TMS320C6657: Power-on sequence leakage on C6657

Omin Chou

Part Number: TMS320C6657

Hi all,

I had a custom board using C6657.

We found 650mV leakage from +V1.0_S0_DSP (CVDD1) to +V1.5_S0_DDR3 (DVDD15) supplies during power-up sequencing. Then also make sure a maximum of 100ms between one power rail being valid, and the next power rail in the sequence starting to ramp. Does this negatively impact device operation or reliability?

Thanks in advance.

B.R.

over 8 years ago

0 Yordan Kovachev over 8 years ago

TI__Guru**** 161600 points

Hi,

Not sure what will be the impact on your device operation in long term, but it is not normal and you should revise your schematic. Are you certain you comply with the power up sequence?

Best Regards,
Yordan

0 Omin Chou over 8 years ago in reply to Yordan Kovachev

Genius 3555 points

Hi Yordan,

Thanks for your reply.
Would you please share power-on sequence waveform from TI EVM? I want to have a comparison between them.

B.R.
OC

0 Yordan Kovachev over 8 years ago in reply to Omin Chou

TI__Guru**** 161600 points

Hi,

I don't have it and I don't have the means to measure it at the moment. I will ask the hw team to elaborate here.

Best Regards,
Yordan

0 Tom Johnson 16214 over 8 years ago in reply to Yordan Kovachev

TI__Mastermind 46460 points

Omin,

You need to follow the power supply sequence shown in the TMS320C6657 Data Manual. The Errata document shows that there is known leakage between the VDDR and VDDT rails. There is not other leakage expected. Please review your board design to detect other paths for leakage.

Tom

0 Tom Johnson 16214 over 8 years ago in reply to Tom Johnson 16214

TI__Mastermind 46460 points

Omin,

I looked into this further. The VDDRx pins are tied to 1.5V which is the DDR supply. The VDDTx pins at 1.0V and tied to CVDDT. Therefore, the leakage that you are seeing is the leakage reported in the Errata Usage Note 10.

Tom

0 Omin Chou over 8 years ago in reply to Tom Johnson 16214

Genius 3555 points

Hi Tom,

Thanks so much for your mention.
Seems this leakage is expected, so the top of power-on sequence should be the normal waveform on C6657, right?

B.R.
OC

Processors

Processors forum

TMS320C6657: Power-on sequence leakage on C6657