First, some background.
The members of time-nuts suggested that the Fury Interface board might benefit
from using a 5MHz OCXO instead of the 10 MHz, that is specified.
I agree with this since I have a bunch of different 5MHz devixes that I'd like to test with the FURY.

Bruce Griffiths suggested that I use the NIST frequency doubler in the paper:
NIST_Ultra-High_Stability_Synthesizer_For_Diode_Laser_Pumped_Rubidium.pdf

This paper, has a tiny schematic of the doubler (above the chart on Page 3, Figure 3) so I had to do a bit of digging to get out the needed
parts.

This paper: Superimposing_Low-Phase-Noise_Low-Drift_Instrumentation_Techniques_On_RF_Design
helped shed some light on the parts.

The critical part is the Siliconix Dual FET. Siliconix, it turns out, was bought by Linear Systems and they make a similar part: LS840.

Here is a copy of the datasheet. This part is in production and they provide a PSPICE model which is what I used to model the circuit.

So, here is the schematic: