No, this is not a commentary on politics. This post concerns vacuum.
Recently I was invited to a fellow glassblower’s shop. He is an accomplished glassblower who has become interested in doing plasma sculpture. Plasma artwork shares many of the same disciplines required of those who produce neon signs and artwork….most notably the glassworking, vacuum, and electrical/electronic knowledge. In his pursuit of this goal he purchased a vacuum pump off of eBay at a bargain. A Welch 1402, which is an excellent belt driven rotary vane pump capable of achieving very high vacuum levels. The factory specs call for about .01 micron. Sea level atmosphere is 760mm of mercury. A micron is a thousandth of a millimeter. So….yes, a hundredth of a micron is very high vacuum. In practice, when fitted to a manifold and associated hardware, the maximum vacuum is never quite that low, but it can be very close, or even better if this rotary pump is used as a rough pump to back up a diffusion pump or turbomolecular type. The higher the vacuum, the better…the goal here is to evacuate as much as possible.
The main purpose of my trip was to take a look at the bargain 1402 and verify that it was working. The pump was not connected to any manifold or other equipment as he has not yet purchased these items. I brought along a thermocouple type vacuum gauge, but it was one that had been obtained for free from a junk box and I was suspicious of its validity.
Lacking any known good instrumentation, we performed a quick check using a short piece of red rubber high vacuum hose and a piece of glass tube with the end sealed shut in a form that resembled a common test tube.
As a general rule of thumb, if you evacuate a glass tube and hit it with a spark coil tester (basically a 50-75kv hand held Tesla coil) you will get a differing color and type of glow in the tube that is dependent upon the degree of vacuum. It usually works out like this:
Initially his pump was not showing good results. I ended up leaving a bit later and tried the same test tube and hose on my pump just to verify that the test apparatus we hastily made was not at fault. I got a good pale blue haze on the wall of the tube, indicating somewhere in the ballpark of 5 microns. Not bad for a cold started pump. I called my friend and let him know….he had, it seemed, already resigned himself to an overhaul and ordered the parts kit. However, being the optimist….or maybe just stubborn….no, let’s call it thorough…that’s better. I suggested he try again but make certain that he had a very good seal between the hose and pump’s barbed connection. Use vacuum grease or a clamp….or preferably both. He did so and found that he would not be needing that rebuild afterall.
All of this seemed an opportunity to give some thought to nothing. Vacuum. Not really nothing but more or less the absence of something…..or at least the presence of less something than the surrounding area……typically a gas, but sometimes equally applicable to the inside of my wallet.
Nature abhors a vacuum, as they say. The tiniest of infinitesimally little openings will eventually allow atmosphere to enter. (sadly my empty wallet does not magically suck cash from the surrounding atmosphere) Sometimes even the materials used in a vacuum system will work to degrade the vacuum–not by leakage, but by outgassing. Everything, no matter how smooth, has some degree of porosity…in these pores resides millions of molecules of air, just waiting to be released when the ambient pressure is lowered. Sometimes it is not the pores, but the material itself. Items with too high a vapor pressure are a definite no-no in vacuum applications.
Another recent discussion in an online neon forum concerned pump down speed. The person who posted noted that when he was processing two tubes connected by a T section to his manifold, the pump-down time was much shorter than when he only processed one single unit.
The initial answer I offered him is a bit of truth in vacuum system design but it is totally counter intuitive. The plumbing was to blame. Having one unit, joined by way of it’s 5 or 6mm diameter tubulation at perhaps a foot long, presented a much slower flow than two full units having only a few inches of that tubulation (due to pump table layout) and then joining a larger diameter T section to reach the manifold. The total volume of the tubes was doubled. The T section, he said, added even more volume in the form of an extra foot or two of (12 or 15mm diameter glass)……yet in spite of this added volume it pumped down quicker. Why? Not in spite of the added volume, but because of it. That is counter intuitive. But here’s why….. Up to a point, a larger diameter pipe will always evacuate faster. It presents less resistance to the flow of molecules of gas. Even at very low pressure, it matters. Bigger pipe equals more room for the molecules to pass each other on their way out and with less opportunity to drag along the edges of the pipe. This is for the most part true up to the diameter of your vacuum pump’s inlet connection. The replacement of the single thin section with the larger T shaped one for most of the distance covered could easily account for this speed difference. Back in the day, we often built small manifolds….with small diameters–partly this was due to material availability (we usually built manifolds in our shops using the same sorts of tubing found in the signs we created, this often meant a maximum of 15mm or so) but it was also due to the fact that most of us were not physicists or vacuum system engineers. As an added factor, the noble gases were supplied in 1 liter flasks at atmospheric pressure….a large manifold volume would mean using up the flasks quicker. Today, many obtain gases in pressurized cylinders that will last a long time before being emptied, and with the advent of large bore pyrex manifolds, it is not uncommon to see main headers of 20 to 25mm. The pumping speeds are markedly faster as a result.
But there was one other piece of this puzzle that I had initially missed. While the information I gave him was undoubtedly true….it was not until he clarified something else that I knew another factor that likely was the primary cause of his observation… It was the way he measured his vacuum. He had no gauge on the system. The method he was using is the same method that I have used in the past and that almost every oldtimer out there used: He flashed the bombarder until the tubes would no longer light. This does work. To a point. When the units are evacuated past a certain point they will in fact no longer light, even when hit with 15-20kv from a bombarding transformer. There are just two problems with this methodology…..well, likely more than two…but two that leap to mind first:
1.) It is hard on the bombarder. The loss of a current path does a couple of things-it can break down the insulation in the transformer windings causing an eventual failure….and it also can cause a flashback or other hazard as all that high voltage seeks another place to go–hopefully not into the operator!
2.) It is not a totally reliable and repeatable measuring technique. The voltage required to light and sustain a discharge in the lamp is a function of the length, diameter, and pressure in the tube. By adding that second unit, connected in series, the operator had effectively doubled the length of glass to be lit. This equates to needing more voltage applied. IF the only variable was the pressure and all other factors between the single unit and the two unit configuration were the same….it becomes obvious that the two unit pair would extinguish at a higher pressure than the single unit! This means that his pumping system would require less time to pump-down to the point that the lamps would no longer illuminate when flash tested!
So there are two good reasons to add a gauge to your system. I also added a horn gap to my bombarder. This is essentially like a small Jacob’s ladder spark gap. If I flash test a tube (which admittedly I still do sometimes, old habits die hard) The horn gap provides a place for the current to flow, reducing the hazard and wear and tear on the bombarding transformer. In addition, having this gap can provide a safety in the even of an unexpected tube failure (such as cracking) during processing as it will allow that power to flow when the failed unit presents an open circuit.
Now, this is not to say everyone doing this work has to go get a gauge. Let’s face it, a large majority of good quality neon units were produced prior to affordable gauges being available. The flash method can work and become somewhat repeatable if other factors are taken into account such as allowing the units to cool down to a consistent temperature after processing and before filling. The lower temp will result in lower pressure. You still won’t know the exact absolute value…..but if you know your system and know it well, you will have a good pretty close idea if it is right or not. Even so, a good reliable gauge can add that extra measure of consistency to the operation. And in some applications (plasma, large diameter cold cathode lighting, and making electron tubes) it is an absolute necessity to verify proper processing and to insure consistent performance of the end product. Seeing that something sucks is fine….but it is important to know how much it sucks.
Glad I’ve finally found something I agree with!