AM335x and IEEE1588 drift

M.A.M

Hi,

Where can we read about the relative timing error/drift (ns) in the timestamp mechanism of the IEEE1588 implementation?

The timestamp itself in Stellaris is using 16bit values, so I guess this is the same in AM335x also giving a limitation in accuracy , i.e 2^(-16).

/Magnus

over 13 years ago

0 Lawrence Ronk over 13 years ago

TI__Intellectual 2820 points

Hi Magnus,

The timestamp register in Sitara devices that support IEEE1588, like AM335x, is 32-bit wide. So, the limitation cited below is not applicable to Sitara devices.

Regards,

Lawrence

0 M.A.M over 13 years ago in reply to Lawrence Ronk

TI__Expert 8805 points

Hi again,

Is there a measurement report of drift/jitter that you can share, on real timing accuracy versus a reference IEEE1588 node?

I.e what values to expect in ns?

Thanks.

/Magnus

0 James McCartney92666 over 13 years ago in reply to M.A.M

Prodigy 10 points

In addition to drift which I will assume is directly related to the quality of the oscillator used in the system, do you have any measurements on stability of the IEEE 1588 PTP subsystem. Specifically you know of anyone who has made Allen Deviation measurements on the steered frequency? It would be interesting to see what this curve would show.

Thanks

James McCartney

Processors

Processors forum

AM335x and IEEE1588 drift