The Budget Vacuum Pump

Copyright 2001, 2002 by Jim Wilson
(Click images to enlarge)

Refrigerator Compressor Vacuum Pump

My budget vacuum pump system is up and running well. Near as I can tell, it pulls and holds about 27" Hg (to perhaps 9.9 kPA, 99 mbar, 74 torr, or whatever at standard pressure), which should certainly be adequate for everything I plan for it.

Some (very expensive) systems are capable of a much superior vacuum, but I really don't need millitorr pressures, and I like to save hard-earned dollars!

What's it for?

I built this little shop vacuum pump primarily for sealing the wooden parts of tools I make, but also for just playing around, basically. Additional intended uses include degassing castings, bagging foam wings for radio control planes, showing my kids that, yes, water really can boil at room temperature, and who knows what else might be dreamt up down the road.

The Basic Parts

The main component is a junk refrigerator compressor I got from the scrap yard for $3. The inlet line is 5/16" OD copper tubing; the outlet is 1/4" OD steel tubing. It was quite easy to put together, although it did take a couple trials to get the plumbing working perfectly. The challenges were purely from my not knowing what would work well.

The first trip to the hardware store convinced me to go with 3/8" flare fittings for the interface to the various fixtures I planned. It has a larger OD than the pump inlet, works well under vacuum, and has the greatest variety of readily available adapters to other sizes and types of plumbing fittings.

I made a mistake in choosing a needle valve to isolate the pump from the evacuated system. The needle valve requires too many turns, gives a poor seal and has a less than positive stop when fully open and closed. I replaced it with a ball valve. Much better.


One of the obstacles I encountered was the 5/16" inlet line -- I couldn't find anything in that size locally, except compression fittings, which I wanted to avoid. Finally, I reamed a short piece of 3/8" refrigeration tubing to 5/16", soldered it to the inlet, and viola -- the inlet was resized to 3/8" OD. From there, the plumbing was easy.

Another challenge was making the final connection flexible. I was worried that a hose would collapse. I think many of those that I looked at would. There were several types of plastic and metal tubing that seemed as though they'd be strong enough, but they were not very flexible. The most promising seemed to be a thick-walled clear vinyl tubing with fiber running through it, but the store I had visited was out of stock in the size I wanted.

So, I tried 3/8" refrigeration tubing, which can be bent, although I definitely wouldn't call it flexible. As I had feared, it was simply too unwieldy. I crossed my fingers and replaced it with a three foot length of standard 3/8" ID 600 psi air hose. I used 3/8" barb to 1/4 NPT adapters on each end. It works great! I had thought the barbs might leak, but they don't -- not yet, anyway. If that problem does come up, I'll install hose clamps.

I curved the outlet tube so it points straight up and flared the end. A couple pieces of scrap tubing and a bearing ball function as a simple check valve.

An unused vacuum gauge from an old swimming pool cleaning system completed the setup. Total out-of-pocket cost was about $30, of which $15 was for a flaring tool, which I wanted anyway.

The Finished Unit

Refrigerator Compressor Vacuum Pump   The compressor-turned-pump

The little black object on the left side and wired to the pump is the pump motor run capacitor. The suction line runs to a ball valve -- you can see the handle above and behind the vacuum gauge -- then to the vacuum gauge and finally to a hose that connects to whatever -- usually the budget vacuum chamber I built.

The ball valve allows the pump to be isolated from the vacuum chamber, which can be handy since some compressors leak a bit when turned off. However, this one hasn't leaked, so the isolation valve isn't really necessary.

Between the chamber and the vacuum gauge is a tee fitting with another ball valve. This one is closed when the chamber is pumped down. Opening it returns the chamber to atmospheric pressure.

On the shelf under the vacuum gauge are some fittings I've used to connect to various other things, such as an air conditioner coil the serviceman thought was leaking. By the way, the coil itself checked good; the problem turned out to be a leaky fitting. Being able to test it myself instead of accepting the serviceman's guess ended up saving me about $400. Are we starting to notice a pattern here? (G)

Check Valve   Sticking up in back of the pump is the outlet line, on top of which is a simple check valve made from junk. The end of the 1/4" tube is flared, and resting in the flare is a 5/16" ball bearing. A short piece of scrap rubber hose is fitted on the tube below the flare. A larger piece of plastic hose, plugged at the end and perforated below the plug serves to contain the ball and complete the valve.

The Vacuum Chamber

Vacuum Chamber   I built a vacuum chamber to seal the mortise chisel handles and other wooden parts for tools I make.

The chamber is made from a 39" piece of 8" PVC pipe, with a 12"x12" piece of 1/2" thick PVC glued on the bottom. I scrounged the pipe from a local commercial sewer contractor. The top is a piece of 1/2" acrylic sheet with a polyurethane seal.

Chamber port - outside   I drilled a hole in the side of the chamber near the top and cemented in a 1/2" PVC female adapter. A 1/2" NPT to 3/8" hose barb adapter connects the vacuum hose.
Also in these pictures you can see the basket for getting parts in and out of the chamber. The basket is made of 1/4" hardware cloth, 24" deep and a close fit to the inside the chamber. You're actually looking at the basket's bottom; it's upside down in the chamber and propped up with a piece of dowel so the bottom can be used as a table. I took these pictures when my son and some of his friends were "discovering" that water does indeed boil at room temperature in a vacuum.   Chamber port - inside

One thing you can't see very well is how smooth (or how flat) the edge of the pipe is. A good chunk of a Saturday morning was devoted to cutting, filing, sanding, and polishing both ends so they would seal well on a flat plate. The seal is very good. When the system is not in use, I pump it down a tad (3 or 4 psi), and close the isolation valve. Even after sitting for several weeks, the indicated pressure is unchanged.

The parts for the chamber cost more than the pump itself! I had to buy the PVC bottom and acrylic top squares, and a couple of fittings. The PVC base was about $8; the acrylic lid cost $10.

To seal the lid, I cut a gasket from some 1/8" polyurethane material I had on hand. It's expensive stuff -- runs about $25 per square foot -- and is supposed to be resistant to oils and solvents. Unfortunately, after a few weeks of sitting above a chamber full of polyurethane finish and mineral spirits, it had swollen and distorted to the point of being useless. I'm still experimenting with various gasket materials. When I find something good, I'll post it here. Some of the alternatives I'm trying are the cheap red rubber sheet you can get at the hardware store and some soft vinyl sheet, such as the stuff used in bath and shower installations.

So How Well Does It Work?

(Apologies in advance to all the scientists out there for these silly units!) The chamber has a volume of 1828 cubic inches. The plumbing adds 17 cubic inches, for a total volume of 1845 cubic inches in the system.

The gauge reads 0 at atmospheric pressure, and negative inches of mercury (in Hg) for a vacuum. The pump evacuates the system to -5 in Hg (-2.5 psi) in 12 seconds. It takes about six minutes to pump the chamber down to -27 in Hg (-13.3 psi). Just how good is that?

My shop is about 2250 feet above sea level, so the absolute (i.e., "true") atmospheric pressure here should theoretically average around 27.5 in Hg, or equivalently, 13.5 psi. (ref)

This turns out to be a pretty accurate number. Although I don't have a barometer to check it directly, I was able to calculate the actual absolute atsmospheric pressure based on barometric pressures given in the local weather report. Incidentally, the barometric pressure reported by weather stations is always adjusted to sea level, so it must be tweaked back to your actual altitude to give your true atmospheric pressure. USA Today explains all this pretty well. The formula to do the conversion is a little messy because of unit conversions, but here it is:

AP = 0.02953*(((33.8639* BP )^0.190284-0.000084288*0.3048* EL )^(1/0.190284)+0.3)

Here is a slightly simpler form:

AP = 0.02953* (1.95472* BP^0.190284 - EL/38924)^5.2553 + 0.008859

AP is the calculated atmospheric pressure (what a barometer at your location would say); BP is the barometric pressure reported by the weather station, sometimes called the altimeter setting; and EL is your elevation. If you want to avoid the math, Richard Shelquist has an online calculator that will do it for you. By the way, temperature and humidity don't affect the absolute air pressure calculation -- they're already accounted for in the value reported by the weather station. Richard's calculator just wants them for other outputs.

Anyway, the calculated values for my location do in fact hover very near the expected 27.5 in Hg mark, fluctuating only slighty due to the weather. The bottom line is, assuming my gauge is reasonably accurate, the system can pull the absolute pressure inside the chamber down to somewhere near 0.2 psi. I think that is a pretty darn good vacuum for the price!

What Next?

Someday, I want to add a diaphragm pressure control switch, similar to THE VACUUM SWITCH, but for now the pump is simply powered by a switched outlet:


I got some of the more important ideas for this system from this web page. The vacuum chamber is similar to one built by Steve Knight.

More Web Pages

Toolmaking  •  Shop Tour