Vacuum System

Standard Farnsworth Fusors operate at a pressure near 5×10^-3 Torr, or about one-hundred thousandth of atmospheric pressure. At that pressure, air tries its best to get back into the chamber. Any small misalignment between parts or pinholes will leak and ruin the vacuum. The cleanliness of the chamber is extremely important, as any amount of water or oil residue (even from fingerprints) will outgas badly and reduce the ultimate pressure of the system.

My system uses a Conflat chamber and a turbomolecular pump to achieve pressures below 1×10^-7 Torr, which is about the same pressure at the altitude of the International Space Station.

Vacuum Chamber

In the world of amateur fusion, there are mainly two ways to go about designing a chamber. Either you can use o-ring seals in the case of KF and ISO flanges, or you can go with metal gaskets for Conflat flanges. The o-ring seal systems normally use a special rubber called Viton, which has a lower outgas rate than common rubber. They are relatively cheap, as well as easy to assemble and disassemble, but have a worse ultimate pressure.

The Conflat system uses low-oxygen content copper gaskets that are crushed between stainless steel “knife edges.” Due to the system not containing any polymers, the system can reach a much lower ultimate pressure than the Viton options and can be baked at a higher temperature. The downside to the Conflat system is that it is all-around more expensive, and the gaskets are one-time-use, unlike the rubber o-rings.

I chose to go with the Conflat system. It serves as a good test of my abilities to follow a very specific procedure of cleaning and system assembly to get close to the Ultra-high vacuum range. The main chamber is a 4.5″ 4-way cross with many other couplings for various inputs and valves. I bought all of these parts second or third-hand, but I ensured that the delicate knife edges were undamaged.

I took the most time cleaning and preparing the chamber parts before assembly. During this process, it is incredibly important to wear clean nitrile gloves and use non-lint wipes, or you can ruin all the cleaning work you’ve done to that point. When I finished cleaning a part, or was between cleaning steps, I placed aluminum foil over the openings and put the parts into a clean plastic tub to keep dust off of them. My procedure was as follows:

• Run a microfiber cloth through all of the parts to remove dust and debris
• Wash and scrub parts with de-greaser to remove oils from the surface
• Wash with distilled water to remove de-greaser
• Dry with new microfiber cloth
• Place in an oven at 350 Fahrenheit for 20 minutes to evaporate the leftover water
• Clean with a microfiber cloth and Acetone in order to remove adsorbed water from surfaces
• Clean with another microfiber cloth and 99% Isopropyl Alcohol to remove any Acetone residue

After the alcohol evaporates, the surfaces are mostly oil and water free. Every day between the cleaning steps and the final assembly of the chamber, atmospheric water vapor will continue to accumulate on the parts. Once the chamber is together and pulled down to vacuum, you shouldn’t expose it to the atmosphere unless it is necessary to open it up.

Assembling the chamber with the Conflat system is much more difficult than with the KF system. The copper gaskets arrive in individual vacuum sealed packets and when placing the seal on the flange, it is important to never actually touch the seal, even with gloves. Therefore, you squeeze the seal out of the package and directly onto the flange, then place the two flanges together and insert the bolts. This procedure nearly always requires three hands and becomes especially difficult when the flanges are vertical, causing the gasket to want to fall out of its place.

After the bolts are hand tightened, they need to be progressively tightened in the specific star pattern applicable for the size of the flange. The gasket is copper, so it can squeeze out on one side if not tightened evenly. These bolts have a torque specification for each flange size and are normally silver plated to prevent galling.

Vacuum Pumps

Vacuum pumps are used in almost all industries and even some arts and crafts. The HVAC industry is where small vacuum pumps became widely available even on Amazon. These are called rotary-vane pumps, where movable vanes ride along an eccentric wall to pump out air. These pumps can be found cheap, but they cannot reach the pressures needed for fusion. My rotary-vane pump can reach a pressure of 2×10^-2 Torr, which makes a great backing pump (also called a roughing pump) for a high vacuum pump.

There are several different type of high vacuum pump, but two are most common for amateur fusors: the diffusion pump and the turbomolecular pump. The diffusion pump uses the boiling of oil to force air molecules out of the chamber. It has no moving parts, so it is robust, but it can track oil vapors into the chamber. The turbomolecular pump is designed similarly to the compressor section of a jet engine, with turbine blades spinning at very high speed. This option starts up significantly faster and does not introduce oil into the vacuum chamber. With both options, it is incredibly important to not expose them to atmosphere after starting up. The boiling oil of the diffusion pump can ignite, and the sudden increase in pressure can rip the blades off of the turbo pump and turn it into scrap. For this reason, I designed in redundant valves and strictly kept to a written procedure. This was important enough that I printed out a valve diagram and labeled each one so I wouldn’t get them mixed up.

I happened across a good deal on a Varian Turbo V-250 turbomolecular pump that I refurbished and used on my system. It spins at 58,000 rpm and is drastically oversized for my chamber, but it works wonders. I am planning on adding an Ion Pump as well as a Titanium Sublimation Pump in the near future to experiment with lowering my ultimate pressure.

Pressure Measurement

I picked up an old set of Instrutech vacuum gauges and associated controller near the beginning of the chamber design. Two of these gauges are Instrutech “Worker Bee” Convection Gauges, which bottom out at 1×10^-4 Torr. These work well as gauges for the roughing side of the vacuum system. The main chamber gauge is a Instrutech CCM500 “Hornet” Cold Cathode Ion Gauge. This gauge works from 1×10^-2 to 1×10^-9 Torr.

Chamber Conditioning

When I first assembled the chamber, my ultimate pressure was 9×10^-7 Torr. When I sealed off the chamber it rose to 8×10^-5 Torr after an hour. This is perfectly adequate to do fusion, but I decided to take on a small side project to improve the quality of my vacuum.

The rise of pressure in the chamber is due to oil and water adsorbed to the surface of the walls, which quickly vaporize at these low pressures. One way to remove these contaminants from the system is to perform what is known as a bakeout. By heating the walls of the chamber while the vacuum pumps are running, the oil and water are driven off and removed from the system. I purchased silicon heating tape for this purpose and wrapped it around the chamber. I encased the heat tape and the chamber in insulation and aluminum foil to minimize heat losses to the atmosphere. Over the span of three hours, I ran the pumps and heated the chamber and monitored the temperature of sensitive items like the viewport and the ion gauge. The chamber only reached 60 Celsius, but it still had great results.

After this bakeout my ultimate pressure dropped to 3×10^-7 Torr, and only rose to 3×10^-6 after one hour. I am very happy with these results and I plan on doing an improved bakeout process soon.