It has the look and feel of a capacitor charging/discharging somewhere, but I can't put my finger on it.
Shorting the output via the front-panel current set button, it always drops down to a few mV instantaneously (I would hope so given it is a physical short!).
But when I release the button, it ramps up to the set voltage. The ramp-time is 1) nearly instantaneous for 1V, 2) about 2 seconds for 9V, and 3) about 6 seconds for 19V. And I'm considering it done when the 1 mV digit on my DMM stops moving (the DMM is set to 1 PLC, so it is updating very fast and is not the cause of the delay). The ramp-up is certainly non-linear, as it gets to about 18.900V within a second before taking another 5 seconds to stabilize at 19.000V.
Same basic behavior when down-ranging to a lower voltage - the greater the change, the longer the time.
As I said before, once it stabilizes, it stays put well within the drift spec. And the noise spec seems fine (at least as far down as I can measure). The 100 uF output cap is new.