By Howard Tring

Introduction

This article talks only about one and two stage “medium vacuum” oil sealed rotary vane vacuum pumps that can produce a catalog ultimate vacuum of about 1 x 10^-2 Torr (0.01 Torr or 10 microns) for a one stage model and about 1 x 10^-3 Torr (0.001 Torr or 1 micron) for a two stage model.

Oil sealed rotary vane vacuum pumps are used in the vacuum heat treating and vacuum furnace industry as a holding pump at the exhaust side of the oil diffusion pump. It keeps the exhaust line pressure low enough to prevent stalling of the oil diffusion pump while the larger mechanical pump is roughing (evacuating from atmospheric pressure) the main vacuum chamber.

In the early days, pre-second world war, there were also oil sealed rotary cam pumps designs, such as the “world famous Cenco ‘Hyvac’ two stage vacuum pump (Fig. 1) and the Nelson Pump Co. ‘Nevaco’ vacuum pump (a). Although the Hyvac pumps and other Cenco models are still manufactured today by HyVac Products, PA, the Nelson Pump Company was bought by Ace Pump, TN, and their vacuum pumps are no longer made.

Fig. 1. Cenco “HyVac” vacuum pump.

There are other oil lubricated rotary vane vacuum pumps made, usually by different manufacturers, that are not sealed (sometimes referred to as “flooded”) with oil. These vacuum pumps allow limited oil into the pump for lubrication and generally produce a best vacuum of around 0.5 Torr. This is not a low enough pressure for many industrial and scientific applications, especially those using an oil diffusion pump, a turbomolecular pump or a cryogenic pump for high vacuum pumping..

Early Rotary Vane Oil Sealed Vacuum Pumps (vee belt drive)

Rotary vane oil sealed vacuum pumps (RVOSVPs) were originally vee belt driven. The pump was mounted on a baseplate with an electric motor on a slide base installed next to it. The pump was driven through vee belts on pulleys mounted to the pump and motor. Pulleys of different sizes allowed the rotating speed of the pump to be set, usually around the 500 to 600 rpm area, about a third of the electric motor speed. The slide base allowed the vee belt tension to be adjusted. Later on, as safety standards were introduced, the vee belts and pulleys had to be covered by a metal cover.

Due to the relatively low rotational speed of the pump rotor, the blades or vanes were generally made of steel in the hope that centripetal force would throw them outwards to contact the inside of the stator and make a seal. To make sure this happened, holes were drilled on the inner edge of the blades and springs were inserted. Often steel pins were added inside the spring to reduce the flexing and prevent buckling.

At Edwards they also used molded asbestos/resin blades and this made the springs and pins more important. The springs because the blades were lighter than the steel ones and require a positive force to ensure contact with the stator; and the pins because a flexing spring could wear through the side of a molded blade creating lots of damage.

There were many manufacturers of RVOSVPs in North America and in Europe and up until the 1950s most manufacturers sold their products generally in their own local markets. There were few European made vacuum pumps sold in North America and most likely, few North American made vacuum pumps sold in Europe. One of the reasons for this could have been the different electrical power supplies, i.e. 60 Hz 115 V in NA and 50 Hz 230 V in EU. That problem wasn’t too difficult to deal with when most of the vacuum pumps were driven by vee belts as the motor was a separate item. In fact Edwards in the UK sold larger US made Kinney rotary piston pumps before they designed their own large rotary piston pumps. Another reason, in my opinion, was the time and cost of travel between the two continents and extended shipping times by sea.

Fig. 2. Welch DuoSeal vacuum pump.

The most popular design of oil sealed rotary vane pumps became a modified version of the Gaede design from 1907 (Fig. 3). Welch Scientific introduced the Wegner Micro-Deka in 1934 which was an improvement on the Gaede design (b). This improvement was the oil seal between the gas inlet and exhaust openings in the stator. This area of the stator was machined at the same diameter as the rotor therefore producing a narrow band between the rotor and stator that would hold oil in it when the clearance was around 0.002 inches. Too little clearance would allow metal to metal contact, and too much clearance would allow the higher pressure gas on the exhaust side of the seal blow the oil out of the sealing area and allow gas leakage across the seal. This seal design is known as an arcuate seal. Welch trademarked the name “DuoSeal” and continues to use that trade name today for its vee belt driven oil sealed rotary vane vacuum pumps. (Fig. 2) Most other manufacturers also used this arcuate seal design as it has proved to be very robust.

With the clearance set at the correct amount the pump blade (or vane), which tends to sweep a small volume of oil in front of it at the tip area, adds this oil to the arcuate seal on the exhaust side and a similar amount of oil would be expelled from the seal area on the inlet side. This means that the oil in this seal area is constantly being changed. The oil being pushed into the arcuate seal causes a distinctive noise on some models of this pump design. Once the pump has run for a minute or so, on a small volume, and the “slap-slap” or “clack-clack” can be heard the pump should be giving a good vacuum (low pressure).

Pros and Cons of Vee Belt Drive vacuum pumps

For some vacuum pump users, especially in quiet laboratory settings, this “slap-slap” or “clack-clack” noise is objectionable and the sound can be reduced by cracking open the gas ballast valve a small amount and letting a small amount of air into the pump. Cracking open the gas ballast valve can cause additional oil mist to be generated from the exhaust of the pump requiring the use of an oil mist filter.

As I mentioned the gas ballast valve a short comment about it is necessary. On early oil sealed rotary vane vacuum pumps there was no gas ballast valve. Not until around 1953 on Edwards UK made vbd pumps (Fig. 4). In those days if the application caused water vapor to be drawn into the vacuum pump and it was enough to condense in the pump oil and to dilute the oil, its lubrication and sealing properties were adversely affected. If a water and oil mixture is left in the pump corrosion will occur and the life of the pump will be compromised.

Fig. 3. Gaede 1907 design.

The cure before there was a gas ballast valve was to add a vapor trap on the pump inlet. The trap I am familiar with, supplied by Edwards, was a multi tray trap containing phosphorous pentoxide (P₂O₅) that looks like a white powder. When water vapor reacts with the P₂O₅ it changes the surface powder to a wet material, phosphoric acid. The material in the trays could be stirred once or twice to bring dry powder to the surface before having to replace the material altogether. It was typically rinsed down the closest sink. This would not be allowed these days of course.

Once the gas ballast valve was invented it first became an optional accessory and then a permanent part of any pump of this type. As a refresher for you, allowing a bleed (ballast) of air (or inert gas in some applications) into the exhaust valve area of the pump allows water vapor to be expelled out of the pump as vapor before it is compressed enough to reach its vapor pressure and condense into a water droplet that will mix into the pump oil. Oil that appears to be an emulsion of milky white liquid with green streaks is the obvious sign of major oil contamination by water.

Most early vee belt drive pumps had an oil lubrication circuit that relied on oil from the “filled to the top” oil box draining through a small hole in the top of the pump by gravity through passageways to the bearings, rotor and stators. These small laboratory sized vacuum pumps do not have ball bearings supporting the rotor. The rotors are relatively light and the vacuum pump only creates very small radial and axial loads so that a simple metal to metal with an oil film bearing will support the rotor. The bearings are straight sections at each end of the rotor that match bored holes in the stator and have close tolerances to hold an oil film between the surfaces. This is the main reason that this design of vacuum pump needs to have clean oil in it at all times. Any contamination can affect the bearing surfaces and lead to pump failure.

A drawback of the simple gravity drain oil circuit is a problem called “suck back” of oil.

If the pump stops either deliberately or due to a power failure it leaves the pump inlet open to the system which is under vacuum. In that event suck back of oil can occur. Atmospheric pressure acting on the oil surface in the oil box slowly pushes oil down the gravity drain opening, into the pump mechanism and towards the inlet side of the pump which is at low pressure. If the amount of oil is large and the system is close to the pump inlet it is possible to have oil “sucked back” into the system causing contamination. Gravity drain lubrication requires lots of oil above the top of the pump oil inlet hole to make it work, hence the problem. An ideal pump set up is one where the vacuum pump is isolated from the vacuum system on shutdown and then the pump interior is allowed to go to atmospheric pressure. This minimizes the chance of oil draining into the pump mechanism.

This “suck back” problem requires vacuum pump users to install an automatic shut off valve on the vacuum pump inlet that to prevent any oil getting past it. One simple non-return valve, used by Edwards, was a lightweight ball inside a housing with an angled top, inside. If oil was drawn towards the system the ball would float until it sealed against the angled inside face. This was not 100% foolproof but worked quite well. A later option was a magnetic solenoid valve that would close on power failure to give a secure seal. A downside of any of these accessories was the space taken up on top of the pump. In many cases an equipment designer had limited space inside a framework to fit the vacuum pump. Adding these accessories took up another three or four inches of space.

Another problem caused by oil draining down into the open volume of the stator occurs when restarting the vacuum pump. When the vacuum pump is pumping gas the gas is compressed through the relatively small exhaust valves of the pump. Trying to expel oil through those same small holes is more difficult because the oil is not compressible.  On a vee belt drive pump it can cause the vee belts to slip on the pulleys and possibly burn out, or it can cause the electric motor to overload its thermal protection device or fuses and shut off.  

Fig. 4. Edwards vbd pumps.

The main positive feature of slow running vee belt drive rotary vane vacuum pumps is their ability to dilute contaminants in the large volume of oil and keep running despite the contamination. Their running temperature is generally lower than that of faster running direct drive vacuum pumps and they are a simple rugged machine. They are still popular in applications such as wet chemistry laboratories where the chance of ingesting a variety of chemical contaminants is high.

Vee belt driven vacuum pumps were labor intensive to assemble, with the baseplate, pump, motor, pulleys, vee belts and beltguard making it time consuming in the final assembly area. In the 50s and 60s the tremendous growth of industry and scientific research applications caused the vacuum pump manufacturers to rethink their designs. There was a push for compactness and a need to reduce manufacturing costs to stay competitive. The age of direct drive vacuum pumps was near.

US manufacturers of vee belt driven rotary vane oil sealed vacuum pumps.

There were a number of companies making these pumps from the 1920s up into the 1970s and 1980s and most have either been taken over by other companies or are out of business altogether. Here are some names that you may or may not have heard of:

• Arthur F. Smith, sold to All Starr Scientific (out of business)
• Central Scientific Co. (Cenco), sold to Boekel and again in 1990 to HyVac Products (still made)
• Marvac Scientific Mfg Co. (closed vacuum pump division in 2003)
• Precision Scientific, sold to GCA Corp and again in 1966 by Jouan (no longer made)
• Red Point Corp. discontinued vacuum pumps in in mid 1970s
• Robinair, now a division of SPX, now make small direct drive pumps for A/C and refrigeration use
• W M Welch Scientific, merged with E H Sargent, bought by Thomas Industries and then in turn bought by Gardner Denver. Welch DuoSeal seal pumps are still manufactured and have been modernized.

In the next article I will discuss the move to direct drive oil sealed rotary vane vacuum pumps and the design changes that were part of it.  

References:

a) History of Mechanical Vacuum Pumps in the United States. Written by D. B. Webb of Vacua Techniques Company, Alamo, CA. This company is now closed. The article was presented at the 2002 Annual Conference of the Society of Vacuum Coaters.
b) History of Vacuum Devices. Written by Pal A. Redhead (NRC, Canada).

Howard Tring / Tel: (610) 792-3505(610) 792-3505 / E-mail: HowardT (at) VacuumAndLowPressure.com / Web: www.vacuumandlowpressure.com

Howard Tring is the owner of Vacuum and Low Pressure Consulting, a company that supplies vacuum pump accessories such as reconditioned inlet traps and exhaust filters and new replacement elements for exhaust filters. Howard also offers on-site vacuum technology and oil sealed vacuum pump repair training and consulting services, customized to the needs of the client. Howard is a member of ASM International and the Heat Treat Society, the AVS, the SME, the SVC and the American Society for Training and Development.

Copyright December 2014, Tring Enterprises LLC – Comments on this article are welcome. I do not profess to know everything about any specific vacuum related subject. However, I have worked in the vacuum pump industry a long time and have seen good, bad and ugly. Please contact me with any comment or question. All messages related to the content of the article will be answered.