I'm running a measurement on the first sample right now. The circuit is attached. R1 is actually a resistor decade set at 515 Ohm, that is a reasonably stable point. The zener is connected through an IC socket, the 10V source is my JVR reference, so it is quite stable, C1 is a Mylar film capacitor. Right now I'm using the 34465A. I might leave it running over the weekend with 3558A though.
Cheers
Alex
P.S. - first impressions are ... interesting. Very little random noise but the device is (relatively) sensitive to temperature variations and air movements. I suspect that it can only be useful with a good temperature stabilizing loop. With ~7.5mA and 6 Volts the dissipation is quite high and so is self-heating. As a result, any air movements produce noticeable voltage shifts (few ppm even with a good foam envelope around the device). The dynamic resistance at 7.5mA is about 13-14 Ohm. We'll see how it performs over the weekend.