I only have experience with uT 3+, but for directly heated cathodes the filament supply must be isolated. If it is a DC supply, the average filament voltage will be Vf/2, so the test results will not be consistent with the nominal supplied Vf.
A transformer-derived AC supply is another option. If I recall correctly, Ronald devised a circuit to synchronize the measurement pulse with the supply crossing of nominal Vf, but I have had success measuring directly heated rectifiers without that, as seen in the trace below. Presumably, the normal averaging of multiple measurement pulses produces a result very close to nominal Vf. Note the Ia-Va-k trace is relatively smooth, about as smooth as expected when measuring this type of tube using a DC filament supply.