Out the Side of the Mountain

In 1994 I built a mass driver.

It was six solenoid coils off some small 120-volt control valves, slid onto a piece of stainless tubing, wired to a chain of magnetic motor starters so that each coil fired as the one before it let go. You dropped a steel screw in the breech and hit the button, and the screw went down the tube and out the end and six to twelve feet across the room, and landed on the floor. It was crude. The timing was done with relays, not electronics, and half the time one fired late and the screw dribbled out the end. But when it worked, it worked, and there was no powder, no heat, no noise to speak of. Just a click and the screw was gone. It did not dent anything and I am not going to pretend it did; how it was wired, and what I would build next, is a separate piece.

I had been reading about mass drivers for years by then, and I wanted to know if the idea was real before I put it in a magazine. It was real. Gerard O'Neill and Henry Kolm had built one at MIT in 1977 out of scrap from a magnet lab, and it hit thirty-three gees. The next version, a few years later, hit several hundred.¹ Mine threw a screw across a room. The principle was the same.

This article is about the version I have been drawing ever since.

What a mass driver is, and what it isn't

Strip the science fiction off and a mass driver is a linear electric motor with nothing riding on it but the load. A row of electromagnets fires in sequence down a track; a carrier with its own magnet gets pulled toward the next coil, then pushed by the one it just passed, over and over, faster and faster, until it leaves the end of the track at whatever speed the track was long enough to give it. No fuel on board. No engine on board. All the energy comes out of the wall, and every part that costs money stays on the ground and gets used again tomorrow.

That is the whole appeal. A rocket carries its own propellant, and the propellant weighs more than everything else combined, so most of what a rocket lifts is the stuff it needs to lift itself. A mass driver leaves all of that behind. The thing on the track is the thing you wanted in orbit, and not much else.

O'Neill designed it for the Moon, where a few kilometers of track and a gentle exit speed would throw lunar rock to a point in space where it could be caught and smelted.² On the Moon there is no air and gravity is light, and a mass driver is close to an ideal machine. On Earth it is a much harder proposition, and it is worth being plain about why.

Why Earth is hard

Three things. Speed, air, and heat.

Speed first. To reach orbit from the ground you need something like eight kilometers a second, and by the time you have paid for the air on the way up, closer to nine or ten. A track can only get you there so fast, and the pace it gets you there is set by how many gees you are willing to put on the load. Do the arithmetic on a twenty-mile track, which is thirty-two kilometers, and it comes out to about a hundred gees for eight seconds to leave at eight kilometers a second.³ A hundred gees kills a person and crumples most machines. It does not bother a sack of seeds, a bar of metal, a tank of water, a spool of wire, or a pallet of printed panels. So this is a freight line. Cargo only, no exceptions. People go up on rockets, the way they do now, and the rockets carry nothing but people, because everything else went up on the train.

Air second. Whatever leaves the end of the track leaves it at hypersonic speed into the thickest air on the trip. That is backwards from a rocket, which is going slow where the air is thick and fast where it is thin. The pod has to punch through with a nose that can take the heat and a shape that does not tumble, and it will lose a good fraction of its speed doing it. Every bit of altitude you can give the muzzle is speed you do not have to buy with magnets. Which is why every version of this idea since the seventies ends the track on top of a mountain.

Heat third, and nobody talks about this one. All that energy comes out of the wall and goes into the load, and a chunk of it goes into the coils on the way. Copper has resistance. Iron has losses. Fire a thousand magnets in eight seconds with the current it takes to move a ton of freight at a hundred gees and the track gets hot, and if you want to fire again in ten minutes the heat has to go somewhere. On a lunar mass driver you radiate it into a black sky. On Earth you are underground, in rock, in a tube you have pumped the air out of, and the heat has nowhere to go unless you build it a road.

Marshall Savage's Bifrost

The version most people know is in Marshall Savage's The Millennial Project, a book from 1992 that I read the year it came out and that I still think is the most ambitious plan for getting off this planet anybody has put between two covers.⁴ Savage called his launcher Bifrost, after the rainbow bridge in the Norse stories. It was a magnetic track in a vacuum tube, starting horizontal under the Serengeti plain, curving up through the inside of Mount Kilimanjaro, and exiting at the summit, nineteen thousand feet up and three degrees off the equator. Because the track alone could not deliver orbital speed, the pod was to be finished off by ground-based lasers focused on a block of ice on its tail, boiling the ice off as rocket exhaust.⁵

I want to give Savage his due, because I am about to disagree with him and I do not want it mistaken for dismissal. The vacuum tube is his. The horizontal start curving up through the mountain is his. The equator and the big mountain are his. The people who took up his work after him have set the mountain launcher aside as unworkable in that form,⁶ and I think they gave up on it too early, but that is a different argument. What I have been drawing is not Bifrost. It diverges in three places, and I think the places matter.

Where I go a different way

Out the flank, not the summit. Savage put the muzzle at the peak because the peak is highest. I would put it on the upper flank, somewhere in the last third of the climb, for two reasons. The first is that a summit is a knife edge and a launch complex is not: you need a muzzle door, a shock structure, a cooling plant, a rail yard, a road, and a place to stand, and none of that fits on a peak without leveling it. The second is that the summit of the mountain I have in mind is sacred to the people who live under it, and I would rather build somewhere else than fight about that.

Down, not through. Savage ran his track through the heart of the mountain. I would run it deep, under the plain and under the mountain both, and come up only at the end. Rock is a heat sink. Rock is shielding. A track ten stories underground does not care about weather, sabotage, or a bad day at the surface, and every mile of it is sitting in a few thousand feet of basalt that will soak up a great deal of heat before it notices.

Cool it like a machine, not like a magnet. This is the piece I have not seen anybody draw. Every coil on the track gets a liquid jacket, plumbed as a closed loop, the way you cool a big electric motor or a welding transformer. The loop runs to a chiller plant underground, and the chiller dumps its heat into the rock and into a water loop that comes out somewhere useful, because a launch complex that is also a hot-water plant is a launch complex that pays part of its own bills. I have spent a career around big electrical gear in cold places, and the rule is the same everywhere: the machine that lives is the machine you built a way to cool.

Why Hawaii

Savage wanted the equator, and he was right to want it. The Earth spins about a thousand miles an hour at the equator and every bit of that is free speed for anything launched eastward. But Kilimanjaro is in Tanzania, and the mountains in Ecuador are in Ecuador, and I am an American writing about something my country could build. The farthest south the United States goes with a mountain on it is the Big Island of Hawaii, at about twenty degrees north, and the mountains there are Mauna Kea and Mauna Loa, both a little under fourteen thousand feet.⁷ Twenty degrees costs you a few percent of the spin bonus compared to the equator. It is not nothing, but it is the price of building on your own ground, and there is nothing else in the country that comes close.

Both mountains are shield volcanoes, which is to say they are enormous and gently sloped, which is exactly what you want for a twenty-mile tunnel that has to start flat and come up slowly. Mauna Loa is active. Mauna Kea is not, and it has a road to the top and a summit full of telescopes, and that is where the trouble is: the summit is sacred, the telescopes have been fought over for years, and I would not put a launch door up there if you paid me.⁸ The flank is a different matter. The lower slopes are ranch land and lava field. There is room.

So: a portal near the coast on the east side, where the freight comes in by ship. A tunnel running west, deep, flat for the first fifteen miles, then beginning to bend upward, and coming out on the upper east flank of Mauna Kea at somewhere around nine or ten thousand feet, aimed east over the Pacific. Twenty miles of track, evacuated, coils jacketed, chiller plant at the low point. Freight pods rolled into the breech one behind the other, the way you roll cars onto a train, because a track that fires once a day is a science project and a track that fires every ten minutes is a railroad.

What comes out the end

A pod, a few tons, going better than eight kilometers a second at ten thousand feet, heading east and up. It punches through the atmosphere in about a minute and comes out the top of the air at something less than it started with. And here is the part Savage solved with lasers and I would solve differently: it is not in orbit yet. A thing thrown from the ground is on a path that comes back to the ground, and it needs a nudge at the top to bend that path into a circle.

Savage gave it that nudge with a beam and a block of ice. I would give it a catcher. A tug, already in orbit, meets the pod at the top of its arc and does the last little bit of work. That is a separate article, because it is a separate machine, and so is the question of what the tug does with the pod after it catches it and how the empty pod gets back down to the base of the mountain to go again. The whole loop, from the ship at the dock to the tug in orbit and back to the dock, is what a trucker would call a backhaul, and it gets its own piece.

What it would take

I am not going to pretend to know what this costs. Nobody does. A twenty-mile tunnel is not exotic: the world has longer rail tunnels and they were dug with machines you can lease. The magnets are not exotic: they are the same things that levitate trains in Shanghai and Japan. The vacuum is not exotic, though holding it across twenty miles with a door at one end that opens every ten minutes is a real problem, and I do not have the answer to it drawn yet. The power is the hard part. Eight seconds at the draw it takes to put a ton on orbit is more than the island's grid can give you, so you bank it: capacitors, flywheels, something that fills slowly and empties fast. That is engineering, not magic. Somebody has already done it at a smaller scale, because SpinLaunch in New Mexico has been slinging test vehicles at five thousand miles an hour off a spinning arm and putting NASA hardware through ten thousand gees to see if it survives, and it does.⁹ They are a long way from orbit and they say so. But they are doing the work with private money, and every test they run is a number I do not have to guess at.

I am a builder, not a physicist, and I have laid out what I have laid out to be knocked down. If you can show me that the heat cannot be moved, or that the door cannot be built, or that the pod cannot survive the air, I want to hear it, because then I know what to draw next. What I will not accept is that it cannot be built because nobody has built it. Nobody had built Biosphere 2 either. Nobody had built a mass driver in 1977 until a few students did it out of the scrap pile, and nobody had built one in my shop until I did.

Twenty miles of tube. Out the side of the mountain. Freight only, and a train that never stops running. That is the plan.

Ideas in Star Bound are published open — take them, build them better. Our open proposal.

Comments

Star Bound is not a forum, so comments run the old way: if you have comments, please write to us at hello@starboundmagazine.com and they will be published here in the next update, with your name on them. Say if you would rather stay anonymous.

No comments posted on this one yet. Yours could be the first.

Share this article

Copy link Facebook X Email

A note on where this comes from. I am not a metallurgist, a spacecraft engineer, or a degreed engineer of any kind. But I am a practical one, with more than forty years in heavy industrial construction: welding, millwright work, hydraulics, mining, and thirty-three years of electrical construction and commissioning on the North Slope of Alaska, in a place that punishes anything built wrong. Much of that time I have worked as a field engineer, which means I am the one standing in front of the equipment when the drawing turns out to be wrong, and the fix has to be made on the spot, with what is on hand, and has to work the first time. Anyone who has done that long enough learns that a lot of engineering gets corrected in the field by people with more experience than the people who drew it. That is the eye I bring to these subjects: not what should work on paper, but what actually gets built, and what breaks. The facts I take from other people are footnoted so you can check them and read further. The ideas built on top of those facts are my own, and I would rather see them knocked down by somebody who knows better than left standing because nobody looked.

Sources

1. Mass Driver 1 (1976–77), built at MIT by O'Neill, Kolm, and students, reached about 33 g; later Space Studies Institute prototypes reached several hundred g, with SSI reporting up to 1,800 g. Gerard K. O'Neill, Wikipedia; "Mass Drivers: From the ones who actually made the real ones," Space Studies Institute. ↩
2. The lunar mass driver concept and its purpose in O'Neill's plan: "Mass Drivers I: Electrical Design," NASA SP-428, Space Resources and Space Settlements (1979), via National Space Society. ↩
3. Editor's arithmetic. Acceleration to reach velocity v over distance d is v² ÷ 2d. For 8,000 m/s over 32,000 m that is 1,000 m/s², roughly 100 g, for 8 seconds. Higher exit speeds to cover atmospheric losses raise the figure in proportion to the square of the speed. ↩
4. Marshall T. Savage, The Millennial Project: Colonizing the Galaxy in Eight Easy Steps (1992; 1994 edition with an introduction by Arthur C. Clarke). The Millennial Project, Wikipedia. ↩
5. Bifrost as described in the book: maglev track under the Serengeti, up through Kilimanjaro, exiting at the summit, with laser boost on an ice block. "Reviews of The Millennial Project," Curriepedia; Bifrost, Living Universe Foundation. ↩
6. The successor project's assessment that the mountain launcher "proved infeasible" in its original form. The Millennial Project 2.0: About; Bifrost, TMP 2.0 wiki. ↩
7. Mauna Kea summit 13,803 ft (4,207 m); Mauna Loa 13,679 ft (4,169 m); the Big Island lies near 19.5–20° N. Mauna Kea and Mauna Loa, Wikipedia. ↩
8. Mauna Kea's summit is sacred in Native Hawaiian tradition, and the Thirty Meter Telescope has been the subject of extended protest and litigation since 2014. Thirty Meter Telescope, Wikipedia. ↩
9. SpinLaunch's Suborbital Accelerator at Spaceport America has completed ten flights at up to roughly 5,000 mph, carried NASA payloads, and qualified hardware at 10,000 g; the orbital system was still in development as of 2026. SpaceNews, Jan. 2023; Interesting Engineering, Oct. 2025; Futura-Sciences, May 2026. ↩
10. The 1994 mass driver, the flank exit, the deep-rock alignment, the closed-loop coil cooling with an underground chiller, the Hawaii siting, and the orbital catcher in place of laser boost are my own proposals and are offered to be argued with.