Electronics > Projects, Designs, and Technical Stuff
Are low powered radio signals ever used to share a timing pulse locally?
(1/5) > >>
cdev:
Suppose I have a GPSDO that gives me a nice, low jitter high precision 1PPS and 10 MHz carrier. Or even just a GPS.

Given two identical GPSs their PPS signals are usually (very very) close but they do vary sometimes, probably its reduced to a very low value within the computers (better GPSs also allow you to adjust the cable delay setting) and then if one is connecting them over a network, the network introduces error. If the packets need to go through a router or traverse subnets within a switch or possibly two switches, thats more possible variability between machines. NTP compensates for it and for the most part, does very a good job. But it has inherent limitations.
For example, it seems that cable modems are inherently problematic in a way other networking equipment isnt preventing any really precise measurement of that last bit of delay.

Also, has anybody ever found that different parts of the world may observe slightly different times?

Two GPSs with 1PPS is probably as good as it gets for networking. And as good as it needs to get. But are there other ways (radio) to get a timing pulse that have been standards in the past, or which are maybe used in some settings? (places where GPS signals dont go on Earth, space outside of the Earth's gravity, and GPS signals reach, or where spoofing is a concern)
dmills:
I can hit 100ns of timing jitter on a 10Gig network connection, but you don't do it with NTP, you want PTP with the network cards having suitable support (Many of the modern ones do), and ideally a switch that knows how to play transparent clock.

As far as RF goes, there are plenty of good frequency references broadcast by the national broadcasters that can be used to derive good local frequency standards, time is a little harder as you really need to know the offset between the time standard transmitter and your location, and that is only sort of predictable (Particularly below VHF the ionosphere can reflect things with somewhat variable delay, which can be very annoying if you also have a groundwave path).

The cell phone network has a relatively good timebase but converting that to wall clock turns out to be annoying.

A telescope pointed at a pulsar or two works in space for just about any space we can reach, and given sufficiently accurate tracking of at least 4 pulsars can give you time as well as position and orientation.

Regards, Dan.
Berni:
Cell phones actually have very accurate timing.

They time there signals so accurately that they can measure the time of flight and determine the distance from the tower. The reason for this is that cell towers have a limited number of frequency channels available to them so they use one radio channel to talk to many phones simultaneously by assigning each one its time window. To allow so many timing windows to still be efficient they pack them together closely. The problem is that when a phone is far away from the tower it gets its signal delayed so much that it overshoots its timing window and might jam the signal of the phone in the next window (that maybe happens to be right next to the tower). So the way this works is the phones synchronize there own oscillators to the accurate rubidium or GPSDO clocks in the cell towers while accurately measuring round trip times to determine the exact delay to the tower. With this they know exactly when there transmit slot begins at the tower and they start transmitting earlier than that so that once there signal reaches the tower its spot on in the beginning of the time slot.

Given the large ranges possible with GSM this gives a pretty large area of time locked devices trough direct radio communication. Tho don't think it you could make much use since you need a bulky GSM radio module for it and you would need to hack it to get it to provide a timing signal.

In any case if maximum accuracy is important you can't beat GPS. The GPS system was designed for timing from the ground up and is incredibly good at it.
madires:
In central Europe there's DCF77 which is used by tons of clocks for synchronizing. You can get inexpensive receiver modules for a few bucks and also professional receivers, all supported by the classic ntpd. A GPS module, a DCF77 receiver plus a few reliable ntp sources (like the ones from PTB in Germany) should keep you up to date ;)
cdev:
Maybe one of you can help. I am having weird lack of DNS connectivity that seems un-amenable to anything, it may have to do with my use of my own DNS servers which are supposed to use DNSSEC. That part is working. But some domains dont resolve. I am trying to trace it down. My ISP claims to not support IPV6.

However if I left everything on default I would be connecting via ipv6 much of the time.

I am not super comfortable with IPV6 because I feel its creating a loss of privacy. Also i am concerned about security. Mobile devices in particular, tablets are convenient for surfing (although I hate typing into them and basically almost never do that) but especially they seem to be problematic from a security point of view and maybe the companies that sell and design them may be doing that on purpose. In any case it seems that the settings I want to turn off certain features dont exist. (or if they do they are very difficult to find)

I am just speculating here, and many of you know I do that a lot -

Because some mobile devices and even some browsers involuntarily seem to do things in the ipv6 domain I am not comfortable with, and I am unable to resolve some domains and cant figure out why, among other things, I am trying to learn more about the network, something I should have done decades ago but never really did. In part I hope to be able to have precise timing as one more bit of info that I could use.  For what its worth.

Except I don't know where to even start.
Navigation
Message Index
Next page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod