The test boards arrived today with the TPS-54231 based buck converter arrived today. The noise and ripple on the +15V line is dramatically better than even the initial prototype boards. At the moment, there's around 20 mV p-p of ripple and just a little bit of high frequency noise.

When powering a 5680, the short term stability is also much, much better. After a brief warm-up, while free-running the unit is staying in a phase corridor ±2 ns wide.

The good news is that I cheated and ordered the next rev of the GPS boards already, so they should be on sale by Maker Faire this weekend. I'll have one there doing calibration duty for the Crazy Clock if someone wants to check it out.