Q1 TANK/BRACKET

I wanted the clock to run in a fairly high vacuum and decided to have a flanged glass tank made with a glass dome at the top. The tank was fabricated to my specifications here in New Jersey by a firm that specializes in laboratory glassware. The tank is made of Pyrex glass with two welded glass flanges, top and bottom.They also made the dome that covers the movement. The tank is closed at the bottom by a circular piece of glass that is ¾ inch thick. I had two 1/2 inch holes drilled through the glass. In one of the holes, I mounted a fixture holding a vacuum valve and a thermocouple vacuum sensor. In the other hole, I mounted a vacuum feedthrough that would accommodate the dozen wires I needed to run the clock. I ended up having serious problems maintaining a vacuum inside the tank. Both the dome and bottom glass were fitted with Buna-N rubber seals specifically made for this type of project. The seals were coated with Dow Corning vacuum grease before closing up the tank. When I initially pumped the tank down, it began to lose vacuum immediately and for weeks I looked for leaks.   I thought perhaps the fixtures I made mounted to the glass had problems and I remade them. It didn’t help. I changed the seals around the glass dome at the top and that didn’t help either. I tried various sealing compounds, all useless. I finally went on the web

Looking for advice and came across a group of vacuum professionals on a commercial web site that seemed happy to answer questions. I posted my question and one of the guys offered to send me new vacuum wire seals which he thought might be the problem. I gave him my address and he responded that his shop was quite close to me and offered to come by and help solve the problem. He stopped in about a week later (in a Rolls!) and he didn’t think the seals were a problem. He felt the glass tank itself was leaching water vapor under vacuum and that I probably didn’t have a “real”leak at all, but a “virtual” leak. His suggestion was to heat the tank and its contents to about 120-140 degrees for a couple of days while running it under a high vacuum. This would “cure” the glass and would probably stop the “virtual” leak. I went to the hardware store and bought several lengths of roof heating cable. I plugged the cables into a thermister sensing heat controller and wrapped it around the tank and dome. I brought the heat up to 140 degrees

The tank in the process of being heated in order to outgas the components. It's wrapped with heating cable connected to a temperature controller set to 140 deg. F. I ran it and the vacuum pump for two days  

up to 140 degrees

The tank in the process of being heated in order to outgas the components. It's wrapped with heating cable connected to a temperature controller set to 140 deg. F. I ran it and the vacuum pump for two days  

up to 140 degrees and pumped the tank down as far as I

and pumped the tank down as far as I could and ran it for two days. It worked! Interestingly, while I was doing this, I had a chart recorder connected to the vacuum gauge. As the tank heated up, I could see it losing vacuum at an increasing rate so that the vacuum pump could not keep up with it in the beginning. After about 14 hours or so, the vacuum in tank started to increase and eventually I could see the “leak” stop. Every component in the tank will outgass and the process of eliminating these virtual leaks is absolutely necessary.

The 3/4 inch thick bottom closure showing the vacuum valve and vacuum sensor on the right and the plug and wire feedthrough's on the left

Custom made Q1 tank

The tank in the process of being heated in order to outgas the components. It's wrapped with heating cable connected to a temperature controller set to 140 deg. F. I ran it and the vacuum pump for two days

THE TANK BRACKET

The bracket was fabricated from aluminum and is designed so that the clock can be leveled in all directions. Three fine adjustment screws are fitted to the floating tank support. This .

in turn rests on a substantial fixed support mounted on a steel stanchion. The steel stanchion is fixed to a 2 ton concrete block mounted in a concrete well in the floor. The block "floats" on a set of vibration isolators. I did this so that the noise generated in the shop as well as extraneous outside noise from the road would only minimally affect the clock.