[Top Cap]

I assume the scope is in reasonable calibration. It shows the sawtooth at 10ms which is 100Hz and the sort of waveform you would get from a poorly smoothed full wave rectifier.
Can you try fitting a suitable resistor across the HT supplies i.e. no valve fitted, the resistor imitating the current drawn by a typical valve. Then check to see if you still have the sawtooth waveform? At least you should be able to ascertain whether it is power supply problem or something happening when a valve is in circuit under test.