I increased the EVK1 voltage from 1.0 V to 3.3 V, which also increases current by a factor of three. I now see some pretty significant PPK2 charge injection artifacts that are larger than the noise.
I configured the JS220 for fixed 180 mA current range to completely eliminate any JS220 switching artifacts. We are only looking at the load and PPK2.
EVK1 supply → JS220 → EVK1 load:
EVK1 supply → JS220 → PPK2 → EVK1 load:
My conclusion is that the JS220 is measuring the real PPK2 charge injection current, which causes the JS220 to autorange more frequently than when the PPK2 is not present.