I'm thinking that a modification of what makes quadrupole mass spectrometers and ion traps work will be interesting (at least) in a fusor setting, as after all these have some interesting properties. They do lack compensation for any space-charge repulsion effects, but they do give a starting place to get into the right octaves of parameter space, which is otherwise daunting. The idea of having your ions moving though a focus cyclically - and at high speed, coming through this point at high density, then slowing as they pass focus (to maybe re-smooth their distribution while at low energy) - seems attractive and would be the long-sought "recirculation" Farnsworth/Hirsch/Meeks thought would be required.
I found this link helpful in having some "worked" examples to make sure I was using the right units.\
However, the master class (which assumes you know the unit system used, the glossary is horrible on that count) is the book edited by Peter H. Dawson,
Quadrupole Mass Spectrometry and its Applications, ISBN 1-56396-455-4, (AIP) which covers a lot of possible configurations. Seems not much changes other than the factors 2 and 4 in the "a" and "q" equations - which are easy to see in the following code - and at any rate, not all over the place for the purposes of knowing where to begin experimenting.
Other measurements made here seem to indicate that in the part of a cycle where the ions are spread out at the tank walls - we CAN treat them as not reacting much to space-charge repulsion. In any cyclic setup they only get "dense" near the grid (we've seen way over 1000::1 "compression ratios) where it might not matter if they've already established a trajectory and gained high energy (a little repulsion won't matter - it's too late for it to have much effect at that point).
So, here's a little perl code you can diddle and mess around with. In my main setup this looks like we're going to have to be cafeful to not mess up a ham band or maritime comm between 160 and 80 meter bands. While it's only a theory, I'd assume that to make fusion happen we'll still need some high voltages and velocities, so while you can in theory trap a few D ions with a few volts, that's not a useful regime for fusion. We might not need the 10's of kv we need with DC drive, however, as measurements here indicate that with ~ 50kv dc drive, due to various effects, most of the D+ ions only get to around 5kv energies - probably part of the reason fusors are so inefficient..
So, here's my simple calculator in cut and paste form:
Code: Select all
#!/usr/bin/perl
use Modern::Perl '2015';
# tell F and ac/dc volts for Deu vs r
my $z1 = 95800000; # e/m of proton in coulomb/kg
my $a = .15;
my $q = .6;
my $tpi = 3.1415927 * 2;
my $em = $z1/2; # assume deuterium or close
my $r = 3; # radius input in inches
my $r2; #radius converted to meters and squared
my $f = 3500000;
my $tpf2; # f as radians squared
my $U; # DCV
my $Q; # AC Pk V
my $dummy;
while (1)
{
say "what radius, inch? ($r)->";
$dummy = <>; # read input line
chomp $dummy; # remove the newline
no warnings "numeric"; # don't bitch if it's now empty
if ($dummy > 0) {$r = $dummy;}
$r2 = ($r * 0.0254)**2;
say "radius meters squared = $r2";
say "what freqency, Hz? ($f)->";
$dummy = <>; # read input line
chomp $dummy; # remove the newline
no warnings "numeric"; # don't bitch if it's now empty
if ($dummy > 0) {$f = $dummy;}
$tpf2 = ($f * $tpi)**2;
say "radians squared = $tpf2";
$U = ($tpf2 * $r2 * $a) / (4 * $em);
say "\n\nU is $U VDC";
$Q = ($tpf2 * $r2 * $q) / (2 * $em);
say "Q is $Q VAC pk-pk\n\n";
}
It looks like this in a "real" text editor:
- Easier to understand with color?
I'm guessing that like those coefficients "2 and 4" used here, things might need to be double or half based on our electrode configurations (this is true in the Dawson book for some that use 3 and 6 or 4 and 8), but since there are two other things that are squares of the parameters...this reduces the search space considerably with just a few iterations...and who knows if my guesses for best "a and q" in the middle of the stability region are the right ones? But given the huge multidimensional search space otherwise, this is a heck of an improvement.