The Liquid-Ring Pump: The Water Is the Piston

Use a length of clear acrylic tube of about 100 mm bore as the casing, closed by two acrylic end plates bolted through with machine screws and sealed with O-rings in printed grooves. Print a 12-bladed impeller in PETG, 80 mm across, with radial blades and a hub for a steel shaft. Make it the full width of the casing less 0.5 mm, so each blade nearly touches both end plates — the end plates are where the gas ports are, and a large side gap lets the gas short-circuit from outlet back to inlet. Mount the shaft in bearings in the end plates **8 mm off the casing's centre**, so the impeller nearly touches the casing at the top and clears it by about 18 mm at the bottom. Nash put the variation in the casing shape instead — the enlarged lobe — but the effect on the water ring is the same.

The ports are crescent-shaped slots in one end plate, close to the hub, and their position is the whole design. - The **inlet** slot goes where the ring is moving AWAY from the hub — where the buckets are emptying — covering the arc from just past the close point to the bottom. - The **discharge** slot goes where the ring is moving back IN — where the buckets are filling — ending just before the close point. - Between the end of one slot and the start of the other, leave solid plate at least one blade-spacing wide at each end, so that no bucket is ever open to both at once. Mark the direction of rotation on the plate before you cut. The same pump turned backwards blows out of its inlet. Fit a hose tail to each slot through a printed manifold and seal it with silicone. Tee the vacuum gauge into the inlet hose.

Fill the casing to about a third with water with a drop of dye in it, and drive the shaft with the cordless drill. Below a certain speed the water simply sloshes. Above it the water climbs the wall and forms a ring, and you can see it: deep at the bottom, thin at the top, and the buckets emptying and filling as they pass. Close the inlet hose with your thumb and read the gauge. Increase the speed and read it again. Then feed the service water: lead a slow trickle of cold water into the inlet side and let the discharge carry the surplus out with the air into a bucket. Measure the temperature of the water leaving. Then do the test that matters: run once with cold water, once with warm. The warm ring will not reach the same vacuum however fast you drive it — the notebook in the next step says exactly how far it will get.

**US 1,091,529, 'Pump and Air Compressor', Lewis Hallock Nash of New York, assignor to the Nash Engineering Company; filed 24 February 1910, patented 31 March 1914.** The specification describes a wheel turning in a cylindrical case chamber with an enlarged portion or *"lobe"*: the water in a bucket *"is caused to flow out into said enlargement by the centrifugal force of the revolving wheel"*, leaving an opening *"through which air or other gas can be drawn"*, and returns as the lobe contracts to eject it. Nash founded the Nash Engineering Company in 1905, in Norwalk, Connecticut; the Nash name is still on liquid-ring pumps today. The liquid ring's great virtue is what it tolerates. Gas saturated with water vapour is no problem — the vapour condenses into the ring. Particles pass through. A slug of liquid is simply added to the ring. That is why it drains the water from the felts of paper machines and holds the vacuum on steam condensers, where a dry pump would drown. **Honest limits.** Its efficiency is low: much of the shaft power churns the ring. The vapour pressure of the service liquid caps the vacuum. It needs a supply of service liquid and gets rid of warm effluent. And the service liquid mixes with what is pumped, so if the gas is toxic, the water becomes a waste stream to treat.