The Atmospheric Railway

(from An Illustrated History of Air Pumps, © David L Brittain, 2014)

In the middle 19th century in England, steam locomotives were having trouble climbing some grades with full loads due to slippage of the wheels on the tracks. In addition, they were noisy, dirty with soot, slow, and mechanically unreliable. An inventive and quite ingenious solution came in the form of an atmospheric railway, and believe it or not, it used vacuum to lower the pressure in a tube so that atmospheric pressure could drive a piston connected to a railroad car to power it without an engine—even uphill.

 Several variations of this design were made, with some put into actual use. The most successful—though rather short-lived—line was built in 1846 and ran between Exeter and Starcross on the South Devon coast of England. It worked like this:  The system used an 8.5-mile long continuous line of cast-iron tubes about 20 inches in diameter and 10 feet long with a continuous slot about 2½-inches wide at the top. Other rail lines used 15-inch diameter tubes (for low gradients), or 22 inches in diameter (for steeper gradients). The tube was laid between the rails. A 15 foot-long dumbbell-shaped piston ran inside the tube, with a connecting plate extending up through the top of the tube. This was sealed with a continuous line of leather to act as a valve, which was weighted with iron plates to seal it against a channel filled with a mixture of tallow (made from animal fat) and beeswax, or soap and cod oil. The piston had small wheels before and after it that opened and closed the valve.

A section of the South Devon Atmospheric Railway

by Chowells

 Illustrations of the operation of the valves

The piston was driven by the differential pressure of the atmosphere (hence, the technically correct term of atmospheric railway) pushing against a vacuum created by large air-pumps driven by approximately 80 hp steam engines spaced at about 3 mile intervals. As the train car passed along the track, the connecting plate lifted the leather valve, and the weights closed it after the car passed, thus—in theory—keeping it from leaking air into the vacuum side of the tube. The only thing that kept that theory from working was the elements (the leather tended to dry out and crack, or rot from chemical reactions); the sun, which melted the tallow mixture; and rats, which ate the tallow (and, I’m assuming, got a great thrill when they got sucked—excuse me, atmospherically pushed—into the tubes). Further problems were associated with the lack of communications between the pumping stations in order to advise when a train was approaching so that they could turn on the air pump, and under-sizing of the pumps, which were sized for the 15-inch tubes, not the 22-inch tubes. The latter was never used.

 

An Atmospheric Railway car used in Saint Germain, France.

Interestingly, in separate tests, one train actually reached a top speed of 68 miles-per-hour, and averaged 64 MPH over 4 miles while carrying a load of 28 tons; another hauled a load of 120 tons! The average travel time between the two cities was 28.8 MPH. The vacuum level in the tubes averaged about 16 inches of Hg. A report was given of a man named Frank Elrington, sitting on the piston car without any passenger cars attached. When he released the brake, the car shot off at a very high speed, averaging 84 MPH over the entire 2 mile distance, much of it uphill.

 Alas, in 1860, the last atmospheric railway of this type closed. It was near Paris, France, and ran for 5.5 miles.

“Like a Sail-Boat Before the Wind”

In 1870, in New York City, Alfred Ely Beach (founder and editor of The Scientific American) proposed an atmospheric railway that he called the Beach Pneumatic Transit. It was to be the first subway in NYC and it was constructed in secret at night, because he knew that Boss Tweed would oppose his effort. It was a 312 feet long, 8 feet diameter tube constructed (in only 59 days!) under Broadway from Warren Street to Murray Street to demonstrate the pneumatics-powered subway. 40,000 tickets were sold to curious customers in the first two weeks, and over 100,000 in its first year of operation. A huge 100,000 cfm “fan”, called The Western Tornado, drove the car in one direction; then, the louvers were reversed, creating a vacuum, which powered the car in the other direction. The fan was manufactured by Roots Patent Force-Blast Rotary Blowers. This is the same company that invented of the type of blower known by its name, which is today a generic term:  Roots, (Connersville, IN). (The blower was referred to in news reports and documents as an æolor; not having ever heard this description I tried finding a meaning to this word, with no success until I looked in an old copy of the Oxford English Dictionary (OED). Æolus (l.) is the god of Wind, and an æolist is a worshiper of that god. That would make an æolor a wind machine; a quite appropriate description for a 100,000 cfm blower used to make wind.) With a pressure of only ¼ psi, the blower theoretically generated enough force to drive the car at a speed of from 60 to 100 MPH (though it normally travelled at 7 MPH)!

 Journalist Helen C. Weeks commented on this blower in her article “What a Bore!” 6, “Down we went into a basement, where we faced at once, the greatest blower ever yet seen on this continent. Not, however, a New York politician, as you may have supposed, but a rotary blower … “

 The exterior of the blower was covered with frescoes, although many reports incorrectly state that the frescoes were on the walls of the station.

 Though it operated for nearly three years, the Pneumatic Railway never became a reality, killed by a stock market crash that caused investors to withdraw from the project, and other issues.

 In 1976, a song by the band, Klaatu, titled Sub-Rosa Subway, made it to #62 on Billboard’s Hot 100. Part of the lyrics:

 Back in 1870 just beneath the Great White Way

Alfred Beach worked secretly

Risking all to ride a dream

His wind-machine

His wind-machine

New York City and the morning sun

Were awoken by the strangest sound

Reportedly as far as Washington

The tremors shook the earth as Alfie

Blew underground

Blew underground

...

As for America’s first subway

The public scoffed, “It’s far too rude”

One station filled with Victoria’s age

From frescoed walls and goldfish fountains ...

To Brahmsian tunes

Written by John Woloschuk and Dino Tome

Beach Pneumatic Railway under Broadway, in NYC

 A humorous anecdote is that when Beach entered his pneumatic tube, he thought to himself, “tu-be or not tu-be”.

 The Pneumatic Railway and all its offices moved out of its building at 260 Broadway Ave. in December, 1875. It is interesting to know that in January, 1876, the station and the tunnel for the Pneumatic Railway became a 100-yard rifle range for a newly formed New York organization named the National Rifle Association.

 “The Western Tornado” was a much larger version of today’s Roots blower. This æolor, or blower, was twenty-one feet high, thirteen feet across, sixteen feet long, and weighed 50 tons; and, with its appendages, occupied quite a suite of cavernous rooms, independent of the steam engine and boiler house. 

Adapted from: Scientific American, March 5, 1870

But the story does not end there; in the late 1970s, Brazil built a “people mover” named the Aeromovel. Lightweight cars ride on rails on an elevated hollow concrete box girder that forms an air duct.

An Aeromovel pneumatic train in Indonesia. (The blue air duct is visible below the track.)

 by Gunkarta Gunawan Kartapranata

 

Each car is attached to a square plate—the piston— within the duct, connected by a mast running through a longitudinal slot that is sealed with rubber flaps. A stationary air pump blows air (vacuum is not used here) into the duct to create the differential pressure necessary to drive the car—just like the system more than 200 years earlier. The Aeromovel systems are still in production and are used in theme parks and airports.

Sources:

http://fdelaitre.perso.sfr.fr/Beach.htm

http://www.columbia.edu/~brennan/beach/

http://www.nycsubway.org/wiki/Beach_Pneumatic_Transit

http://www.klaatu.org/klaatu11.html

The Remarkable Pneumatic People Mover, by Alan Bellows; www.damninteresting.com

Atmospheric Railway; www.wikipedia.com

The Applications of Vacuum, from Aristotle to Langmuir; by Theodore E. Madey; Journal of Vacuum Science Technology, April-June 1984.

The Atmospheric Railway; Mike’s Railway History; hppt://mikes.railhistory.railfan.net/r027.html

Unless otherwise noted, all images are in the public domain.

 Images so marked are used with permission according to the Creative Commons Attribution-Share Alike 3.0 Unported license. 

Why "Q"?

OK, so why is your display name Q? 

Two reasons:  First, a number of years ago, a few people at my place of work started calling me "Q". This began as sort of a joke—as do most nicknames—due to a presentation that was made at one of our company's sales meetings. We used a James Bond theme, and I played the character Q, the go-to guy at the home office who came up with all the solutions, usually in the form of gadgets. Since I played that part in our office, it stuck. 

Second, Q is a symbol used in vacuum, known as throughput, or torr·cfm. It's use should be more widespread among us dirty vacuum guys than it is. It's a rule-of-thumb mnemonic that permits one to convert from mass-flow (lbs/hr), to scfm, or to acfm at one pressure to acfm at another pressure with one simple intermediate step. I teach this in our vacuum school. The symbol caught on. 

Here's the rule-of thumb using Q:
Q = torr·cfm = torr x acfm* = Throughput
   ex:  Q = 150 torr x 50 acfm = 7500
Q ÷ 760 = scfm

   ex:  7500/760=9.87 scfm
Q ÷ 169 = lb/hr (air equivalent, mass flow) 

   ex:  7500/169 = 44.37 lb/hr
If you have the same load, but want to operate at a different pressure, say 210 torr, then:
   Q ÷ new pressure = 7500/210 = 35.71 acfm. Therefore, you can use a smaller pump. Mass flow remains the same.

* ACFM at the operating pressure

This should be used only as a rule-of-thumb, since there are other factors (e.g., temperature, molecular weight) that must be considered.

My Gauge is Better than Your Gauge

Q: “A new digital vacuum gauge I’m thinking of buying has an accuracy of ±1%, but the more expensive Capsule Gauges have only ±2% accuracy. Why should I buy a capsule gauge?

A: Because a capsule gauge is more accurate, of course! No, really, it is. OK, it's another one of those "it depends" answers, which means I may be opening a can of worms. But it's a good question, because gauge accuracy is stated in many ways and has many variables.

Generally, the majority of gauges measuring “gauge” (or, relative) pressure are rated with percent-of-full-scale, according to an ASME standard that ranks gauge accuracy by grades. Grade D is the least accurate and Grade 4A is the best. To further complicate the grading process, some grades have different accuracies depending on where you are on the scale. 

Accuracy depends on a number of factors including resolution, readability (size and scale length), repeatability, friction, range, parallax, and hysteresis (variation difference from ac­tual pressure between rising and falling pressures after tapping to eliminate friction errors), and whether the error is percent of reading (%R) or full scale (%FS).

To begin with, %R is always better than %FS for two gauges with the same percentage of error. A %R gauge can even be better than a %FS gauge that has a smaller (better) percent­age of error, depending on where the pressure is read.

I know that sounds confusing, but it’s because with a %FS gauge, the total percentage of the error at any indicated pressure gets worse as you approach the zero point. Let’s assume we have a ±1%FS 0-30 in.Hg relative vacuum gauge; one percent of the full scale num­ber (say 30, for 30 in.HgV) is 0.3% at close to the 30 in.HgV point on the scale. At 1 in.HgV (nearly atmospheric), the error of 0.3 in.Hg is 33 percent! Add to that, the fact that some gauges have different accuracies at the middle versus both ends—the high and low quar­ters—of the scale, and it gets worse. For example, one brand’s 2-inch relative vacuum gauge has an accuracy of ±2/1/2% of span (which is another way of saying full scale). That’s not two-and-one-half; rather, it is two/one/two, and it means that the quarters clos­est to the two ends of the scale have a ±2% accuracy, but in the middle half of the scale (7.5 to 22.5 in.HgV) the accuracy is ±1%. If you are operating at, say, 25 in.HgV, then the actual pressure can be anywhere between 24.4 to 25.6 in.HgV.

A ±1%R gauge, however, has the same 0.3 in.Hg error at near the 30 in.HgV reading, but at 1 in.HgV, the error is only 0.01 in.Hg, or one percent. That’s a big difference from 33 percent. The percentage of the error did not change, but the possible actual pressure variation did.

Consider four different gauge styles, each with a different accuracy.  A 2½” liquid-filled gauge has an accuracy of ±2/1/2%FS, while a 4” liquid-filled gauge has ±1%FS accuracy. The capsule gauges have ±2%R accuracy, and the digital gauges are ±1%FS accurate. From reading this, one would expect that the digital gauge would be best. But being digital doesn’t ensure accuracy, only readability. Since it has only one pressure reading and often has large digits, its readability is unparalleled. Any error resulting from trying to interpolate between the printed numbers on the dial face by counting the lines indicating the increments between those numbers, is eliminated.

Digital gauges usually “sample”, or take a pressure reading every second or so; therefore, you get readings that look like they are jumping all over the place since instantaneous pressure is always changing (pressure fluctuations are less noticeable if the system volume is large). An analog gauge reacts to minor pressure fluctuations much more slowly, and therefore gives more stable readings. Speaking of stability, a liquid-filled gauge is not nec­essarily more accurate. The liquid mostly smooths out pressure fluctuations and vibrations.

Now, let’s look at the capsule gauge. The capsule gauges can be ordered with different scales, so you can select a range that best meets your operating conditions. Most im­portant, the capsule gauge measures absolute pressure. That means it is unaffected by effects of altitude, but gauge accuracy is also affected. The zero point is at the opposite end of the scale from a scale on a relative pressure gauge. That means that the gauge accuracy is best at the deepest vacuum level; add to that the fact that we can improve readability by choosing a gauge with a smaller portion of the total pressure range—say 20 to 0 torr—and the accuracy of a 2%R gauge can be ±0.2 torr at 20 torr, or ±0.02 torr at the 2 torr reading. That’s several orders of magnitude better than a ±1%FS relative vacuum gauge at about 29 in.HgV, and a lot better than the ±7.5 torr variation of the ±1%FS digital gauge. These capsule gauges are made using jeweled bearings, which eliminates most friction and adds to their accuracy, repeatability … and cost.

Of course, you should always purchase the gauge that best suits your application. If you are a thermoformer working at, say 22 in.HgV, then a bourdon tube type gauge should be sufficient. If you are working with evaporation of water or solvents, then an absolute pressure gauge like the capsule gauge would be your choice.

One final note:  Please, USE gauges. In my experience of nearly 45 years, very few users even have a gauge present on their vacuum systems or processes. In trying to trouble-shoot application  questions, when I ask what vacuum level users have in their processes, most don’t know. In these cases, any gauge would have been helpful—regardless of accuracy.

Would You Like Ice in Your Drink?

Q: “Would you like ice in your drink?
A: “Yes, but make it ice Ih, please.”

Many of us may recall our high school or college physics class, and learning about water and its physical characteristics. We learned from our instructors that it has three phases: liquid, ice, and vapor; and that it has a triple point: a single specific pressure and temperature at which all three phases of water can coexist, any of which can abruptly change from one state to the other.

They lied.

As vacuum engineers, many of our applications deal with water in one of its forms. It may be drying, distilling, or packaging, to name a few. Thus understanding its characteristics is important.

Water is a very complex compound, and we are still discovering new things about it. For example, we now have identified at least 14 phases of water, not just the three we were taught about in school. Other than the liquid and vapor phases, all other phases are of different forms of ice. These phases depend on pressure and temperature. Each phase of ice is called by its number (e.g., ice VI is ice six, ice XI is ice eleven). The ice you may be enjoying in your iced-tea glass is called ice Ih, as in “one h” (the “h” stands for hexagonal—the shape of the crystals; it’s why snow crystals have six sides). Crystals in some forms of ice are cubes (but that’s not why we have ice cubes), such as ice Ic; some are trigonal, some are tetragonal.

We were taught also that ice floats in water, with nine-tenths below the surface; that’s because it has a density of 0.92, which is lower than liquid water’s 1.0. But some ices would sink straight to the bottom, because their density is far greater than liquid water—up to 2.51. Higher density ice is created by higher pressures—the kind of pressures that might be found at the bottom of 800+ miles of ice on other planets.

Rather than the single triple point, we now know that there are at least 12 triple points for ice and water. Most of these triple points are for the various phases of ice itself. Alas, all these other triple points and phases of water will remain a mystery for most of us in our day-to-day engineering; we’ll still be dealing with the triple point and three states of water that we were taught about in school.

But, the next time you are asked if you want ice in your drink, you can now say that you want ice Ih.

Then, sit back and chill.

A dirty vacuum guy

This is a beginning. Of what, I don't know yet. I intend to write about the world of industrial vacuum. From time to time, as subjects or applications arise I will post some notes. I don't know how I'm going to do it without giving away too much detail (industrial spies, you know), which could affect my work, but I'll try. It will be a learning process, for sure.

I have spent 40 years in the field, working for a number of companies that produce a wide variety of types of vacuum pumps. The one thing that they all had in common was that the pumps they produced were used in the rough vacuum field. This covered vacuum levels from atmospheric to 29.9 in.HgV (0.5 torr). We would occasionally combine them with roots-type blowers to pump down to about 20 microns.

In my experience, folks who work in the high vacuum field tend to look at us with scorn, since we are in the "dirty" vacuum range. But, I would contend that they would have difficulty with many applications in our range. Some things get more complicated in our world; take pump-down.

Pump-down equations use average volumetric flow (acfm) to determine time to evacuate a closed chamber. this is OK when a pump that can pull down to a deep vacuum (low absolute pressure) is used, since the acfm curve is flat, thus the average acfm is, well, average. What I mean is that it stays constant, whereas a rough vacuum pump may have a curve that is, well, a curve. Being a curve, the average is a variable. In order to calculate the average it is necessary to perform the calculations in steps, then total all the averages. The points between the two pressures must become narrower as you approach the pump's ultimate pressure, since pump-down is logarithmic. It should be noted that the higher the pump's ultimate absolute pressure, the more pronounced the curve.

For the high-vacuum guys, the equation is simple. They perform it one time. For us dirty guys, we must perform it many times. A spreadsheet makes this process quite easy, however. It is also possible to create a list of correction factors, based on the pump's ultimate pressure. This can get complicated, however, since there are many possible ultimate pressures.

Anyway, that's the first post. I welcome your comments--if you don't mind getting "dirty".