Re: Helium-3 Fusor Thread
Posted: Thu Mar 22, 2018 9:45 pm
Jackson,
Some very important things to note about these extra high voltages. Once you start getting up into the +100KV range, things get much, much more challenging to deal with. Working with a fusor at 50kv or lower is very different from 100kv, and don't assume that once you have worked with one level the other level will be the same. At these voltages, electricity can act quite bizarrely and counter-intuitively. Surface tracking for one becomes a major issue. High voltage at these levels will no longer necessarily take the shortest path between two points, and can track very long distances along a surface. Another major issue is corona losses. At these voltages, you will have significant corona discharge spraying from any metal surface that is not smooth or has a very large, even diameter. Any sharp edges drastically increases field enhancement effects, making these voltages even more difficult to deal with. Also, depending on the amount of corona losses you have, your load will suffer from fluctuations and instabilities. Large corona losses will also load down your power supply.
The injector of our facility's electron linac is biased anywhere from -125kv to -150kv. The whole gun and all of the controls, monitoring, pulsing, and drivers also float at this voltage on an isolated deck on top of a large isolation transformer. The entire deck, along with the injector has to be contained in a specially designed corona shield, which has very large diameter corners and smooth surfaces. The whole thing sits inside a large metal shack with several dehumidifiers and a temperature control unit to control ambient temperature and humidity. Even at the long distances, between the corona shield and overall enclosure, depending on the humidity, it can still arc occasionally. Especially on very humid days, in our second test injector, we can observe very large corona losses and instabilities at the higher end over 125kv. Also, current at these levels have a major impact, especially on corona losses, surface tracking, and arcing. We run our injector at probably less than a mA normally, but with humidity it can draw several mA from the supply. Turning up the current even just a couple of mA can introduce very large corona losses, and with it, system instabilities and arcing. Running a power supply for this type of fusor not only at +100kv voltages, but currents in the tens of mA range is incredibly difficult, and dangerous to control.
From the x-ray standpoint, our entire test bunker has several layers of almost 2 feet thick of special iron-weighted concrete blocks, stacked to prevent direct line-of-sight between seams of adjacent bricks to minimize radiation leakage (the actual linac injector, not the test stand, is inside a massively shielded room with many feet of shielding, but this is mainly for the neutron shielding.) Everything is controlled remotely from outside of the controller. It is not a small setup, and is much more complex than running a typical fusor on a workbench. At these voltage levels, you will be looking at a similar setup for the fusor.
If your insulator is only rated for 25kv, you may get away with running it at 50kv, as you have without issue. Over 100kv, this will not be the case. At these voltages, if the insulator is not long enough, the arc can very easily track along the surface (the reason why these insulators have wavy structures is to increase the surface distance between two points to reduce surface tracking possibility as opposed to a straight insulator.) At voltages and current you will be looking at, it is also very difficult to insulate, and can punch through even seemingly thick insulation with ease, especially if there are any gaps or defects. Even applying insulation yourself over a connector or between two points requires very special step gradations with various insulators and thicknesses to deal with and manage field potential distributions and gradients. The arc can snake in between cracks, seams, and joints. Something known as the "triple junction effect" also becomes an issue, where localized geometric field enhancement occurs when insulators of two different permittivities (such as air and ceramic) meet at an electrode, which can initiate discharges that can propagate along a surface. Add high currents to this and it becomes a very formidable challenge.
You also have to consider your vacuum system at this point too - all of the points above are only for the outside of the system, we haven't even gotten to issues actually in the system. Pressure and breakdown voltage are related - look up the Paschen curve. Under certain conditions, an arc can travel extremely far at even very low voltages. You also have to consider all inside surfaces that they are flawlessly smooth, especially your grid - any tiny sharp edge not perfectly ground down and smooth will cause immediate issues. You also have to deal with the effects of water vapor and surface contamination in your system - to run a fusor like this at these levels, proper conditioning is an absolute must. It takes us many, many hours of conditioning at high voltages before we can even turn on our electron beam - if it is not conditioned well enough at ultra-high vacuum levels, it will arc over internally when the beam is on.
For your idea on lining the inside with ceramic, this may not actually prevent arcing as you expect. Charged particles bombarding insulator surfaces in vacuum systems will build up a charge on the insulator, and can easily cause it to arc and flash over. This is very observable in systems such as high power RF windows as well as insulating structures in beam systems. RF windows actually have to be coated with a special metal layer to prevent charge buildup and secondary field emissions which in turn cause breakdown across the ceramic. In fact, one of the major issues for DC accelerators with a stacked ring-insulator topology, is that the beam cannot be allowed to "see" the insulators. If it does, charge will build up, causing arc over. Therefore, special shaped rings are placed internally to guard over the insulators to prevent charge from directly depositing on them. It's also a phenomenon in sputtering systems: one of the reasons why you cannot sputter insulators with DC is that charge will build up on the insulator surface from ion bombardment from the plasma. RF is needed to prevent this and sputter these materials correctly. Just adding ceramic insulators will not necessarily prevent arcing in your system, and depends how and where they are set up. If all of the walls are lined, it may also cause issues with actually running the fusor itself, and could lead to unforeseen instabilities.
Again, a fusor, or any high voltage system, is very different at tens of kv vs. over 100kv, and there are many things to consider, both inside and outside the system. It will become a much more serious engineering effort and require a lot more planning, research, and money. It is do-able, but you are entering the realm outside of what can be done on a desk at home.
Some very important things to note about these extra high voltages. Once you start getting up into the +100KV range, things get much, much more challenging to deal with. Working with a fusor at 50kv or lower is very different from 100kv, and don't assume that once you have worked with one level the other level will be the same. At these voltages, electricity can act quite bizarrely and counter-intuitively. Surface tracking for one becomes a major issue. High voltage at these levels will no longer necessarily take the shortest path between two points, and can track very long distances along a surface. Another major issue is corona losses. At these voltages, you will have significant corona discharge spraying from any metal surface that is not smooth or has a very large, even diameter. Any sharp edges drastically increases field enhancement effects, making these voltages even more difficult to deal with. Also, depending on the amount of corona losses you have, your load will suffer from fluctuations and instabilities. Large corona losses will also load down your power supply.
The injector of our facility's electron linac is biased anywhere from -125kv to -150kv. The whole gun and all of the controls, monitoring, pulsing, and drivers also float at this voltage on an isolated deck on top of a large isolation transformer. The entire deck, along with the injector has to be contained in a specially designed corona shield, which has very large diameter corners and smooth surfaces. The whole thing sits inside a large metal shack with several dehumidifiers and a temperature control unit to control ambient temperature and humidity. Even at the long distances, between the corona shield and overall enclosure, depending on the humidity, it can still arc occasionally. Especially on very humid days, in our second test injector, we can observe very large corona losses and instabilities at the higher end over 125kv. Also, current at these levels have a major impact, especially on corona losses, surface tracking, and arcing. We run our injector at probably less than a mA normally, but with humidity it can draw several mA from the supply. Turning up the current even just a couple of mA can introduce very large corona losses, and with it, system instabilities and arcing. Running a power supply for this type of fusor not only at +100kv voltages, but currents in the tens of mA range is incredibly difficult, and dangerous to control.
From the x-ray standpoint, our entire test bunker has several layers of almost 2 feet thick of special iron-weighted concrete blocks, stacked to prevent direct line-of-sight between seams of adjacent bricks to minimize radiation leakage (the actual linac injector, not the test stand, is inside a massively shielded room with many feet of shielding, but this is mainly for the neutron shielding.) Everything is controlled remotely from outside of the controller. It is not a small setup, and is much more complex than running a typical fusor on a workbench. At these voltage levels, you will be looking at a similar setup for the fusor.
If your insulator is only rated for 25kv, you may get away with running it at 50kv, as you have without issue. Over 100kv, this will not be the case. At these voltages, if the insulator is not long enough, the arc can very easily track along the surface (the reason why these insulators have wavy structures is to increase the surface distance between two points to reduce surface tracking possibility as opposed to a straight insulator.) At voltages and current you will be looking at, it is also very difficult to insulate, and can punch through even seemingly thick insulation with ease, especially if there are any gaps or defects. Even applying insulation yourself over a connector or between two points requires very special step gradations with various insulators and thicknesses to deal with and manage field potential distributions and gradients. The arc can snake in between cracks, seams, and joints. Something known as the "triple junction effect" also becomes an issue, where localized geometric field enhancement occurs when insulators of two different permittivities (such as air and ceramic) meet at an electrode, which can initiate discharges that can propagate along a surface. Add high currents to this and it becomes a very formidable challenge.
You also have to consider your vacuum system at this point too - all of the points above are only for the outside of the system, we haven't even gotten to issues actually in the system. Pressure and breakdown voltage are related - look up the Paschen curve. Under certain conditions, an arc can travel extremely far at even very low voltages. You also have to consider all inside surfaces that they are flawlessly smooth, especially your grid - any tiny sharp edge not perfectly ground down and smooth will cause immediate issues. You also have to deal with the effects of water vapor and surface contamination in your system - to run a fusor like this at these levels, proper conditioning is an absolute must. It takes us many, many hours of conditioning at high voltages before we can even turn on our electron beam - if it is not conditioned well enough at ultra-high vacuum levels, it will arc over internally when the beam is on.
For your idea on lining the inside with ceramic, this may not actually prevent arcing as you expect. Charged particles bombarding insulator surfaces in vacuum systems will build up a charge on the insulator, and can easily cause it to arc and flash over. This is very observable in systems such as high power RF windows as well as insulating structures in beam systems. RF windows actually have to be coated with a special metal layer to prevent charge buildup and secondary field emissions which in turn cause breakdown across the ceramic. In fact, one of the major issues for DC accelerators with a stacked ring-insulator topology, is that the beam cannot be allowed to "see" the insulators. If it does, charge will build up, causing arc over. Therefore, special shaped rings are placed internally to guard over the insulators to prevent charge from directly depositing on them. It's also a phenomenon in sputtering systems: one of the reasons why you cannot sputter insulators with DC is that charge will build up on the insulator surface from ion bombardment from the plasma. RF is needed to prevent this and sputter these materials correctly. Just adding ceramic insulators will not necessarily prevent arcing in your system, and depends how and where they are set up. If all of the walls are lined, it may also cause issues with actually running the fusor itself, and could lead to unforeseen instabilities.
Again, a fusor, or any high voltage system, is very different at tens of kv vs. over 100kv, and there are many things to consider, both inside and outside the system. It will become a much more serious engineering effort and require a lot more planning, research, and money. It is do-able, but you are entering the realm outside of what can be done on a desk at home.