TMS320VC5502 does not always boot from I2C EEPROM

Thierry Cachat

Other Parts Discussed in Thread: TMS320VC5502

Dear all,

We have designed a board with a TMS320VC5502 configured in I2C EPROM boot with an EEPROM 24LC512 from Microchip. Most of the time the DSP boot and run without problem, but some time it does not boot. When it fail, there is no activity at all on the SCL pin of the I2C. We have checked the following at power up :

the level of the BOOTM and other configuration signals (GPIO 0, 1, 2, 4, 6, 7 each has an external pullup or pulldown)
DVDD, CVDD, PVDD,
the Reset_b is low for many milliseconds using a TPS3808G33MDBVREP,
the crystal between X1 and X2 is oscillating.

Everything seems correct, except that the DSP does not start to read the content of the memory.

Is it possible that the interrupt pins INT0_b to INT3_b affect the startup code? We have left NMI_b floating.

Thank you

over 12 years ago

0 Amrutheshwara sannadyavanannavara over 12 years ago

Prodigy 10 points

we all faced this problem but one thing is mainly about handling board and use PCB(silver or gold plate). if your presently using silver. please go to gold finish.i hope it will help you.

Thanks

0 Steve Tsang over 12 years ago in reply to Amrutheshwara sannadyavanannavara

TI__Mastermind 23785 points

It is always a good practice not to float any input pins. NMI is enabled by default and is non-maskable. Thus it could cause an interrupt. I would also recommend to scope these signals to be sure.

Regards.

0 Thierry Cachat over 12 years ago in reply to Steve Tsang

Intellectual 280 points

According to the datasheet, there is an internal pullup on NMI... We have also tried to add an external pullup, but it did not solve our problem. When we reboot the DSP by software, the boot is always correct.

A few days ago we had (erroneously) an external pulldown on NMI, and the problem appeared quite often.

0 Steve Tsang over 12 years ago in reply to Thierry Cachat

TI__Mastermind 23785 points

Have you scope the signals? With you new data, I have these thinking, 1) with the incorrect pulldown, it happened often, now with the correct pullup, it still happens but less often; 2) a software reboot always work. This tells me something is incorrect signals right after reset, do you meet the required stabilization periods for reset signal to propagate through the DSP?

Hope this help.

Regards.

0 Thierry Cachat over 12 years ago in reply to Steve Tsang

Intellectual 280 points

We have scope many signals and we did not find anything suspicious. Attached is a screen shot of the scope.

Green is the crystal: it start oscillating early.
Red is the reset signal.
Yellow is the 3,3 V power supply, and
blue is the SCL (I2C). Here it fails. When the DSP correctly starts, we see activity on the SCK right after the reset goes high.

(We have also checked the 1,2 V power supply)

I should also mention the following: after 5 minutes without power, the board always starts correctly.

When we try to restart after only a few seconds power down, it can fail. And when it fails to start, if we shutdown power again for only a few seconds, there is high probability that it fails again.

0 Bill Finger over 12 years ago in reply to Thierry Cachat

Genius 3735 points

Have you looked at the I2C data line? I've had issues where one of the slave devices on the bus got into a state where it was holding the data line low, and only a complete power down could restore the system to proper operation.

0 Thierry Cachat over 12 years ago in reply to Bill Finger

Intellectual 280 points

We have checked the data line, it is correct. There is only the EEPROM (and the DSP) on the I2C bus.

0 Steve Tsang over 12 years ago in reply to Thierry Cachat

TI__Mastermind 23785 points

If it always work after a ~5 mins power down, I would suspect some components is still powered by large capacitor(s). The time is for the cap to discharge. You might want to check that out.

Regards.

Processors

Processors forum

TMS320VC5502 does not always boot from I2C EEPROM