to
gowda
Loading thread data ...

Hello,

I am designing a board with IBM PPC 405EP and TI DSP TMS320C6711C. I want put those devices in a single JTAG chain so that i can carry out boundary scan as well do the emulation by having jumpers appropriately.

But i have one problem with defining a initial state for nTRST signal. In DSP it is pulled low internally and datasheet specifies not to put a pull-up on this line. But PPC datasheet says to put a pull-up on this line. Can anybody help me sort out this issue.

I had copied the references made in both datasheets below. Pls help...

TI Dsp "The TMS320C6711/11B/11C/11D DSP requires that both nTRST and nRESET resets be asserted upon power up to be properly initialized. While nRESET initializes the DSP core, nTRST initializes the DSP's emulation logic.Both resets are required for proper operation. While both nTRST and nRESET need to be asserted upon power up, only nRESET needs to be released for the DSP to boot properly. TRST may be asserted indefinitely for normal operation, keeping the JTAG port interface and DSP's emulation logic in the reset state. nTRST only needs to be released when it is necessary to use a JTAG controller to debug the DSP or exercise the DSP's boundary scan functionality. For maximum reliability, the TMS320C6711/11B/11C/11D DSP includes an internal pulldown (IPD) on the nTRST pin to ensure that TRST will always be asserted upon power up and the DSP's internal emulation logic will always be properly initialized. JTAG controllers from Texas Instruments actively drive nTRST high. However, some third-party JTAG controllers may not drive nTRST high but expect the use of an external pullup resistor on nTRST. When using this type of JTAG controller, assert nTRST to initialize the DSP after powerup and externally drive nTRST high before attempting any emulation or boundary scan operations. Following the release of nRESET, the low-to-high transition of TRST must be "seen" to latch the state of EMU1 and EMU0. The EMU[1:0] pins configure the device for either Boundary Scan mode or Emulation mode."

IBM PPC 405EP " nTRST on the PPC405 based processors shouldbe pulled up with a 10K resistor. Also note that nTRST must be asserted low in response to a power-on or system reset or else the processor may not boot reliably. It is recommended that nTRST from the JTAG connector be logically ORed with power-on reset of the board before being connected to nTRST on the processor"

Pls suggest some solutions....

Reply to
gowda

Jeeze, could they make it any more complicated??? TI was one of the original founders and promoters of JTAG boundary scan and JTAG debugging. But for some reason, they do a lousy job of making their development tools easy to use with a boundary scan capable board. Worse, none of the vendors seem to care if their parts can be debugged with the part in a boundry scan chain.

I have to suggest that you keep the two JTAG ports separate. I looked into this and that was the suggestion from everyone I spoke to in the industry; chip vendors, debugger vendors and JTAG test tool vendors, especially the JTAG test tool vendors. I expect they would know best. In the factory test, you can always use multiple test ports simultaneously, or so I was told. I eventually gave up on the idea since my board has multiple power zones and it is too hard to bring it up so all the parts can be tested together. I am going to use on board software to test the board at the factory.

But it does sound to me like you can tie the two nTRST pins together and let the TMS part generate the power on reset for the PPC. I don't see where it says if the PPC nTRST can be held low after power on reset. They just say it has to be reset and you should use a pullup. Actually, it sounds to me like the PPC JTAG is not designed correctly. What if you are running a debugger that pulls nTRST high actively and you do a system reset? Will the PPC hang if the nTRST is not pulled low?

--
Rick "rickman" Collins

rick.collins@XYarius.com
 Click to see the full signature
Reply to
rickman

Sorry about the spelling and grammatical errors... ;-)

I guess this doesn't help too much, but I'd like to share some experiences that I have had with TI TMS320C6000 platform and it's BScan implementation. I've been working with a design that has several TMS320C6416 devices, PPC440 and some other BScan devices in a sigle chain for BScan testing. For emulation purposes TI DSPs are separeted from the chain (two DSPs/emulation chain). I won't go into more details about the design, but some notes on the BScan functionality with TI320C6416 in heterogenous environment, namly:

In order to get the darn DPSs into BScan mode you need to have EMU[1:0] pins correctly set and the pins are not latched only by the raising edge of _TRST, but prior to that at least one TCK pulse has to be applied when _TRST is active i.e. low.

TMS-reset (max. 5 TCK pulses with TMS held high) of the test logic (TAP-controller into test logic reset state) in the TI TMS320C6000 doesn't obviosly work unless the chip is first got into Boundary-Scan mode.

_TRST should be pulled-up as it actually should be done also internally on any BScan component. The dot1 standard doesn't mandate this, but in order to access the test logic even when external access to the _TRST is missing due to a fault, there should be an internal pull-up instead pull-down that would disable the usage of the test logic. One could easily think, that pull-down would make the chip functionally usable even if the external access was missing (which is of course true). However, as there are means to test if the access is there or not it is strognly advisable to assure that the test logic is usable also when the _TRST pin would have e.g. an open fault.

Getting the TMS320C6000 chips into BScan mode with BScan tester might require small additional logic. For instance Göpel tool doesn't have active TCK when _TRST is asserted. This means that EMU[1:0] pins doesn't become latched correctly. However, if I recall correctly the tool from JTAG Technologies has the TCK running also when _TRST is active, so most likely the TI DSPs would get into BScan mode w/o problems.

It really is annoying that a company, which joined the JETAG (-> JTAG after yanks came along) working group back in 1986 to come up with the nice IEEE1149.1 standard, is now doing things a lot more difficult than they should be...

Reply to
Mika

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.