CCS/TMS320VC5409A: Converting from 5409 to 5409A

Hi Mark,

Our team will need some time to gather data. Please allow us a few days and we will get back to you.

BR
Tsvetolin Shulev

Hi,

I've notified the design team. They will post their answer directly here.

Best Regards,
Yordan

Hi Mark,

It can be difficult for us to support these older devices, but we'll try to help you solve your problem.

Did you perform the device swap on one or more than one board? Could the issue have been a reflow problem where not all balls are making a good connection? Or did you populate the C5409A on a new board or remove and replace the processor? If so, do all the surrounding components appear intact?

Have you checked voltages, clocks, reset?
There are changes in the clock modes between C5409 and C5409A. What are the states of the CLKMD1-3 pins?
C5409A: DIVFCT is configured as CLKOUT divided-by-4 mode following reset
Can you observe any clock coming out from CLKOUT? It should most likely be toggling.

According to this thread, the latest version of CCS that supported C54x is CCSv5.5, so dont try any CCS version newer than that:
e2e.ti.com/.../592581

Have you confirmed that the XDS510 emulator works with other boards - is CCS able to connect to other boards with this emulator?

What emulator and target have you set in your target configuration selected in CCS?

Regards,
Mark

Hi Mark,

Your response seems to indicate that what I have done should be working. The reason I am going to the 5409A is that we are supporting a commercial system that runs on the 5409. We’re trying to avoid a redesign.

I am using the PGE100 package so it has pins, not a ball array. The 5409 chip from a working board was pulled and replaced by a 5409A. I had that done by a PCB company. When I couldn’t get the board to connect to the debugger, I had the same thought as you – that the rework may not have been good. I had the 5409 from another working board replaced by a different PCB company and the result was the same. All voltages look fine.

I have been using CCS 3.3 and CCS 5.3.  CCS 3.3 doesn’t distinguish devices beyond being in the 5400 family.  In CCS 5.3, I have selected the 5409A chip. CCS 3.3 is useful in that you can connect the debugger without having the program load. With the CCS 3.3, I am using a XDS510 emulator. With CCS5, I am using a XDS510 plus emulator. The emulator scan test fails frequently with the XDS510 and the Spectrum Digital CCS3.3 version test software. The scan test is always successful with the XDS510plus in CCS5.

CLKOUT is showing approximately a 2.5 MHZ square wave, 3.3 V in amplitude. It is identical to the CLKOUT signal with the 5409 installed. The external crystal oscillator is a 10 MHz device.

I did get the debugger to connect once to the 5409A with CCS3. I had unplugged the JTAG from a working 5409 board and plugged into the 5409A board. Haven’t been able to replicate that though.

I’m thinking of getting some 120 MHz 5409A chips and trying them with the 1.5V CVdd.

Mark

I think I have found the issue – the not RESET on the supervisory voltage chip. In my working system with the 5409, that pin is high. In the modified board, it is low. In changing the CVdd from 1.8 V to 1.6 V, I switched out a TPS3305-18 for a TPS3306-15. The WDI, RESET, and not MR pins are not connected in my design. The not RESET is connected to the DSP. On the first page of the TPS3306 datasheet, the not RESET pin shows a pullup resistor to 3.3 V. My design with the TPS3305 doesn’t have that. There’s no discussion of the pullup resistor in the TPS3306 datasheet. I put in a 10K resistor from 3.3 V to not RESET and now I can get the debugger to connect. I also have read in the datasheet that the PFI pin on the TPS3306 should be grounded if it is not used so I have done that rather than leave as NC.