Let me try to specify some issues even further.
First of all, BLP does NOT always pulse their Suncell. Yes, they often pulse it in order to try and have some control of the reaction rate and perhaps for other unknown reasons. However, on YouTube there is a video of Randell Mills giving a lecture at a university during which he speaks about the negative resistance regime (using different words but explaining the inverse voltage/current relationship and how it offers a self re-inforcing reaction rate increase) AND shows a video of a reactor producing a GIANT plasma ball while, he claims, all power is CUT OFF. That's right, a self sustaining plasma ball producing massive excess energy with no input. Of course, he also goes onto mention how he at first bragged to an investor about the huge energy output and how they vaporized tungsten components of the reactor. The investor wasn't happy at all and made a comment that a reactor that destroyed itself wasn't viable (or words to that effect).
Secondly, I'm suggesting that we don't use a conductive metal or liquid environment in the reactor. BLP utilizes a conductive liquid metal environment in the Suncell to reduce the voltage required to get the discharge going. Otherwise, they'd have to use a higher voltage to ionize the plasma especially if they are operating at normal atmospheric pressure. Please note that Andrea Rossi had to use a powerful high voltage jolt to initially ionize the interior of the E-Cat QX before he dialed it back down and allowed the ion acoustic oscillations to self generate for several seconds before repeating the cycle (to avoid overheating).
Finally, what I'm suggesting is to use a GASEOUS atmosphere (nothing electrolytic and no liquid metal) in combination with a higher voltage to initialize the plasma ONCE - not over and over again like Paulo Correa's Pulsed Abnormal Glow Discharge device. Once we ignite the plasma with a high voltage, then we reduce the voltage dramatically and LIMIT THE CURRENT so that we stay within the negative resistance regime and produce a steady state plasma ball. Once this occurs, we can adjust the properties of the circuit so that it continues to increase in self organization and detaches from the cathode or anode and becomes free floating.
In theory, in my mind, although technically this may or may not be a good method, we could theoretically even use two batteries: an extremely high voltage voltaic pile and a much higher current battery along with a control box. In this mental model (which I realize is NOT practical) we use the high voltage battery to ionize the gaseous atmosphere for a brief instant and then switch over to the high current battery to sustain the plasma discharge. Somewhere between the battery and the discharge tube should be a variable resistor to make sure we do not allow the current to go so high to create a true arc discharge. Once we create the macro-scale EVO, we start "tuning" the circuit. In the simplest of my mental models, we would just add some turns to the wire leading to the discharge tube to create some resistance and probably even less capacitance (unless we made that segment of wire really wide to create greater capacitance between turns in the solenoid). We wouldn't be seeking to quench the discharge and then re-trigger it again like with the PAGD. Instead, we adjust it just enough to make the fireball detach from the electrode and make the ion acoustic oscillations on the oscilloscope increase in amplitude. If this is done in a calorimeter, we should be able to determine pretty quickly the most optimal settings to produce excess heat.
I hope this clarifies my views somewhat.