ADS5281 behavior on clock loss

Jiri Kral

Other Parts Discussed in Thread: ADS5281

Hello,

I would like to ask what is the expected behavior of the ADS5281 after the driving clock has been suspended for substantial time (miliseconds - seconds)?

I have order of 400 ADS5281, ~10 mounted to a single PCB always. 395 out of them recover well after loss of driving clock (for whatever period). 5 ADCs, randomly distributed through the boards and positions, do not recover after the clock loss. "Do not recover" means that the returning word clock is dead and there is no data output. Slow control is functional.

Normally I do reset the ADS5281 after clock losses. For these 5 "misbehaving" ADCs I do have to fully power down/up with the PD pin or slow control, to get them back to operation. As I have figured out the way to get them online, there is not an urgent problem. What bothers me is why would these 5 decide to behave different than the others (adjacent on the PCB, sharing voltage, clock, resets, ...). One of the options is preemptive replacement (but before I do that, I would like to hear your opinion .. maybe someone encountered the same).

Thanks for suggestions

- Jiri

over 10 years ago

0 Richard Prentice over 10 years ago

TI__Mastermind 29220 points

Hi,

i have not seen this before nor heard of it, but i will check with the design team about it.

Regards,

Richard P.

0 Jiri Kral over 10 years ago in reply to Richard Prentice

Prodigy 50 points

Hi Richard,

thanks. That would be highly appreciated.

- Jiri

0 Richard Prentice over 10 years ago

TI__Mastermind 29220 points

Hi,

this is the response from the design team:

I think the problem may be related to auto-frequency detect circuit of PLL getting confused after clock loss and stuck to an invalid range after clock is restored.

See a similar note on page 36 of datasheet which also gives a software fix for this.

So, instead of powering down/up the device, you may write 0x01, value 0x0010, disable PLL auto switch, and write appropriate register for sampling clock range. Please let us know if it solves the problem.

Regards,

Richard P.

0 Jiri Kral over 10 years ago in reply to Richard Prentice

Prodigy 50 points

Hi,

this was tried before posting here. Sorry that I did not mention it specifically. The behavior was identical. Only the power up/down did help.

Thanks

- Jiri

0 Richard Prentice over 10 years ago in reply to Jiri Kral

TI__Mastermind 29220 points

Hi,

Thanks. I've forwarded that back to the design team as well. I have a question though - did you just try disabling the PLL autodetect adn setting the range to see if it would *clear* the issue, or did you also try it initially to see if it would prevent the issue?

Regards,

Richard P.

0 Jiri Kral over 10 years ago in reply to Richard Prentice

Prodigy 50 points

Hi,

I did try the prevention (not the recovery, as I do prefer to recover through the PD pin anyways, if the problem already happens).

Cheers

- Jiri

0 Jiri Kral over 10 years ago in reply to Jiri Kral

Prodigy 50 points

Hello,

just curious, if there was some reply form the developers team. I have to make a call soon if to replace/keep the affected ones.

Thanks

- Jiri

0 Richard Prentice over 9 years ago in reply to Jiri Kral

TI__Mastermind 29220 points

Hi,

The design team has set up their characterization board to try to replicate the issue, but this is their response as of today:

Quote:

Can we get two failing units shipped back?

The behavior of failing units is unexpected and will need lab testing. Unquote.

If that is acceptable, then may I ask where you are located? Do you have a local TI or distributor contact that we can make arrangements?

Regards,

Richard P.

Data converters

Data converters forum

ADS5281 behavior on clock loss