Thanks everyone for the helpful comments. I have also been measuring the AD587 on a Geller SVR (without any of the trim circuit). Attached shows two plots, the first without twisting the leads and with Cout=0 (like a stock SVR), the second with twisted lead connections, shorter leads to the two 9V batteries and also added Cout=0.1uF (X7R) to the SVR. Room lights were off during these tests and no cellphones were nearby. In the second plot I am not getting the periodic voltage spikes so this appears to indicate either (or both) of the mods (added Cout and twisted+shortened leads) fixed the problem. Measuring at the ppm level can have it's challenges!
Best regards,
Jeff