Page 1 of 1

FAQ - Deuterium gas inlet port location.

Posted: Fri Jan 11, 2013 5:24 pm
by Richard Hull
When designing a spherical fusor, the position of the gas inlet port should actually be on the opposite side of the fusor, (180 degrees), from the bore of the vacuum manifold leading into the fusor chamber.

The gas, (D2) that you are effectively "leaking" into the fusor needs some small amount of time to diffuse into the current, extant chamber gas load.

Placing the gas inlet in the vacuum manifold or close to the vacuum line inlet would probably just waste a lot of gas. A great idea would be to admit the gas tangentally at about the 135 degree point from the vacuum inlet such that the tangental flow headed for the 180 degree point, hugging the chamber wall.

Naturally, at the pressures involved, we are in molecular flow and it would seem reasonable that this port placement is not all that critical, but mechanically, it seems to make sense to not put you gas inlet inside or right at the vacuum manifold.

Currently, on fusor IV, my port is at 90 degrees from the vacuum port.

There are many points in the mechanical layout and design of the fusor that should be adressed. There is already an older FAQ that covers this in some detail.

Richard Hull

Re: FAQ - Deuterium gas inlet port location.

Posted: Fri Jan 11, 2013 6:13 pm
by Chris Bradley
I'm not sure I necessarily agree with that, Richard. Feel free to delete this post if you think it is wrong and detracts from the FAQ, but the mean free path, especially in the sub micron range you enter before back-filling, would mean gas [at least, a significant %age] would simply run straight through the fusor to the other side and straight out, would it not?

I have two flat bulkheads at the distal ends of a cylinder. I've put inlet and outlet in the same bulkhead, but apart by some 200mm offset, so that the incoming gas has a chance to enter the chamber and bounce around a bit first. I feel this is about optimum, given the mean free paths involved.

For viscous diffusion to take place readily before there is a chance of gas molecules heading straight out the exhaust opposite before they collide with anything else at all, the MFP should be substantially less than the chamber dimension. As you have said, in a way you should try to arrange it so that the input flow will immediately start bouncing around the chamber, which will better randomise the directions and help diffusion. This could be done equally by putting a baffle over the input vent. A simple way to achieve this, which I have actually done before but I did not get the impression it made a hill-o-beans difference, is to bend and wedge a piece of material in the input port that will fill a half of the circumference of the port so that the half closest to the output is 'shielded' towards it, and scattering is then encouraged away from the exit port but rather more into the rest of the chamber. In this way, the path the gas molecules have to take has to be a 'longer route' around the chamber and several scatters later, thus should then be better mixed.

Re: FAQ - Deuterium gas inlet port location.

Posted: Fri Jan 11, 2013 6:58 pm
by Richard Hull
You assume that all the gas molecules are columnated in some fashion. They are fully thermal and go in all directions, however, the tangental entry or right angle entry I mentioned would even obviate the MFP argument assuming absolute columnation was occuring at the inlet port.

The average fusor operates (fuses) at 10 microns and the MFP for the cool inlet gas is on the order of a centimeter, at most, so a six inch chamber should not create an inlet facing an outlet issue.

The average gas temperature in the chamber is probably slightly more than the wall temperature, but is a crap shoot guess at best.

I am sure the gas will diffuse fully due to the spacing in a 6" or larger device. Much of this could be a "hill of beans" discussion, but it points out that careful mechanical layout is something to mull over a bit before drilling into the hemispheres to locate ports.

Richard Hull

Re: FAQ - Deuterium gas inlet port location.

Posted: Fri Jan 11, 2013 7:07 pm
by Mike Beauford
I know this may sound stupid, but does gravity come into play here, or as been stated before, it's all thermo/brownian motion.

So does putting the gas inlet on top of the fusor matter vs. putting it on the side and having the vacuum side on the bottom of the fusor?

Re: FAQ - Deuterium gas inlet port location.

Posted: Fri Jan 11, 2013 7:31 pm
by Chris Bradley
Richard Hull wrote:
> You assume that all the gas molecules are columnated in some fashion.
Not really. What I am thinking is that as the gas emerges from whatever it emerges from, if you imagine that the vacuum port on the other side occupies, say, a 50th of the solid angle that the gas sees around it, so you loose a 50th immediately. Then there is a 'one reflection' loss in which a given area of the shell might be struck so that the gas molecules are reflected directly into the vacuum port. Only after several bounces do we begin to see what would be conventionally called 'randomised diffusion'.

The objective we are after in locating the input port is to ensure that what is drawn out of the output port is each constituent of the contents drawn out of the chamber in respective proportion. But with a port directly opposite the exit, I think this would be biased towards drawing out proportionately more of the freshly admitted gas if there is a line-of-sight between input and output along which a gas molecule may miss hitting anything at all en route.

I think a baffle arrangement that, geometrically speaking, means the admitted gas has to bounce around several times would encourage better randomisation of the gas molecule's velocity and direction together with the existing gas in the baffle area and, thus, diffuse more randomly into it.

Re: FAQ - Deuterium gas inlet port location.

Posted: Fri Jan 11, 2013 7:33 pm
by Chris Bradley
Mike Beauford wrote:
> I know this may sound stupid, but does gravity come into play here
The gas molecules are moving at the speed of sound (for the temperature of the gas at the point of admission). This is too fast for any significant deflection of a gas molecule's trajectory over the dimensions of the chamber.

Re: FAQ - Deuterium gas inlet port location.

Posted: Fri Jan 11, 2013 10:22 pm
by Richard Hull
Did anyone note that I figured 10mm as the MFP! there are ~12 MFPs to the other wall.
The stuff is randomized well without a baffle. the speed of sound or the speed of light means nothing if the physics says your MFP before collision is 10mm! After 4 or 5 MFPs, (~2 inches), you are not going in any particular direction based on what you were doing way back at the inlet.

Richard Hull

Re: FAQ - Deuterium gas inlet port location.

Posted: Fri Jan 11, 2013 10:45 pm
by Chris Bradley
Richard Hull wrote:
> Did anyone note that I figured 10mm as the MFP!

At 1 micron, the MFP of a 64 pm diameter (0.64 Angstroms) molecule [which I am lead to believe is the size of a di-hydrogen molecule] at 20C is 166 cm, according to the MFP calculator on; http://hyperphysics.phy-astr.gsu.edu/hb ... enfre.html .

That may amount to 16 cm at 10 microns, but when back-filling the deuterium you would want to drop the pressure lower still than 1 micron, at least to start with.

Re: FAQ - Deuterium gas inlet port location.

Posted: Mon Jan 14, 2013 5:34 pm
by Richard Hull
What?!

The molecular diameter of Hydrogen is ~2.9 angstroms!

http://web.anl.gov/PCS/acsfuel/preprint ... 3_0481.pdf

going to..............

http://hyperphysics.phy-astr.gsu.edu/hb ... re.html#c3

plugging in:
1. 2.9 x10e-10m (2.9A)
2. 0.01mm Hg (we fuse at around 10 microns or more in a Fusor producing a lot of neutrons)
3. 20 degrees C (~ inlet temp)

We get 8mm MFP!

The gas is scattered effectively over the 150mm to the other side of a standard 6" fusor. No baffles needed.

Fusors run for a while to reach one million n/s (2million fusions/sec) The shell of my fusor has heated on occasion to 150C. Assuming all the gas inside, by that time, is at the shell temperature and not 20C, then the mean free path stretches to 9mm

Richard Hull