It remains to be demonstrated that a direct gate input will perform worse than the closed loop version as applied in Guinevere's plate load circuit.

We're looking at another CCS for the cathode load, but that one is a drain-end rig instead of the source end as applied in plate load apps. Actually, if it could be easily arranged to deliver + and - voltages at the same time, a pentode would make an acceptable CCS. I like the heaterless action of the MOSFET's for this level of experimentation.

Low voltages do have their attraction.
regards,
Douglas

The circuit is simple enough that there can't be too many troubles.

On to the FMEA...
Hum can be from two sources, the plate supply or the input. THe cathode counted as an input.

If it is the plate supply, I suggest MOSFET failure. It is not likely that a working CCS lets enough hum through. Check plate voltage, and if possible measure out-of-circuit gate-source resistance. it should be infinite, excepting some capaciatance( but we're talking DC...)

If it is the input, there are more troublesome things to check. grounding is too wide a net, look at individual ground points, and check physical integrity. Of lesser likelyhood, a filament-cathode short would cause hum of some, given that we're not running pure DC.
regards,
Douglas