Computing stuff tied to the physical world

Now THAT’s a clock!

In Hardware on Mar 10, 2012 at 00:01

Triggered by a video on EEVblog of a Rubidium frequency standard and its teardown, I decided to get one myself. These things are available from eBay for around €40 these days, as recalls from cellphone towers, apparently:

DSC 2919

It’s about the size of a hard drive, it’s completely closed, and there really is nothing to it – connect power, wait a few minutes for things to stabilize, and out comes a 10 MHz signal – a perfect black box:

SCR25

There’s a capacitor in the output circuit, so the resulting signal is AC-coupled and hence centered around 0V and 1.5 V peak-to-peak. There’s also a 1 pulse-per-second (PPS) output signal, with a 1 µs pulse (at first I thought it didn’t work, but the pulse is really there).

The big deal about such a Rubidium-based atomic clock is its accuracy. More on this tomorrow…

  1. Are you planning a new time zone? Jee Standard Time? A few more hours in the day would be nice :-)

    Maybe this explains my weak mobile signal, have the pikeys stopped stealing the copper and taken to shifting bits of mast equipment via ebay?!

  2. “it’s completely closed, ” Still closed ? Or did you look inside.

    • Nah, didn’t want to unmount it from the base plate it came on – there’s a teardown here if you’re interested.

  3. No, not the pikeys (for once) ! These units are on the market due to the tightening standards required for cell networks and reaching their predicted (very conservative) end of service life. Though what actually wears out inside puzzles me.

    Just to get a sense of scale, Telcos use a heirarchy of clock precision, the top level requirement is a stable, 1 part in 10 to the eleventh power. Still struggling with the billions?

    That is about one second in 3,000 years….

    Beats Tag Heuer any day…. ;-)

    • Most of the FE-5680A units now on the market use a 60 MHz VCXO inside, which is locked to the Rb resonance and is also divided down for the 10 MHz output. This crystal oscillator has a long-term drift and eventually the center freq. is offset enough that it can’t be steered into lock at 60.000000 MHz. I suspect that is one of the first parts to “wear out”, or at least exceed a specified tolerance window. One of my 3 units would not lock for this reason, until I adjusted a trimmer (see also, FE-5680A FAQ http://www.ko4bb.com/dokuwiki/doku.php?id=precision_timing:fe5680a_faq ).

  4. A tad more exact on time: http://www.physorg.com/news184517462.html (one second off each 3.7 billion years)

  5. …and you think interfacing to SMD’s is tough – this is a single atom !

  6. I have a few 5680-A’s here , and a Trimble Thunderbolt to verify them against. They are nice units. Don’t reverse the 15v polarity … I just killed one :-( And get a Heatsink for the unit (even when mounted on the PCB), it can generate heat but needs the ability to “Dump it” to the surroundings.

    Timenuts has a page here http://www.ko4bb.com/dokuwiki/doku.php?id=precision_timing:fe5680a_faq

    And they also explain why the 1-PPS disappear if you drive a LED directly from the LOCK signal.

    /Bingo

Comments are closed.