Cross-discipline matching network

[Doug Coulter]

Well, I thought we needed more in this section, and I had this, so...

3" OD 400ml beaker, 2" winding at 36 tpi

This isn't going to handle real high power, but I don't think the coil form is gonna die from heat...I haven't put on the high temp goop yet as I may want to adjust this somewhat.
And yes, I drilled a hole in the bottom of the beaker to have a nice way to mount it. It's not hard with a diamond hole saw and some modelling clay to make a little dam to hold some water. Just go real easy on the feed pressure. If you're working with more shattery stuff, you can glue a backing like plastic on the backside first (breaking through is when things break usually).
Decent Q, here shown in-place with fusor stray C as the load. Like I said, at high power this is gonna fry, maybe. What seems to happen when there are ions around is that the net Q goes down and the input impedance up, so...we'll see.

MiniVNAPro sweep around resonance in reflection mode.

In other news, you can see my house from here! Why this matters is that the camera is southeast of the solar panels, and this gives me start of breakeven or better about an hour earlier than before. As well as the firewood. It also turns out that just seeing a bigger piece of sky is good in "bright overcast" which we have a lot of, and this was the most of that for the least cutting.
I now rarely reach or need full panel output on a decent day. By the time I've got real full sun on the whole array, the batteries are mostly charged up and so the controllers cut the current back.
Of course, that means I can turn on the AC or charge the car no sweat...

Better solar situation.

[Doug Coulter]

Anyone who knows me knows how I feel about that one, but a neighbor sent me this for all our amusement.

But can you start a fire over the internet? Ask Dan Tentler...(Or the CIA or those evil Russians).

[Doug Coulter]

More matching.
The original network and the one after (better coil) just weren't getting it. Not enough Q under load to get the stepup ratio I need...and all the energy going into waste heat.

So I made a better one, or so I think/hope. This one has < 1/2 the inductance, much nicer wire, and 50 more pf (bringing total to 80) across the fusor/load.
The measured Q near resonance is around 10 times the previous version.

If things go as usual, at least I'll learn a little more. At low drive, ions look like a capacitor. When you get enough of them for scattering to happen, they start to look like there's a resistor in there too (kinda makes sense). I expect that C reactance to change and maybe even become an L once the voltage is such that the ions aren't all staying near the grid, and when the DC to AC ratio is closer to what my little ion.pl program predicts should work...of course, the software only knows about "actual field" which can be and usually is a lot different than what I'm putting on there....but you gotta begin somewhere. I bought a really nice RF choke so I can inject the DC without frying the ballast R in the Spellman (and that thing wasn't cheap but it measures just about perfect), and I took some measurements with the whole mess hooked up and "cold" as well as just the fusor + L + C. Looks pretty good, or at least where I think I might want to be for the next trials.

So, all lashed up, it looks like this. Yeah, I know, RF guys gonna cringe - but really for short tests, the parasitics from the sloppy wiring ain't gonna matter that much, and it's not going to radiate that much with a foot of exposed conductor either...or so I hope.

Ready for smoke, but I hope for no smoke...check the nice vacuum cap on the right...thanks Bill.

Here's the VNA plot with just the fusor connected.

No load, around 2 ohms at resonance...

And with the whole schmear but on the fusor side of the 750 pf coupling cap from the amplifier (with the RF choke to the Spellman in circuit too)

At the end of the coil, all else is hooked up

Now, on the RF amp side of that coupling cap, which is the 1/3 turns tap on the autoformer which is also the DC plate choke (type 61 ferrite, about 1mh end to end).

Of course I also saved the more detailed pdfs of these, but they won't display inline here...at any rate, when the power is on I'll be tuning the usual way and only using these numbers as a ballpark place to begin.

That was a lot of work for something "conceptually trivial". There are still too many PhD's running around alive because we engineers didn't strangle them for using those words in a sentence like "why aren't you done yet, it's....".

[Doug Coulter]

All this foolery is based on the idea that at least in part, this can be made to act somewhat like an RF-DC ion trap. But rather than the values used for a mass spectrometer where one wants the edge of stability so ions slightly off in e/m fly out - we want to keep them all, so I chose numbers in the middle of "stable".
As you might expect, higher frequencies and higher voltages go together when you're worried about hitting the tank walls, which is kinda harmful to recirculation.

And higher voltages make repulsion less of an issue out where the density is lower, far from the grid at least.it's going to be an issue near focus no matter what, but the hope is that a lot of the trajectory destiny can be set before that. Higher voltages and higher frequencies are harder to get to in the lab - I now understand why the cyclotron guys put their matching networks in their own vacuum tank (corona, radiation...arcing). So I picked some initial numbers that while on the low side of what might work, are at least achievable in the lab.

And if you can run perl programs (linux or mac make it easy) here's one that gives the numbers based on a "stable" region.

Ion.7z

make executable and run from a terminal, ctrl-c to exit.

(708 Bytes) Downloaded 300 times