Yes,I use a common DVM as the output voltage is goes through a very high resistance voltage divider of about 300 megohms.
Then it is buffered by a unity gain non-inverting opamp stage.
This method completely isolates the meter impedance from the voltage divider that would otherwise cause an error in the measured voltage output off of the resister divider.
A multi-turn trimmer pot is used at the end of the resistor string to provide a fine exact adjustment.
The only problem I have is that my meter doesn't have a high enough resolution to measure it's accuracy to less then 10v or even 1v of its output depending on the range that the meter is set to.
But by using a scope it is rock solid even on a overloaded spark situation.
On the discharge test the spikes in the waveform is the lowest voltages during the discharge.
The bottom solid line is the actual output voltage as it is a negative voltage.
The Crosshair at the top is 0 volts and is where this test was calibrated.
The X is the voltage measurement and maybe off for this picture but you get the idea as I was just testing for the recovery time.
The flat side of the pulse is the start of the arc and shows the point of when the current in the arc stops.
Then the supply starts to recover its voltage and is what the curved side of the wave form is until it reaches the flashpoint of the gap as shown by the bottom solid line.
This also shows the regulation of the output voltage of before the next breakdown in the gap occurs.
I have been able to modulate this power supply for up to 70Hz as well and with a few changes there may be a possiblilty to use this supply as a bass amp for a very large surface area ESL system or at least the power supply for such a system anyhow !!
Here are some more pics !!