[time-nuts] worst case jitter of PTP (IEEE1588) derived clock?
magnus at rubidium.dyndns.org
Thu Mar 3 05:21:24 UTC 2011
On 03/03/2011 03:54 AM, Chris Caudle wrote:
> Has anyone found reference which derives the worst case clock jumps to
> expect when using PTP (IEEE1588-2008) and how to derive a phase noise
> spectrum from that?
> I'm looking at a protocol which uses 1588-2008 to create a common clock
> between networked devices, and those devices are supposed to derive other
> clocks (e.g. sampling clocks) synchronized to the 1588 clock. If you are
> locking a sampling clock to that, you will need to know the worst case
> phase noise so you can design the PLL bandwidth, etc. So far haven't
> found the right reference which derives that, just some very gross maximum
> offset type of numbers, nothing which shows whether you might go +/-
> maximum offset in one cycle, or whether it drifts slowly to that level of
> offset, etc.
> I don't recall seeing much about PTP (Precision Time Protocol) on the
> list, compared to NTP. Are many people using it yet?
If you view IEEE1588-2008 as a frame-work standard you are closer to the
a true picture. Various implementations can derive quite different
properties from it. If you do native Ethernet or over IP/UDP is one
thing, hardware clocking support in the PHY, in the MAC or separate
controller also cares... as do details in the PLL algorithms. I've seen
some presentations showing the improvements with new algos.
Your mileage will vary, and it will vary with what kind of support you
have in involved interfaces on computers and on the switches in between.
You are asking a quite reasonable question, but in practice there is no
simple answer, you need to try out the equipment you want.
I think the PTP business is not mature, so you have to tread carefully
and verify your own steps. You will see different quality level
equipments out there from "we can stamp another cool standard on it"
almost up to "we provide highest possible quality technically possible".
So, with that in the back of your head, I think you will find articles
showing practical results which give you an indication of what can be
achieved and that's about what to expect. Vendors white-papers etc.
should be viewed as the marketing material it is, maybe a bit to "ideal"
More information about the time-nuts