Tracking Down FF-2B’s Problematic OscillationsSeptember 13, 2019
LPPFusion’s research team continued to track down the causes of the asymmetric gas-to-plasma breakdown (start of the electrical current flow) and the associated, problematic, early current oscillations in FF-2B’s pulses. We’ve made progress in isolating the causes of these early oscillations. This is now the main focus of our research and it will probably take a while to complete a solution. Getting a symmetrical breakdown at the start of the pulse is a necessary step to getting the higher densities and fusion yields that will result from a strong compression at the end of the pulse. Thus, these early oscillations are preventing the expected higher fusion yields.
Experiments in early August showed that the oscillations persisted even when we used nitrogen rather than neon as the mixing gas with our deuterium fuel. Nitrogen allowed a stronger preionization current to smooth the way for the main current, but we still did not get a swift, even breakdown. Instead the pulses got off to a slow and unsteady start.
LPPFusion Chief Scientist Eric Lerner noticed that the oscillations in current from different shots looked almost identical (see Figure 1.) This was a big clue that the oscillations were caused by something unchanging in the device, not a random process. Was this some asymmetry in the device that we could fix?
Figure 1. (a, left) The Main Rogowski Coil (MRC) signals for shot 1 (orange) and shot 2 (blue) of August 9, 2019 measuring the rate of change of the current, are nearly identical in the first 400 ns at the start of the pulse. This shows that whatever is causing these big oscillations is not random. (b, right) The MRC signal for the whole of shot 5, July 9 shows the early oscillations dying down, but then a poor pinch at 1500 ns, indicated by a relatively small downward spike in the signal. A small spike (a good one would go down to -10 kA/ns) indicates poor compression and low fusion yield, confirmed by neutron measurements.
To take a closer look we decided to take photos of the gas-to-plasma breakdown from a window below the electrodes, so we could look for a cause of asymmetry in the current sheath formation. The only way we could do this with an ordinary camera (instead of our complex ICCD camera) was by using a trigger pulse instead of the main, mega-amp current pulse. The trigger pulse is used to trigger the switches when the capacitors are charged. But even if the capacitors are uncharged, the trigger pulse by itself can create a breakdown very similar to the start of the main current pulse. The advantage is that the trigger pulse is too small to damage a camera the way the main pulse could by excessive light, X-rays and radio waves. The photos we obtained in this way (Fig. 2) show the anomalous speck formations on the anode, near the insulator (presumably bits of melted beryllium) were indeed causing an uneven, asymmetric gas-to-plasma breakdown. To remove these pesky specks without opening the chamber, we fired the main bank several times, using a heavy mix of nitrogen with the deuterium. This, we calculated, would cause the current to linger longer on the bright-white specks, vaporizing them. After each shot, we checked with the trigger photos (Fig. 2) to see if the specks were gone. After four shots, we saw the specks were first swept down the anode and then away altogether. This is an important step, as it shows we can clean up these specks if they form again without the time-consuming opening of the chamber.
Figure 2. Trigger photo images taken from the window below the electrodes show the start of a pulse with the trigger as the source of the current. The images were taken after “cleaning” shots with the main capacitor bank. After shot 2 on Aug. 27 (upper left) current is concentrated on two specks (bright spots) sitting on the anode near the end of the insulator (brighter ring). After shot 4 (upper right) the spots near the insulator disappeared, but one spot appeared near the end of the anode (in dark area). After shot 5 (bottom) the spots had disappeared and the breakdown appeared symmetric in these long-exposure photos. The dark outer objects touching the purple rays are the cathode vanes.
However, the early current oscillations only decreased by about 25%, so it turned out that the specks were not the biggest problem. We then looked at the frequency spectra of the oscillations (Fig. 3). These oscillations spectra show clearly that the oscillation at 16 MHz was larger in the 2019 shots (blue line) than in the shots in previous years (orange line, 2016). We know that these 16 MHz oscillations are caused by some current sloshing back and forth between the switches and the electrodes. But the spectra also showed a relatively narrow peak at 40 MHz. Such a narrow peak—like a pure note in music—is probably produced by some resonance of the current in the external circuit. Once we find out where this resonance is, we should be able to fix it by either eliminating the resonance, or increasing the energy absorption (damping) of the circuit or both. A well-damped circuit will not oscillate at all.
Figure 3. Frequency spectra of the early current oscillations shows that in a typical recent shot (July 9, shot 5, blue line) the peak at the 40 MHz frequency mark also appears in the oscillations of X-ray power during the pinch of the same July 9 shot (white line). This indicates that at least the oscillations at this frequency have an effect on the entire current pulse. By comparison, in a typical shot from 2016 (June 6, shot 5, orange line) the 40 MHz peak is entirely absent and the peaks at 16 MHz and 26 MHz are much smaller. The tall, narrow spikes in the spectra of the 2019 shots indicate resonances in the external circuit of FF-2B. We are researching this issue. The 7 MHz peak in the X-ray plot just shows the width of the X-ray pulse and is not of concern.
While we are hunting down the source of the early oscillations, we are finalizing our designs for the new switches that will increase current in the device. We are also installing the remote controls of the trigger and charging circuits that will allow us to experiment with our proton-boron fuel.