The "second filament" is just another fusor grid, a poor one for fusion, but it's a great ion source due to the increased d in Paschen's pxd rule. It's glowing in the pic because it's a lousy fusion grid and the wires get hit a lot by ions due to poor focus.
This 2nd grid, when things are "biased up" shows a power gain of about 100 if you think of the main grid and tank as a PNP tube (sort of, it's a plasma triode), with the "control" electrode being that second grid. This is just electrical power gain, like most active devices have.
This means you can make the thing oscillate (maybe more than one frequency at a time) and I'm moving in the direction of a pops type of mode. More on my website/forums under theories and actual fusor runs threads. Most fusion seems to happen at the onsets - it then gets into a dynamic equilibrium that is poor for fusion (which no on has a credible theory for, but it's there to observe). Thus, getting the thing to oscillate (see https://www.youtube.com/user/DCFusor
) has just shown us gains in Q (defined as fusion per watt input) of around 2800 better than we had before - another record, and that was "first light" and no tuning of anything - just oscilatioin at some random frequency. I expect it to get a lot better once we learn a little more and tune appropriately. That was a relatively huge leap. Incredible for a "first go" situation. You'll do better watching my forums and youtube channel than here - this is getting awfully crowded and disorganized.