[time-nuts] Low-long-term-drift clock for board level integration?

Bill Woodcock woody at pch.net
Mon Feb 20 05:08:26 UTC 2012


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256


On Feb 19, 2012, at 7:28 PM, Dennis Ferguson wrote:
> 10, or even 100, microseconds is tough with NTP.  I don't think it is impossible, but it
> requires a good, reliable network connection…

We will have a very large mesh of devices, but the connections between them may be poor.  It's my assumption that some of them will be able to get enough GPS signal (or GPS via a GSM BTS, as we also have a Sierra Wireless GSM chipset onboard) and would thus be able to act as Stratum 1 servers for the others.  Not that there's any big shortage of Stratum 1 servers out there.  It's been a while since I've tried a large mesh of NTP servers, so I don't have much sense of what a reasonable degree of accuracy to expect is.  

And...

> I'm not sure putting a better oscillator
> on the board is likely to help all by itself since ntpd's "magic" internal constants are
> organized to work with the class of oscillators you typically find in computers, and this
> would need to be redone to do anything useful with something better.

…you're probably right about that, so my experience probably isn't applicable at all.  We've supported a lot of open-source development over the years, so I guess I need to figure out who's most actively doing NTPv4 development now, and see if they want to take a whack at it.

> - If you are deploying this stuff in the US, and if cell phones (particularly Verizon or
>  Sprint phones) work where you are installing the stuff, you might look at this for a time
>  source:
>  http://www.endruntechnologies.com/time-frequency-reference-cdma.htm

Nice, I hadn't seen that before.  Only a small portion of our boxes will wind up in the U.S., though, so we're using a Sierra Wireless SL6087 receiver in essentially the same role.

> - Failing that, look at IEEE 1588.  The trouble with this is that it severely constrains
>  the kind of network the equipment is attached to, and the gear used to build that network,

Yeah, I did look at it a bit, more to see if there was anything useful to be learned from it than in an attempt to actually use it, since we're connecting through the general-purpose Internet.

On Feb 19, 2012, at 7:48 PM, Chris Albertson wrote:
> So you can tolerate 10 parts in 3E13 or 1 part in 3E12 drift per year. And you
> have a $300 budget. Somehow I think either the spec of the budget
> will have to move by orders of magnitude.

…or we use something to discipline the clock more often, which is why we've got GPS, GSM, and NTP…  I just have to assume that some or all of those either won't work, or won't work well enough to be useful, in many cases.  With hundreds or thousands of units in the field, it's much more about trying to make a reasonable general solution than trying to get one thing exactly right…  There'll be some sort of bell-curve distribution for the reliability of each of those methods of improving the time, and I'm just trying to figure out the best way to maximize the area under all of those curves within a budget that still allows us to build a useful number of measurement devices.

>  So it's not like you can sync time to GPS in an instant.  it takes
> at least a few hours if you care about microseconds.

Yeah, the Trimble guys ran some numbers and figure that we can get single fixes with ~100 microsecond accuracy in twenty minutes, with the chipset they're selling us.  The tradeoff in battery size to get that down one more order of magnitude isn't worthwhile…  It would push the battery up from $35 to $200, and more battery doesn't have much collateral benefit in our application.

> Again this becomes easy and within $300 if you can have an outdoor antenna.

The number of sites where anyone would be able to install and maintain an outdoor antenna would be way out under one of the skinny ends of those bell curves, so essentially not worth spending any time thinking about.  If wishes were horses, etc.

Thank you very much for the advice.

                                -Bill




-----BEGIN PGP SIGNATURE-----
Version: GnuPG/MacGPG2 v2.0.17 (Darwin)
Comment: GPGTools - http://gpgtools.org
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=QCcW
-----END PGP SIGNATURE-----




More information about the time-nuts mailing list