Liquid HV Feedthrough

For posts specifically relating to fusor design, construction, and operation.
Niels Geerits
Posts: 56
Joined: Tue Jun 13, 2017 10:25 pm
Real name: Niels Geerits

Liquid HV Feedthrough

Post by Niels Geerits » Sat Sep 02, 2017 8:19 pm

I have been thinking about a way to make my own HV feedthrough given that most commercial feedthroughs are quite expensive. I came up with a liquid feedthrough: instead of using a difficult to produce air tight ceramic insulator one simply uses a vertical SS pipe filled with a non conducting low vapor pressure liquid. So which liquids can be used (answer is in the 3rd paragraph)? My first idea was to use the mineral oil I use for my HV supply. At room temperature mineral oil will vaporize at a few microns. However if you cool the SS pipe (e.g. using a peltier element) you could reduce the vapor pressure.

Coconut oil is another liquid I looked at. It is an excellent insulator and will actually become solid slightly below room temperature. The problem is that I cannot find any Information on its vapor pressure. All the saftey sheets list N/A under vapor pressure and if I test it myself I might ruin a good diff or turbo pump by getting that gunk in there. So if you happen to have vapor pressure curves for coconut oil I would appreciate it.

It turns out ultragrade 19 vacuum pump oil is perfect for this purpose. Its vapor pressure is around 10^-8-10^-7 Torr and as an oil it cannot be a very good conductor. In addition one can freeze the oil and decrease the vapor pressure using multi stage peltier elements. I will be doing tests with this oil. I do not know how easily air can diffuse through the oil, so freezing it may help.

With diffusion pump oil one could gain one or two orders of magnitude in vapor pressure at the cost of the oil costing a few orders of magnitude more.

What follows is a paint drawing of the feedthrough:
liquid feedthrough.jpg

ian_krase
Posts: 447
Joined: Mon Nov 28, 2016 7:48 am
Real name: Ian Krase

Re: Liquid HV Feedthrough

Post by ian_krase » Sat Sep 02, 2017 9:10 pm

A common method is a cheap feedthrough that is inadequate, surrounded by oil. On the outside. With this method the oil need not care about vapor pressure or being sucked into the chamber.


Also you would need a very tall column to keep the oil from being sucked into the chamber - even oil as dense as mercury would need a meter tall column.

MatthewL
Posts: 48
Joined: Mon Mar 06, 2017 12:13 am
Real name: Matthew Lallinger
Location: Houston, Tx

Re: Liquid HV Feedthrough

Post by MatthewL » Sat Sep 02, 2017 10:47 pm

To make a high voltage feedthrough I think that the best option would be to use a compression port and a glass or ceramic tube similar to the one that Doug Coulter makes here: http://www.coultersmithing.com/AuxCP/FT.html.
As Ian suggested you would need a very tall column to keep the oil out of the chamber. If you try, coconut oil probably isn't the best option as, depending on you source, it will probably contain impurities with a higher vapor pressure than the fatty acids that make up the oil.

Niels Geerits
Posts: 56
Joined: Tue Jun 13, 2017 10:25 pm
Real name: Niels Geerits

Re: Liquid HV Feedthrough

Post by Niels Geerits » Sun Sep 03, 2017 12:31 am

Ian I thought of using a lower rated feedthrough as well, but thought a fully liquid feedthrough would be even simpler.

Matthew: I think that Dougs design requires a laythe I read it a while back. Sadly I don't have one.

Removed the math on the column because it was wrong and may therefore be misleading. Originally I thought the Ratio of the area exposed to the vacuum to the area exposed to the air would matter (again due to wrong math). Here a correction:

The pressure at the top of the column is 0. The pressure at height x measured from the oil surface in the Container equals: P=P_atm-rho*g*x. Solve for P=0 --> x=1bar/860kg/m^3/9.81m/s^2=12m Therefore using an open oil Container is not feasible.
Last edited by Niels Geerits on Mon Sep 04, 2017 11:48 pm, edited 2 times in total.

John Futter
Posts: 1291
Joined: Thu Apr 22, 2004 2:29 am
Real name:
Contact:

Re: Liquid HV Feedthrough

Post by John Futter » Sun Sep 03, 2017 1:27 am

niels
Wrong the ratio of inside vacuum to outside has no bearing on the problem

try it out and see

Rex Allers
Posts: 236
Joined: Sun Dec 30, 2012 8:39 am
Real name:
Location: San Jose CA

Re: Liquid HV Feedthrough

Post by Rex Allers » Sun Sep 03, 2017 3:28 am

If I understand what John was saying, I think he misspoke. The inside pressure (vacuum) vs. outside pressure certainly does matter, otherwise a mercury barometer wouldn't work.

But Niels seems to think that the ratio of tube area (if different) matters on inside vs outside. That's not true. Pressure is force/area but the pressure in the liquid is what pushes up the column of liquid. As the area increases the force increases, but so does the mass of the liquid as its volume increases proportionally. The area cancels out; only pressure (force / unit area) matters.

Think of it this way, in your diagram the oil is contained on the bottom by a dish. If what you are thinking about area was true, the diameter of the dish would have a strong effect. It doesn't. Is the dish size a thing you need to control when making a mercury barometer?

Only the relative density of the liquids should matter. I saw 0.87 g/ml for 19 oil and 13.7 g/cc for mercury. A cc essentially is a ml so the ratio is ~15.7 times 760 mm for mercury = about 12 m, as you said.

That should kill the idea but another factor I don't think you addressed is the dielectric factor for the oil. What's the length of the shortest path that the oil will sustain for the voltage you plan to use. It doesn't matter because of your failed column length calculation, but would be needed if you really could be building this.
Rex Allers

John Futter
Posts: 1291
Joined: Thu Apr 22, 2004 2:29 am
Real name:
Contact:

Re: Liquid HV Feedthrough

Post by John Futter » Sun Sep 03, 2017 3:34 am

Yes Rex
I wasn't trying to be so obvious

Niels Geerits
Posts: 56
Joined: Tue Jun 13, 2017 10:25 pm
Real name: Niels Geerits

Re: Liquid HV Feedthrough

Post by Niels Geerits » Sun Sep 03, 2017 2:43 pm

Yep, if you just do a pressure calculation with the appropriate boundary conditions and assert the pressure must be continuous, the area drops out and you get a 12m column regardless. As a result changing the shape of the feedthrough does not help either.

I would not say the idea is dead. If you epoxied the container to the SS pipe creating an air tight container for the oil you would have less expansion. If there is no air in the container at all there will be no expansion. If there is a little bit of air the oil will be pushed into the chamber a bit. How much can be calculated by calculating the pressure of the air and using the ideal gas law to calculate the resulting expansion.

The air will go from 1 bar to rho*g*h (where h is the height of the oil in the feedthrough relative to the height of the oil in the container when a vacuum is pulled). The ideal gas law: P*V=C --> P_1*V_1=P_2*V_2. We know that V2 must equal A*h where A is the area of the feedthrough: P1*V1=rho*g*A*h^2 --> h=sqrt(P1*V1/rho/g/A), so we can control the height of the oil by increasing the area of the feedthrough and decreasing the volume of air trapped in the top of the container.

Let us look at some realistic values: A=19.63 cm^2 and V1=18.9 cm^3 (5cm diameter feedthrough with a cylindrical container +1cm radius and 1cm air column) --> h=34cm. Which is something that can be worked with. The volume of the air will probably be much smaller. You do want a small air gap to maintain some pressure which will prevent the epoxy seal from outgassing too much.

Rex I plan on measuring the dielectric strength of the oil since I cannot find any data on it. However most oils are excellent insulators so I do not expect that to be a problem. to increase the dielectric strength of the feedthrough one could add a few glass tubes to the oil.

User avatar
Dennis P Brown
Posts: 1488
Joined: Sun May 20, 2012 2:46 pm
Real name: Dennis P Brown
Location: Glen Arm, MD

Re: Liquid HV Feedthrough

Post by Dennis P Brown » Wed Sep 06, 2017 12:56 pm

Not knowing your design requirements nor available room, not easy to say too much but making a high voltage and vacuum feed thru cheaply is not difficult. I simply use a glass plate (fairly inexpensive depending on diameter - I have provided a source in other posts), drill a hole thru it (core drills are cheap and I use a hand drill), get a ceramic tube (again, not very expensive), a solid rod of brass (with an outer diameter a few thousands less than the inner diameter of the ceramic tube) and using epoxy, assemble all this into a very good HV feed-thru. If you are interested, I can reference my past posts.

I do not think a liquid feed-thru is a good path (for the reasons that Rex correctly points out. Also, having a source of oil vapor in a fusor isn't necessarily a good idea but that is another matter. Another and possible show stopper is water vapor; any oil exposed to air will pick up water; for a fusor that could be a very bad problem - that, will affect not just the fusor but also the dielectric properties of the oil.)

However, if you are determined to try to use a new method realize this could be a major project in of itself - fusors are difficult enough without adding complexity. That said, if you are just looking for a side project, and a fusor is a secondary goal, then this might be a fun area to pursue; in which case, your best approach is to try various designs and experiment to get a feel for what has the best chance to work.

Niels Geerits
Posts: 56
Joined: Tue Jun 13, 2017 10:25 pm
Real name: Niels Geerits

Re: Liquid HV Feedthrough

Post by Niels Geerits » Wed Sep 06, 2017 4:17 pm

Dennis, I plan on confining the fusor and its components to a 2mx2mx3m space if not less. I use my homemade 40kV supply.

I would like to see your posts about feedthroughs. What has me wondering though: I thought that epoxy is not a good vacuum material, which is why in my design idea the epoxy is on the air side.

As for the water vapor in the oil. This will only be a problem during the first few runs I think. Since the oil is sealed of from the outside air (in design 2.0) and the vacuum will remove water over time. However the oil vapor is concerning and is why I am looking for ultra low vapor pressure oils. Ultragrade 19 is still winning when it comes to affordable oils. Eurovacuum makes a parrafin based oil (great insulator) but lists two different vapor pressures (10^-9 torr and 10^-5 torr).

If you reference your post I will check it out. If it is easier than this I will probably go with your feedthrough and leave this for a future project.

Post Reply