I'm writing this up as kind of a post-mortem. Not long ago, someone told me that the next time I strip something down for parts, I should write it up. (ok, Fabienne made me do it.) alas, I didn't remember that until after I'd done the deed... but at least this time I hadn't thrown away all the pieces.

Every now and then you might have a need to integrate a radio controlled clock, or maybe you just like the pretty pictures on your 'scope... or something. Anyway, OEM WWVB modules are apparently stupid-expensive, and quite rare. Thankfully there's another source: consumer "atomic" clocks. I die inside just a little whenever I hear them called that... too bad fry's doesn't carry rubidium or cesium frequency standards.

Today's victim is an Atomix 13131. It's very accessible for the curious - only 4 phillips screws, and the back comes cleanly off. Just $13 at a couple of online retailers. As my dad would say "I couldn't build it out of plywood that cheap!"

Inside you'll find a small control board. Not much to see, really - there's a blob of black epoxy covering the clock chip on the other side. Note the cleanly labled GND and VDD pads in the lower left corner - as if the red and black wires (which I'd de-soldered by this point) weren't a big enough tip-off.

On the right side, you'll see the 4-conductor ribbon cable which connects to the WWVB detector module. I just cut the cable rather than de-soldering it since the other side has some kind of rubbery spooge on it. Strain relief, I suppose. From the top, the conductors are labeled P, O, V, G - enable (active-low), binary out, V+, and ground. I marked the ground conductor with a black pen so I knew which signal I was connecting.
Here's the actual WWVB detector, still glued to one of the clock's panels (couldn't be arsed to break out the dremel). I'm going to guess it's a C-Max CME6005 chip under the epoxy. The "P" signal is consistent with the notation and purpose ("power-on") in the data sheet, and C-Max will sell these chips as bare dies for mounting directly on circuit boards.

You can see some of the rubber spooge on the cable. It's a pretty tough compound and it was getting in the way of my scope probes. Good thing I marked the ground wire, since the pads weren't labeled on this end.

After a little bit of soldering (ie. P and G to ground, V to V+) and some heat shrink tube, we have this little mess. The black wire is ground, the white wire is binary out.
Speaking of binary out, here's a sample of the raw data you get. The 5th pulse from the left - with the chip on it's shoulder ;) - is an 800ms sync bit, the skinny pulses are 200ms ("0") and the wider pulses are 500ms ("1"). Since the sync bit only comes once a minute, and you're never guaranteed to get good/correct reception, it can take multiple minutes to determine the time. Then you still need to parse the bits to drive your clock.

See also http://leapsecond.com/pages/sony-wwvb/

up to parent directory