Drop Tube Elements

[Enthalpy]

Hello everybody !

Some experiments (clocks, atom interferometers, tests of satellite parts...) must run under zero gravity. Among the methods:

• Space station - Back seat on a satellite - (Small) own satellite
• Seat on a launcher's side booster - own sounding rocket
• Parabolic flight aeroplane
• Drop tube.

A drop tube provides among the best microgravity quality if not the longest. It's up to 150m tall, evacuated to eliminate the air drag, and the experiment drops through it for 5s, or 10s if it's thrown upwards from the base
http://en.wikipedia.org/wiki/Drop_tube
http://de.wikipedia.org/wiki/Fallturm_Bremen
that's unusual engineering and I hope you find it puzzling too.

----- Vacuum tube -----

Vacuum vessels are common, but not of this size... 110-140m long and I suppose 4m diameter. Wiki tells it's separated from the surrounding tower to decouple the wind force, and made of steel.

Aluminium extrusions, assembled by welding, can make this tube as well
http://en.wikipedia.org/wiki/Extrusion
A tailored die costs few k€, the preferred alloys resist corrosion well and are easy to weld. Profiles are commonly 15m long.

At 83t for 140m, it's lighter than steel and possibly cheaper.

----- Shuttle -----

It may host varied experiments, protect them from the brake hardware, keep the experiment in high vacuum outside the drop tube...

Here are shuttle shells made by diffusion bonding followed by welding, similar to
http://www.scienceforums.net/topic/73798-quick-electric-machines/#entry827449

 

It makes individual shell elements more expensive, but nice to assemble in a smooth three-dimensional shape. A cylindrical shuttle instead would suggest extrusions again, here an example weighing 59kg for D=0.8m L=2.4m:

 

----- Components cooling -----

Standard satellite integration must be helpful to cool electronic components in vacuum and hold the circuits when braking:

 

On my satellite's transmitter, I soldered on the printed circuit a 1mm copper plate, photoetched the same way, to evacuate heat to the sides instead. Other methods include heat pipes within the boxes and at the chassis, for bigger power.

Marc Schaefer, aka Enthalpy

[John Cuthber]

What is puzzling?

[Enthalpy]

----- Windshield tower -----

The decoupled tower that surrounds the vacuum tube has, in Bremen, about 10m diameter of concrete, which must weigh many 1000t and need long construction work at height.

A first alternative would use a decommissioned mine shaft, the mine itself, or caves.

An other alternative makes the tower of aluminium extrusion (or of a covered truss): much lighter, possibly cheaper, assembled horizontally. Among many, the following design needs 115t of aluminium and 830t of ballast which can be a part of a surrounding building, to break at twice the force of 240km/h wind:



The windshield tube and the 2.5m diameter legs have their wall of the same extruded profile:



Marc Schaefer, aka Enthalpy

What is puzzling?

 

Thanks for your interest!

 

I find puzzling that people build 150m tall towers to get few seconds of microgravity - the experiment must be designed around this constraint. I also imagine that many experiments measure subtle and faint effects but decelerate from 160km/s within very few meters. I like also the two 150m tubes in an other, hold only by their basis.

[Enthalpy]

----- Put the shuttle in vacuum -----

At least in Bremen, the whole vacuum tube is evacuated from 1atm to 10Pa before each experiment. This takes much power and time.

A first alternative would keep the tube void and introduce and remove the shuttle through a siphon:

For instance vegetable oil has a good vapour pressure. While economic, quick and reliable, this method wastes height and is dirty.

----------

The other alternative is of course an airlock. Maybe with doors that slide to the side and are pressed upwards by a separate hydraulic or electric actuator; a hydraulically inflated seal ring is natural in this context.

I described an airlock with saving tanks, as inspired by water saving basins
http://www.scienceforums.net/topic/78116-airlock-with-saving-tanks/
and here's a sketch for the drop tube:

The saving tanks are evacuated in advance to a logarithmic scale of pressures. To evacuate the airlock, it's allowed to discharge in the saving tanks of decreasing pressure. This saves time and peak pump power.

One mode of operation that saves energy takes air from the still not re-evacuated tanks when the airlock is refilled. This evacuates some tanks for free, but not all: for a given pressure step in the airlock, the tank that refills it has a pressure 2-3 positions higher than the tanks that evacuates it. Here from 100kPa to 10Pa it still looks interesting.

From 100kPa to 5Pa with 9 tanks (only 5 sketched), dividing the airlock pressure by 3[super]9[/super], the relative pressure could be at each step:
Begin with 3.0*P in the airlock and 0.6*P in the 5 times bigger tank, end with 1.0*P in both.
The total volume of the tanks equals 45 times the airlock, not little. Since the discharge process is isothermal only in books, air should be circulated between the airlock and a tank once the pressure is even, both to evacuate and refill the airlock. Limited temperature swing is an advantage of the saving tanks.

Marc Schaefer, aka Enthalpy

[Enthalpy]

----- Catapult -----

The drop tube at Bremen can catapult the shuttle upwards from the tube's bottom in order to double the microgravity time. A pneumatic jack of direct action is uncomfortable, for instance 10m stroke and 0.25m diameter. Building it lighter than the shuttle, and braking it smoothly enough at the stroke's end, must have been a challenge.

Bungees are a viable alternative. I used some to propel 60kg at 120km/h (and a kg much faster
by accident) and the extrapolation to 300kg and 180km/h is reasonable - I didn't claim: "very small".

Mines were half a dozen pairs (or better, U shapes), as thick as a thumb, ~60m long stretched, possibly 20m loose, and gave their energy in the last 10m essentially. Extrapolation here would need some 70 pairs if keeping this diameter.

For the microgravity catapult, I'd stop them with ropes, still stretched after 10m stroke. They can run in smooth metal rings for all their length but the last 10m to keep them out of the way and help stop them. The ropes can pull them to the side, out of the falling shuttle's way.

Bungees are non-linear, non-reproducible, and vary a bit over time. The use must adapt to individual characteristics given by the manufacturer. I'd provide anchor points moved by chains or screws, adjusted individually for each bungee pair. The actuator that stretches the last 10m then measures the force over the way to determine the precise starting point.

Using the bungees to brake the fall is tempting, provided something prevents the rebound.

Drawings should follow.

Marc Schaefer, aka Enthalpy

----- Free floating -----

Instead of evacuating the whole drop tube, we could transport a vacuum bubble around the falling experiment without touching it. Aircraft in parabolic flight achieve a better microgravity by releasing the experiments and let them float in the cabin's air. Floating in vacuum would improve.

Air drags the bubble's shell a lot. Even a 300kg unaided shell with Cx=0.17 would move by 2-3m during the way up relative to the free floating experiment inside, so this is an option for a much heavier drop tube rather. At present sizes and masses, it's better to compensate all drags - but thanks to free floating, this doesn't need much accuracy: the bubble's shell can move around the experiment if it doesn't touch it.

I wouldn't let the bubble shell move unguided in the tube. Among others, I know no safe means that keeps the aerodynamical stability in two directions. But for crash tests, polyurethane wheels in rails guide a test car to 120km/h, so 180km/h must just be tried. After all, their manufacturer refuses to give any information over 5km/h anyway . The rail must be reasonably straight and smooth.

For parabolic flights, the pilot steers the aircraft around a free floating item to keep it centered in the cabin. Though, I feel this too difficult for a machine and would rather have accelerometers at the bubble shell and control its speed and position to zero g.

Drawings should follow.

Marc Schaefer, aka Enthalpy

----- Drag compensating engine -----

It must work at +50 to 0 to -50m/s, provide an estimated +300 to -300N, but survive the initial 20g and 120kN if it doesn't provide them: this combination is challenging.

A steel cable does it for crash tests, up to 120km/h, with enough position accuracy. The cable can stop within some 10m (compare with the catapult) and is about as heavy as the car; it takes big wheels, a strong pre-tension, strong brakes and so on. This would provide noth the initial speed and the free float control. While maybe feasible, I don't prefer this solution: it worked after a long development, doesn't fit the 10m acceleration naturally, and is oversized for the free float phase.

Rocket engines expelling lukewarm nitrogen, carbon dioxide... are more or less possible. They need much propellant, like 50kg, plus the containers. Not convenient.

A linear electric motor would work. I'm not quite sure all experiments accept its magnetic leaks.

----------

The bubble's shell has wheels at a guiding rail. These can propel the shell. Mean 50% of 300N over 150m need 23kJ, which one 10kg 200m/s flywheel per direction can provide through well-controlled clutches: electromagnetic, hydraulic, possibly adapted to a gas. The wheels need enough pre-compression to have 150m/s at the catapult's exit. Hard teethed wheels might be an option though unpreferred.

----------

The rail that guides the bubble's shell can have slotted tubes. The bubble's shell plunges to a controlled depth one of two anchors that deviate air moved at 200m/s in the slotted tubes by fans immobile in the tower. Air can climb in one tube and fall in an other, with the anchors scooping some from one to deviate it by 180° and possibly inject it in the other, similar to a Pelton turbine. Deviating 1kg/s with 0.5dm² scoop section would suffice.

The air can have a smaller side speed in the slotted tubes in addition to the vertical speed. This side speed may or not be a swirl movement within the tube. The side speed helps to keep the flow in the slotted tube if the tube's lips have slightly different heights; it also helps to inject the quick air in the tube and evacuate it. I expect several fans must spread along the track; possibly they rotate just over 200m/s and let the air turn by 180°, taking it from one slotted tube and injecting it in the other - many arrangements exist.

A similar Pelton blade could deviate a liquid jet instead as the catapult, but how dirty.

Drawings should follow.

Marc Schaefer, aka Enthalpy

Edited October 1, 2014 by Enthalpy

[Enthalpy]

One more drag compensation engine:

A turbine can rotate a wheel, directly on one shaft. Energy comes from pressure gas in a tank travelling with the bubble's shell.

About 1kg air at 12bar (0.4m sphere) can pass through a small De Laval turbine. Air expands to 1bar 150K 550m/s, the turbine mean blade diameter is twice the wheel diameter hence runs at 100m/s, air exits at 350m/s which exploits ideally 59% of its enthalpy. For 300N and 50m/s at the wheel, 0.17kg/s suffice.

Alternately, nitrogen at 350bar (210 bar for servovalves) can push a liquid (hydraulic oil) out of a bladder. A nozzle to 1bar makes a 300m/s jet. The Pelton turbine has the wheel's diameter, so for 300N and 50m/s at the wheel, the liquid exits the turbine at 200m/s, exploiting ideally 56%, and 0.6kg/s liquid suffice, again <1kg for the complete trip.

The hydraulic option is more compact and I expect less leakage worries. Have one engine per direction.

Drawing, yes...
Marc Schaefer, aka Enthalpy

[John Cuthber]

You might want to compare

the cost of the electrical energy used in pumping down the system to a good vacuum

and

the cost of the night watchman.

 

I suspect it's not worth worrying very much about the cost of running the pumps.

The saving tanks would be roughly as expensive to produce as the main tower.

 

Siting it in an old mine shaft would be a sensible idea- unless there just don't happen to be any mines near the University. It's a lot easier to build a tower than to dig a well.

 

You have pointed out that there are 4 ways to get microgravity.

All 4 get used.

You say "I find puzzling that people build 150m tall towers to get few seconds of microgravity"

I'd be a lot more surprised if we didn't.

We do a lot of very much more expensive and difficult things that that for the sake of research.

Why not have a drop tower? It's cheap (in comparison to a lot of research equipment..

[Enthalpy]

Here are drawings for the drag compensation engines.



I'd have the two flywheels for up and down trip at different locations on the shuttle. A gear can power also the second propulsion wheel. Permanent force against the part of the rail gives grip. The motor that accelerates the flywheels prior to the trip can be on board, the energy source not necessarily.



This is the liquid version; for a gas, imagine a De Laval turbine. The used liquid lows to a container. The pump to the bladder accumulator can be outside the shuttle.

Yes, a jet of liquid exiting the shuttle could have propelled it without a turbine nor motor wheel. Yuk!



The rail is at left and the shuttle at right on this sketch. The shuttle plunges its scoop more or less deep in the wider tube through the slot to control the propulsive force. More sketches should follow for this scoop engine.

Marc Schaefer, aka Enthalpy

[Enthalpy]

You might want to compare

the cost of the electrical energy used in pumping down the system to a good vacuum

and

the cost of the night watchman.

 

I suspect it's not worth worrying very much about the cost of running the pumps.

The saving tanks would be roughly as expensive to produce as the main tower.

 

At Bremen, they pump to vacuum before each shot, and this takes 1.5 hour, limiting the repetiion rate. The pumps are already big: 32 000 m³/h. More frequent shots are a hot desire there.

 

The drop tube sits anyway in a building, and a vacuum tank can stay unwatched as much as oxygen bottles are presently.

 

The main tower consists only in part of a vacuum tube. You must add the wind shield tube, the experiment catcher and catapult, the instruments, and so on. Just by permitting the pumps to run with a proper pressure ratio (a single centrifugal stage gets possible) one saves on acquisition and electricity costs.

[John Cuthber]

Essentially, all research is very expensive.

Simplicity (and hence reliability) may be much more important that absolute efficiency.

[Enthalpy]

This is a way to blow air upwards in a slotted tube - the downwards tube is reversed, and both can close the air circuit. This especially advantageous embodiment circulates the scoop where the fans are not, which avoids destroying them all.

Nevertheless, the scoop gets good wind thanks to a whirl in the tube, so the air alternates between the fans and the slot zone. The first cause is the fan rotation; blades at the stator can adjust it, as well as tilting the fans up and down.

All fans in a tube could be moved by two shaft lines, or have several motors. One tailored AC line for many squirrel cage motors looks simple, but noise may demand to synchronize all fans.

The shuttle can have several scoops. Their spacing, the fans spacing, and the whirl geometric period better have an optimized relation to smoothen the operation.

Marc Schaefer, aka Enthalpy

Essentially, all research is very expensive.

 

Discoveries made by chance tend to be cheaper than oriented research. For instance, the synthesis of nitrocellulose by Schönbein cost only an apron - and difficult explanations to his wife probably. But is this an argument favouring random research?

[Enthalpy]

An other layout that blows below the slot for a drag compensation scoop. This is an axial view; the air speed is mainly axial, that is, through the screen.



The blowers are skewed to achieve the whirl - or use stator blades for that.

Marc Schaefer, aka Enthalpy

[Enthalpy]

----- Bungees as catapult and brake

Arresting an object smoothly from 180km/h without destroying it nor any brake part is very difficult. Bungees may have this capability to restart a flight immediately, and then brake more gently than crushed material. Combined with the scoop drag compensation that isn't limited by energy storage, they permit long dense sequences that cumulate hours of microgravity for experiments wanting to accumulate data. That's worth trying hard, which includes much experimenting.



A ring spreads the bungee force on the shuttle. Made of woven polymer fibre (aramide and so on), it catches the guided shuttle with a limited shock and hopefully no degradation. Polymer lines to the sides and down hold it centered and well open despite the wind when it awaits the shuttle; stoppers at the main bungees instead are doubtful.

More polymer lines make a rake equivalent that hold the ring and shuttle at the low point after braking, until the shuttle is brought to the proper height for the next operations. The lines rub for several turns over immobile parts, so they move only if their trailing end is loose. As the falling shuttle sinks the ring, it loosens the lines which are pulled at good pace by smaller extra bungees, whose force suffice to keep the ring in low position thanks to rubbing. Slow and accurate hardware then takes control, brings the shuttle to position, and the lines recover the wait length by loosening the bungee end combined with some positive action.

It could also use a standard rake of hard material. Impossible to operate at 180km/h, this one must come in contact with the shuttle when the speed is low. It needs active decisions, but failure here isn't destructive.

Individual bungee adjustment was already mentioned, as well as summing the force over the catapult stroke. Descriptions of the railway and the shuttle should come.

Marc Schaefer, aka Enthalpy

[Enthalpy]

I want a railway to guide the shell that carries the experiment floating in vacuum... Here's a railway's truss that supports the vertical 150m railway and is hold at its bottom, to decouple it from the windshield tower.

Though, I'm not quite sure that the railway and the bubble need decoupling from the windshield tower, especially if flights start only from ground level. The tower moves little in reasonable wind, <1m for 240km/h hence <0.1m for 80km/h and still improvable, and the experiment moves freely in the bubble. Fastening the railway to the windshield tower would gain stiffness and simplify the design.

Anyway, here's a truss design that looks possible. It keeps 1m radius clearance with the windshield tower and hosts a D=2m bubble. This is a horizontal cut, the kind reader must imagine that some beams are inclined.



Stress on the railway is hard to assess, so I've taken arbitrary beam cross-areas - warning - but at least these give pleasant resonant frequencies and are individually safe from buckling. The material considered is S355 steel; the truss resists 0.2g earthquakes, and AA7020 aluminium or a wider tower would improve that.

The outer eight columns are OD=178mm t=5mm round tubes, the inner four W=140mm t=5mm square tubes. The horizontal and biased beams are OD=70mm t=4mm round tubes. The truss has 3m high stages and weighs around µ=560kg/m, totalling 84t.

The truss' inertia is 0.17m⁴. The first flexural mode, holding only the base, corresponds to k*Z=1.88 so Z=150m injects the wave vector k=0.0125rad/m in the wave equation
EI*k⁴=µ*w²
from which the truss resonates at 0.20Hz. At this lowest frequency, a flexural wave propagates at 101m/s, exceeding the bubble's maximum 50m/s, which I believe is important to keep the lateral position stable as the bubble falls down.

The truss' shear compliance is, with only two inclined beams resisting in each direction, 35nm/N for each 3m stage or mean 12*10^-9/N for the truss, so shear waves propagate at 380m/s and the first shear mode, holding only the base, is 0.65Hz. Torsional modes are faster.

The 3m free column lengths can vibrate as half-waves with k=1.05rad/m. The 8*10^-6m⁴ and 21kg/m square tubes resonate at 50Hz where they propagate a flexural wave at 300m/s.

Marc Schaefer, aka Enthalpy

[John Cuthber]

I suspect that the reason why people use drop tubes is their simplicity. An empty tube is a very easy concept to grasp and a fairly easy one to engineer.

Your increasingly complex design reminds me of this cartoon

http://pages.uoregon.edu/ftepfer/SchlFacilities/TireSwingTable.html

[Enthalpy]

Thanks for your interest!

 

A 150m tall and 10m wide vacuum tube is simple before it's developed, built and operated. Once you give some engineering thoughts at it, it gets nasty. Certainly worse than the railway truss, and I'd say (...until the serious engineering worries pop up...) worse than the approximate drag compensation. Remember I consider a free-floating experiment, so the enveloping vacuum bubble doesn't seek high-quality microgravity, and without compensation at all it would drift by only 2m.

 

For instance, the tube at Bremen takes 1.5 hour to evacuate. I still ignore the operation details, but the free falling experiment is braked by crushed material. This makes it difficult to catch and put the experiment in place for the next flight - maybe it happens in vacuum, maybe it need humans.

 

As opposed, a guided vacuum bubble can be caught automatically (it's really difficult at 180km/h!) and be ready for restart in less time than a flight. This means that experiments can accumulate data for most time. Some experiments can take advantage of that and do need much flight time, for instance atoms interferences. <10s reload and 10s flight alternating for one day would be hugely better for them than 10min or 1h reload alternating with 10s flight.

 

I haven't put figures yet, but it's possible that the vacuum bubble that moves with the experiment gives a better microgravity than the experiment falling through a fixed "vacuum", as the remaining air has zero point little relative speed instead of 50m/s.

 

To me, the mobile vacuum bubble looks like a game-changer that brings a drop tube radically better capability.

----------

 

Maybe I should have made clearer that not all the messages here apply to to the moving vacuum bubble.

 

Many are for an evacuated drop tube: the extruded walls, the airlocks...These would not add their complexity to the moving vacuum bubble.

 

A catapult exists already, at least in Bremen, to double the fight time - same complexity at the moving vacuum bubble. And if bungees double-serve as brakes, the combination can be simpler, especially as they permit a quick restart.

 

So a guided mobile vacuum bubble adds the bubble, the drag compensation, the railway - but saves the big vacuum tube and its airlock or huge pumps.

Edited October 20, 2014 by Enthalpy

[Enthalpy]

Here are some hints about how to guide a shuttle - for a vacuum bubble surrounding a free-floating experiment, not for an experiment falling in a vacuum tube.

Plain bearings are doubtful at 180km/h, so I consider polymer wheels instead, especially polyurethane that absorbs small bumps and misalignments. They're good at 120km/h; azimuthal or radial fibres as in tyres could reinforce them if needed.

The sketches, side and bottom views, are not to scale:

The wheels roll against rails separated from the truss to permit a finer alignment. Pouring some resin or concrete between the rails and the beams would bring stiffness if needed. This arrangement of wheels and rails lets clearances depend little on the truss' dimensions; a pre-load is possible.

For quick repetitive operation, an electric linear motor as well could compensate the drag, but I've sketched the scoops and the wind channels instead, here with still another arrangement of the fans. Small hydraulic turbines could rotate the fans, as an alternative to the means already suggested.

Marc Schaefer, aka Enthalpy

[Sensei]

Vacuum vessels are common, but not of this size... 110-140m long and I suppose 4m diameter. Wiki tells it's separated from the surrounding tower to decouple the wind force, and made of steel.

The easiest solutions are usually the best one - why not put drop tube in old unused mine, or oil drill. They can be even a few kilometers long, but below ground. You don't have to worry at least about wind force.

 

Standard satellite integration must be helpful to cool electronic components in vacuum and hold the circuits when braking:

To cool down electronics, I would try surrounding them by some heat conductive (but electric not conductive) liquid.

[Enthalpy]

Earth's rotation lets the free-floating experiment drift during the flight, hi Coriolis. The vacuum bubble must offer enough room for this drift. But how much?

At arbitrary 53° North, the ground rotates at a=43.8µrad/s around the meridian. The bubble's eastward speed increases as height*a, while the experiment's eastward speed decreases as height*a, because the experiment's angular momentum versus Earth's center is conserved:
http://en.wikipedia.org/wiki/Kepler's_laws_of_planetary_motion

With these axes orientation and origin

the difference of east-west speed between the bubble and the experiment is

when the flight starts from the ground - this matters. Integration from t=0 to the half-flight duration T gives the half-flight westward drift X

or 2X=97mm for Z=150m.

----------

This Coriolis drift can be lindered against varied criteria. For instance, the experiment would land at the center of the vacuum bubble if it takes off with 9mm/s eastward, and would then drift most near apogea. Some mechanism could pull the experiment gently to the side during the acceleration, or more reliable, the railway could be permanently tilted by 1.7mm over the lower 10m.

This, of course, if no other cause overshadows the 97mm and 9mm/s...

Marc Schaefer, aka Enthalpy

[Enthalpy]

[...] why not put drop tube in old unused mine, or oil drill. They can be even a few kilometers long, but below ground. You don't have to worry at least about wind force.

 

Thanks for your interest!

 

A decommissioned mine shaft looks an interesting option. I easily agree with it because I suggested it here on Sept 30, 2014. I checked the coal mines in Germany, where the last ones are closing, and they're commonly 600m deep, which would double the flight time. Difficulties:

• People willing to conduct microgravity experiments are not always where abandoned mines are.
• At least in the densely populated Germany, the shafts are filled when a mine is abandoned.
• Many shafts are tilted. Are the vertical ones straight and broad enough? I haven't found accurate descriptions of that.
• At many places, mine shafts use to fill quickly with water and stones.

A plan of a cave appeared on the Internet few days after I suggested here on Sept 30, 2014 to use a tall cave or mine...

• A natural cave >150m tall is a natural marvel. I wouldn't damage it.
• Some mines can be really tall. Maybe salt mines (but natural and manmade marvels again), diamond mines in South Africa (miners let the ceiling collapse to gather ore), some mines in Canada, and more.
• How dangerous the place is remains to check...

I like an oil well less for the symmetric reason...

• The bore is narrow and usually skewed.
• Does it withstand the pressure of stone and water? I believe not, even after cementing. Oil-well operation takes great care to keep the proper mud density in the well to avoid its collapse.
• The new user would have to separate the bore from the oil or gas reservoir and empty the bore if feasible. Oil-well operations are expensive.

A tower isn't difficult nor very expensive with the materials I suggested, and can be built about anywhere. It can exceed 150m easily, and if aviation dislikes that, the tower can be built from a pit. Finally, a mine or mine shaft may be better or cheaper, but it depends much on a lucky finding.

[John Cuthber]

Pumping the water out of a mine shaft (an ongoing cost) might take more effort than building a tower.

[Enthalpy]

To cool down electronics, I would try surrounding them by some heat conductive (but electric not conductive) liquid.

 

It has been done, for instance at one old Cray supercomputer immersed in liquid nitrogen. This method is avoided whenever possible:

• It's dirty. Commercial components are not tested to work in a liquid.
• Electronics engineers badly want to access the boards, and even the components, when it's in operation at its final position with its final connections. Mandatory to assess and correct electromagnetic worries. Already the integration as a rack is bad and leads to boards that work approximately for being difficult to observe and tinker.
• As opposed, the satellite-type integration cools the components well enough for a computer and offers a decent access.
• Where needed, heat pipes are more efficient than a liquid bath and are clean. They suffice for the worst processors and even for radio transmitters on satellites.

The method presently, for supercomputers too dense to be cooled by air only, is to conduct heat away from the components through metal to pipes where a liquid flows actively. I suggest elsewhere to develop long branched alkanes similar to phytane for that purpose; they would have more uses, like transformer oil, hydraulic fluids and lubricants for aeroplanes and vacuum machines.

[Sensei]

Pumping the water out of a mine shaft (an ongoing cost) might take more effort than building a tower.

 

I was just thinking about reusing already made hole for mine's elevator.

http://en.wikipedia.org/wiki/Zero_Gravity_Research_Facility

I don't see any water.

Hole should be properly sealed from either side.

Edited October 26, 2014 by Sensei

[Enthalpy]

The slightly tilted lower portion of the railway, suggested on Oct 25, doesn't work. Changed my mind for the (2N+1)th and hopefully last time. But pushing slightly the experiment eastwards at take-off within the vacuum bubble would.

----------

This is how the free-float within the vacuum bubble could look like. Only a suggestion needing more ideas to be a viable concept. And since the experiment shell is only 1m wide for 0.2m radial free-float margin on this sketch, I'd add 1m to every radius up to the windshield tower.

During gravity, acceleration and braking, the experiment shell rests by cones or cone segments on a catcher that holds at the vacuum bubble. The experiment shell takes off the catcher by ~0.2m after the acceleration, for instance because the scoop control let the bubble be late. Few meters before braking, the scoop control slows down the bubble to land the experiment shell in the catcher.

The catcher is stiff vertically but moves easily to the sides. For instance a suspension under 1m solid steel wires gives at 15g a 0.5s period, longer than the complete acceleration, so that shocks at the railway during the acceleration are filtered out, and the experiment shell starts its free float with a precisely vertical speed.

When landing excentered, the experiment shell hits only the light catcher, which can have lateral rubber bumps in the vacuum bubble beyond its free movement, so the contact shock at the experiment is small.

Instead of hanging under wires, the catcher could roll on two-dimensional ball slides.

The catcher could have lateral actuators, and then optionally be narrower and follow actively the experiment shell. This would reduce the lateral acceleration at landing, as the catcher would center the experiment after braking, and could also linder the Coriolis drift by acting during the acceleration. It would also gain diameter at the shell. It needs very soft actuators, possibly through rubbers, to transmit no shock from the railway. The catcher needs lateral dampers if not actuators.

The catcher could have a vertical actuator; for instance ball screws are fail-safe. Then, the catcher can actively move down to begin the free-float phase and move up to land the experiment shell. This reduces the vertical shock at landing.

More active catcher designs are possible, but they must be very stiff to impart no side speed to the experiment shell - metal cones are good at that - and must insulate the shell from the railway shocks. The experiment's angular speed after take-off is also critical, like 0.1°/s, for good microgravity - again in favour of metal cones. For 1000 flights a day, the catcher must behave gently when its control fails.

I'd rather leave some clearance between the wheels and the rails over most of the track. Though, the acceleration and brake segment can have thicker rails if made of single parts, in order to suppress the play there. The thickness transitions must be very long for 180km/h. I suppose the wheels can have a suspension.

Several experiment shells, possibly several vacuum bubbles, accelerate the experiments turnaround.

Marc Schaefer, aka Enthalpy

[Enthalpy]

Here are hints for some components.

---------- Rake

I mentioned on Oct 06, 2014 a rake - difficult design at 50m/s. Maybe hard parts with proper wedge shape can rub, near wheels and rails where they would thread in more easily, and using a long slope to linder the contact shock which shouldn't be underestimated. Or these part are put actively in contact when the observed speed has dropped.

Instead, the following sketch uses the rubbing of the ropes that retain when desired the catching ring of Oct 06, 2014.

On such a winding, similar to a boat's winch, the rope glides if the pulling force exceeds exp(µ*a) times the holding force. If a=1.5*2pi radians and the coefficient of friction µ=0.25, the amplification exceeds 10.

As the shuttle pushes the catching ring down, it releases this end of the rope, so the weaker extra bungee at the other end can move the rope untils it's stiff at the low position; there, the extra bungee aided by the multiplication factor holds the ring until a separate mechanism brings the ring and the shuttle to the proper restart position.

Before restart, the rake bungee is actively released, the rope actively brought to the wait position, and the rake bungee stretched again.

This design isn't easy to adjust, is not fail-safe, but it may work at 50m/s.

---------- Belt

The vacuum bubble and the experiment shell must be hermetic but quick to open and close. Better than numerous bolts, they could use a belt as the ones that hold satellites on launchers, here a cut of the right side:



The jaws' material must glide well at the lid, cylinder and belt, like polymer bearing material against Cr. A hand wrench can tighten an M16 screw (100kN) to stretch the belt, which then clamps the lid with ~200kN: enough for vacuum inside, enough for air inside up to D~1.3m. As special screw achieves more, but at some point a hydraulic closure is better.

---------- Release mechanism

Already proposed there
http://www.scienceforums.net/topic/58924-magnetized-target-fusion/
it may release an experiment falling free through vacuum and minimize the side kick. The guided vacuum bubble doesn't need that.



350bar on a nickel membrane rubbing over D=20mm h=40mm on a nickel-coated cone hold safely 1t. 1cm³ releases the pressure much faster than the 5ms to open the valve. The pressure equalizes in 20µs so it's <<0.5% even. The <<0.7N*s side kick lets a 1t*3m experiment rotate at <<1.3mrd/s or <<3*10^-7g.

Marc Schaefer, aka Enthalpy