How to build vacuum engine

Power is transmitted in everyday life most often by electricity. There are other means of power transmission such as high-pressure air, high-pressure hydraulic oil, and, on industrial sites, steam. However, electricity dominates: it is the most versatile form of energy. It can be converted efficiently to any other form of energy, something that is not true of other types.

Most kinds of power transmission have a certain degree of tangibility. The rotating propeller shaft of a large truck leaves no doubt that considerable power flows from the engine at the front to the axles at the rear. Wander near a highpower electric system, and you’ll readily hear the low but insistent hum at the 50 or 60 Hz line frequency; sometimes you’ll even feel the hairs on your body react. The apparently inert wires and cables of high-power electric systems can produce huge and mortally dangerous flashes and sparks if they are disturbed. Similarly noisy and spectacular gas jets signal the presence of even small leaks in compressed air or steam systems.

By comparison, the transmission of power through a vacuum in a pipe seems a peculiarly intangible concept. How can power be apparently transmitted by nothing? But in this project we show that a vacuum can indeed transmit power, and that we can demonstrate a motor rather like an old-fashioned steam engine, an engine that can turn the power transmitted by a vacuum in a pipe into mechanical energy.

The Industrial Revolution that transformed the Western world, starting about 1700, needed mechanical power. At first, increased use and more efficient designs of watermills and windmills could provide that power. But it gradually became evident that the continuous power which steam could provide was going to be needed. It is easy to appreciate the expansive force of steam when you see a kettle boil. However, none of the early steam engines used that expansive power. Instead they used atmospheric pressure (they became known later as “atmospheric” engines), with the steam being used to create a vacuum so that the atmosphere could push a piston. We might today, perhaps less accurately, call them vacuum engines.

There have been times when vacuum power transmission has been used. Perhaps the first example was the system used by Matthew Boulton and his partner James Watt. Near the Boulton and Watt engine factory in Birmingham, England—the world’s first engine factory—was the Boulton and Watt mint, a coin factory operated by one of the company’s own engines. Engineer John Southern devised a system in which a steam-driven vacuum pump partially evacuated a huge pipe, known then as the “spirit pipe.” Individual coin presses were powered by cylinders and pistons connected to the spirit pipe.

Since the time of Watt and Southern, vacuum power distribution has occasionally resurfaced in different places. Vintage automobiles from the 1920s on were sometimes fitted with a kind of vacuum engine to operate the windshield wipers, using the vacuum from the gasoline engine’s inlet manifold. It cannot have been an ideal system: if an engine turns over slowly, the vacuum from the engine would decrease and the wipers would operate more slowly. If you were driving one of these old cars and saw an approaching hazard, you would naturally slow down. And just when you needed more wipes of the windshield to see what was going on, the opposite would happen: the wipers would slow down and you would be left peering through rain-swept glass at exactly the wrong moment!

Today this principle is still being used in at least one application (albeit rarely): vacuum cleaners. In some models of cylinder vacuum cleaners with a rotating brush, the brush is powered by a simple turbine device that is turned by air sucked into a vacuum created by a centrifugal fan in the cylinder.

Our vacuum engine is a “steam engine” type of device. Unlike most steam engines, however, it does not require a fully equipped workshop with lathe, milling machine, and so on. Neither does it need the thousandth-of-an-inch accuracy required of a working model steam engine. The vacuum engine only requires a few hand tools, pieces of wood, plastic tubing, and easily obtained metal hardware, and you don’t need to make anything more accurately than within a millimeter. You won’t burn your fingers, either—because you don’t need steam! It is also easy to make—you can probably assemble one in an afternoon. Nevertheless, it well illustrates all the main working principles of steam engines: piston and cylinder, crank, flywheel, valve gear, and valve timing. Take a look at books like that of Semmens and Goldfinch if you want to know more about steam engines.

What you need
  • Vacuum cleaner (ideally the horizontal cylinder kind)
  • Short section of 18-mm (3/4-inch) hose
  • 300-mm-long, 32-mm-diameter plastic pipe
  • ca. 150-mm-long, 31-mm-diameter round section of wood to fit snugly in pipe
  • Flywheel pulley from an old washing machine
  • Brass rod that will roughly fit the hole in the flywheel
  • Metal shaft and brackets
  • Conrods (e.g., 8-inch by 1/2-inch Erector set strips)
  • Wood pieces
  • Electric drill
  • Bolts and nuts
  • Hot-melt glue

How to build
The basic idea of the vacuum engine is that a piston is propelled up and down to push a crank that connects to a flywheel. The piston is activated by atmospheric pressure on its connecting rod (conrod) side, with periodic pulses of vacuum applied to its piston-head side. The pulses of vacuum pressure are applied by intermittently connecting the low pressure from a vacuum cleaner to the piston. The intermittent connection is made by a slide valve. The valve is synchronized to the flywheel rotation and hence to the piston movement, by being actuated 90 degrees out of phase with the piston in terms of flywheel position.

I found a piece of wooden dowel that fit snugly inside the drainpipe I had chosen. I then used this rod and pipe for both the piston and the slide valve. I suggest that you aim for a piston that is about 1 mm smaller in diameter than the cylinder, both for the piston and for the slide-valve assembly. Try to find plastic pipe that is close to precisely round. (You will find occasional pipes or sometimes even entire batches that are appreciably noncircular; perhaps they have been squashed in storage or loading at the factory or supplier.)

The piston, if it is the right size, needs no preparation at all other than to bevel the edges and to screw on the conrod bracket. The slide valve and its cylinder are more complicated. The cylinder needs two or three holes (an air inlet hole is optional) as shown in the diagram, which all need their edges smoothed. You must drill through the valve body for the vacuum port and then make a slot with a chisel for the transfer port. The transfer port allows air into the drive cylinder after it has completed its power stroke.

I have two suggestions for alternative, simpler slide-valve designs: First, you can omit the air-inlet hole and the transfer port channel, relying on air leaking around the piston and valve. Second, you can omit the air hole and transfer port from the valve body and also cap its end. You can now switch the valve on and off with a simple cylindrical piston (exactly like the power piston), by arranging that the piston just uncovers the holes in the valve body as it reaches top dead center (TDC). You will need to cap the end of the valve cylinder in this design too.

I used Erector set parts to construct the crank plate and light steel strips for the conrods. The pipe work was completed with a washing-machine drain hose, which is typically a fairly generous-bore 18-mm (3/4-inch) corrugated pipe. You can minimize the Erector set parts by making your own bearing for the flywheel and fitting the crank pin directly into a small hole drilled into the flywheel. The bearing for the flywheel can be made using a piece of 6-mm (1/4-inch) steel and a piece of brass rod around 15 mm (5/8 inch) in diameter, glued with epoxy adhesive into the center of the flywheel central hole. Bore out the middle of the brass rod with a 6-mm (1/4-inch) drill, deburr it if necessary with an oversize drill bit or just a sharp knife, then run the drill up and down it a few times until the rod will fit snugly but freely rotate around the 6-mm (1/4-inch) steel rod.

The position of the cylinder on the base plate is not critical. The position of the valve body, however, is more sensitive: it must just begin to open to vacuum when the piston is closest to the flywheel (the position conventionally known as TDC).

You must ensure that every part can move freely. Check that the edges of the holes in the valve piston cylinder are smooth and that the pivots on the pistons and the crank are not binding. If rotated vigorously by hand without the vacuum applied, the engine should turn over at least three or four times. If you find that the engine slows more quickly than this, you should check for excess friction in one of the parts.

What you do
Without a plentiful supply of vacuum, your engine won’t work, so make sure that your vacuum cleaner has powerful suction. The stronger the vacuum—meaning the larger the negative pressure relative to atmosphere—the better the vacuum engine will perform. If you hold any doubts concerning the performance of the vacuum cleaner, try to find some means of measuring the negative pressure it produces. The flow rate that the vacuum cleaner can produce is rather less important, as the flow rate needed by the vacuum engine is fairly low and, unless your fabrication of the device is more precise than I have suggested, much of the air flow will go to supplying leaks rather than to propelling the engine. If your vacuum cleaner has a low flow rate, you can still operate a vacuum engine, but you must make the piston and valve pieces a tighter fit within their cylinders.

Now position the flywheel just a little past the TDC. Apply the vacuum. With luck, you should find that the flywheel should begin to turn of its own accord, rushing down toward bottom dead center (BDC) and then beginning to slow down. But it should be going just fast enough to rotate one complete revolution at low speed, after which the process can repeat. The next time the engine will reach TDC a little faster, and the flywheel will complete its revolution more quickly. With the dimensions given here and a reasonably powerful vacuum cleaner, your vacuum engine should build up in speed until it is whirling around at 300 to 400 rpm or more.

How it works
The vacuum engine works by atmospheric air pressure. When the flywheel is at TDC, air pressure is the same on both sides of the piston, so no force is applied. With the flywheel turned a little, so that the valve opens to the vacuum, air is removed from underneath the piston. With no air pressure below but atmospheric air pressure above, the piston is forced downward.

Curiously, in the engines I have tried, the rather rough-and-ready fit of the wooden piston to cylinder may help, in that the air inlet and the transfer passage in the valve gear did not seem to be necessary. As mentioned earlier, this means that you can simplify the engine and use a piston as the slide valve. With a better standard of construction, you will need a proper slide valve with an air inlet.

Like the original steam engines, your vacuum engine may need a little adjustment before it will run properly (or perhaps run at all). You do need to ensure that all parts run smoothly and are lubricated with a little light oil such as bicycle oil.

The highest friction forces, assuming that all your components are smooth running under freewheel conditions, will be developed when the vacuum is applied to the slide valve. With a fairly close-fitting slide valve, this force will be reasonably low. If, however, like me, you started with a rather loose-fitting slide valve, you will find that it tends to bind. What is happening here is that the valve piston is being pulled hard against the valve cylinder because of air pressure on the side opposite the vacuum cleaner connection. Some oil may fix the problem. If the fit is really loose, worse than 1.5 mm smaller than the cylinder, then it may be necessary to start again and make another slide-valve piston with a better fit. A simpler solution if the fit is not too bad is to glue a cap onto the end of the slide-valve cylinder. This blocks half the flow of leakage air to the slide valve and reduces the force needed to operate the valve. (Thanks to the kids at the Saturday Activity Center in Guildford, U.K., for that tip.) Of course, if you have used the simplified piston-style slide valve, then you will have blocked off the end anyway.