I think I may have gotten as close to the elusive universal design that can accommodate both a DOC020V and a DOT050V on the same board as I can.

Mind you, it looks like the DOC020V isn't worth the extra $20, but it's the thought that counts.

I've been using the performance of the current released version as a benchmark. It has a nominal ADEV at tau 1E0 of ~1.5E-10 using a DOT050V and an LDO for power. I tried testing the DOC020V with the same circuit and saw some sort of weird oscillation on the VC pin of the oscillator and the stability went to hell. Adding an RC low-pass filter to the output of the analog section made that go away and I got the very best ADEV reading so far with that part of just under 3E-10 at tau 4E0. At tau 1E0 it was 6.6E-10, so it's still not as good as the TCXO part.

Meanwhile, attempting to put the DOT on the same board with the SC189Z switching supply was yielding much worse ripple on the power rail. That, in turn, was pushing the ADEV up to something like 5E-10 at tau 1E0. I may have figured that out, though. Switching supplies are very load-sensitive. If they don't have a certain amount of load, their ability to regulate the output voltage becomes quite poor. That's why the power supply was much more stable with the OCXO - it was drawing much more current. Tacking on a 39 ohm 1/2W resistor across the power rail fixed that. Testing that result yielded slightly worse performance than the LDO version in the short term (from ~1.5E-10 to ~3.5E-10). Given that adding the load resistor reduces the efficiency for the TCXO case, and the DOC020V isn't doing as well as the DOT050V, it looks like the future is going to be the LDO and the DOT050V, and a separate board designed just for the OH300 with the switching supply.