[0.25]KSP Interstellar (Magnetic Nozzles, ISRU Revamp) Version 0.13

[RabidMonkey]

Yes, and the key is: How much is this fraction exactly? What does a nuclear detonation? It produces a lot of radiation(from heat to gamma) + some particles that where the bomb before + a decenct neutronflux. In the atmosphere, the radiation (and the neutrons) are superheating the atmosphere nerby, that causes the shockwave which do the "work".

But in space, there is no atmosphere.

That means, all radiation and neutrons and particles are going in every direction equaly. That means, even if the distance from the exploding bomb to the "bumpplate" is zero (which can't be), only the half of the "explosion energy" is hitting the plate. From this half, only a very tiny amount will be fast moving particles, that can actualy produce some sort of thrust. From these tiny fraction, only a tiny fraction will have the right angle to push the plate (the vast majory will be reflected without transferring kinetic energy). But the most energy of the detonation will be radiation + neutron flux. What does that with the plate? It heats (and activate) it! Yeah! (Actual, the vaporization of some plate material will provide additional thrust, but it will be also very tiny)

So yes, i don't think the "orion drive" can work as told. Simply because it don't transfer the energy it produce in some kind of reaction mass properly. And it don't direct the reaction mass properly. And undirected energy/reaction mass don't produce thrust. No matter how many energy you have.

You simply need a kind of nozzle, that help you to direct the reaction mass, and you need a device that can transfer a decent amount of your energy into the reaction mass. (and not in the vessel superstructure)

The entire body of the bomb is the reaction mass. For a 1MT detonation, the actual mass that has been directly converted to energy is less than 47 grams. If we take the W53 (America's highest-yield nuclear weapon ever in service - 9MT), then 418.6g of mass are converted to energy in the detonation. That warhead weighed in at 2800kg. That leaves ~2799.6kg which is now plasma moving at ridiculous speeds, because there is no atmosphere to impede it. In a vacuum, the speed of the plasma is proportional to the square root of the change in temperature of the debris. Since this is tens of millions of degrees in only a few milliseconds, the velocity is HUGE. The diameter of the pusher plate then becomes the determining factor for the specific impulse, expressed as what fraction of the plasma will hit the plate, which also depends on the distance from the fireball.

So, let's try this with a bog-standard W53 and a pusher plate 100m in diameter, with the explosion happening 1km behind the craft. I can't find any exact figures for temperature other than "tens of millions of degrees", so we'll go with a 15 million degree change as a nice lowball figure. The fireball radius is ~1.2km, giving a surface area of ~1.8e7m^2. Our pusher plate is 100m in diameter, so our fraction is .0004. I also can't find anything noting exactly what is the proportional factor for the temperature to plasma velocity calculation, so I'll lowball that at .01, giving us a plasma velocity of about 1.5e6 m/s, or about .005c. Looking at bhangometer photos of nuclear tests, this feels about right. This (Isp = (C * Ve)/g where C is the fraction and Ve is the plasma velocity) gives us a specific impulse of about 61s. Fairly inefficient in this particular design, but I'm worst-casing everything here, including design (plate much smaller than fireball, distance from fireball, etc). Plugging that into our handy-dandy thrust equation of F=Isp*Mf*g (specific impulse times mass flow rate times g), we get a thrust from a single W53 hitting a 100m plate 1km away of about 1.7MN. That's only 1 bomb per second on a design produced by idiots. Optimize the plate/fireball diameter ratio with a spherical explosion, and the fraction goes to .5. That gives us an Isp of about 76452.6s. Using the same bomb (W53) at the same distance (1km), we now get 2GN of thrust. From one bomb per second.

Atmosphere vs. vacuum has nothing to do with it. Atmosphere is actually worse, because you encounter air resistance and compression heating. Look up the Pascal-B test from Operation Plumbbob: A 900kg steel plate was accelerated to at least 66km/s by the blast. It's figured that it didn't actually leave the atmosphere because of compression heating. Strip away the atmosphere, and it would have been single-stage-to-interstellar-space - the Sun's escape velocity at Earth is only 42.1 km/s.

Nuclear Pulse Propulsion: Where bigger really is better.

[Jasmir]

The entire body of the bomb is the reaction mass...

Yes, i know. The problem is, the reaction mass can't catch all your produced energy. (To be honest, it even can't get more than a tiny fraction of the energy). So we have to put the explosion far behind the vessel to avoid to much varporization from the bumperplate.

So, let's try this with a bog-standard W53 and a pusher plate 100m in diameter, with the explosion happening 1km behind the craft. I can't find any exact figures for temperature other than "tens of millions of degrees", so we'll go with a 15 million degree change as a nice lowball figure. The fireball radius is ~1.2km, giving a surface area of ~1.8e7m^2. Our pusher plate is 100m in diameter, so our fraction is .0004.

I'am ok with that...

I also can't find anything noting exactly what is the proportional factor for the temperature to plasma velocity calculation, so I'll lowball that at .01, giving us a plasma velocity of about 1.5e6 m/s, or about .005c.

This may be true, BUT there will be a lot of internal collisions in the plasmacloud with produces energy (x-rays?) again and _lower_ this speed. So the interersting value is the overall expansion-speed of the plasmacloud in the moment it hits the bumperplate. And i think, this speed will a little lower, but i havn't found any examples for it.

Looking at bhangometer photos of nuclear tests, this feels about right.

I thought, a bhangometer is a non-imaging device? anyway....

This (Isp = (C * Ve)/g where C is the fraction and Ve is the plasma velocity) gives us a specific impulse of about 61s. Fairly inefficient in this particular design, but I'm worst-casing everything here, including design (plate much smaller than fireball, distance from fireball, etc). Plugging that into our handy-dandy thrust equation of F=Isp*Mf*g (specific impulse times mass flow rate times g), we get a thrust from a single W53 hitting a 100m plate 1km away of about 1.7MN. That's only 1 bomb per second on a design produced by idiots.

I don't think, that we can get a much better fraction. As i said: the excess-energy dictates a minimal distance to the bumperplate, and the weight & strengthness of steel dictates a maximal diameter of this plate. In this design, -given a thickness of 2meters- our stellplate will weight 123464 tons. This will accelerat the plate at 0.01 m/s. (Given, the mass of the other ship + bombs is zero)

Optimize the plate/fireball diameter ratio with a spherical explosion, and the fraction goes to .5. That gives us an Isp of about 76452.6s. Using the same bomb (W53) at the same distance (1km), we now get 2GN of thrust. From one bomb per second.

This means, our bumperplate will have now a spherical shape with a 2km diameter now. The thinkness has also to increased. I don't think, this constuct will perform better due the really HUGE mass now.

[Donziboy2]

Fractal_UK, your changes look pretty awesome. Will they be save breaking?

[Fractal_UK]

You shouldn't scale down the specific impulse by the fraction of the pusher plate surface area against the surface area of the explosion, regardless of what fraction of energy you are receiving, what does arrive comes with the same specific impulse. The fraction instead scales down the power available.

Assuming we have a 1 megaton bomb at 25km/s and 0.0004 of that 1 megaton (4.184x10^{15), that corresponds to thrust of 133,888,000N. Actually, since we've used an energy rather than a power, that is an impulse rather than a thrust and we should scale by the timescale of the explosion to get the actual thrust. In vacuum, this timescale would be very short so 1 second might be a reasonable guestimate anyway.}

^{Regardless, it would be a very efficient drive, I'd guess only the very best Interstellar engines - DT Vista and upgraded 3.75m antimatter drives would outclass it.}

^{Donziboy - no, changes won't break saves but, as always, new models may mean some ships need small modifications in the VAB to fit new attachment node positions.}

[Srilania]

Are we just gonna get teasers, or we going to see some new stuff available for playtime soon?

[Patupi]

Jasmir, admittedly I'm not that kind of engineer and don't know the maths involved, but from what I understood the mass of the pusher plate was mainly as a heat absorption mechanism, not structural integrity. I'm sure there would be a heck of an ablation if you use such a drive a lot to need thickness, but for structural strength a solid plate is not neccessary. In fact a lot of the mass of such a plate is the shock absorption mechanism that evens up individual pulses into something that won't shake the ship apart.

Also, I thought that the old designs included adding propellant around each bomb to increase efficiency of transfer of momentum to the vessel. This doesn't include the whole 'shaped charge' idea which generates a lobed blast, fore and aft less to the sides. True, you could never get 50% transfer even with a close blast, but with shaped charge you could start getting in that range. I was never sure how you 'lobe' a nuclear charge. It's hardly the same as chemical shaped charge after all which is vastly different in purpose and (I assume) mechanism.

Plus, nuclear detonations do not vaporize as much as people assume. 1km seems an awful long way away for any kind of blast, especially as most pusher bombs were rated as (comparatively) small devices at less than a kiloton. With weapons tests in the ocean, and defunct vessels at less than a kilometer from the blast (and much bigger yields than we're talking here!) the hulks retained their form surprisingly well, despite not being designed to withstand that kind of force. A lot is simply because, just as in an orion ship, they could move freely and would be accelerated by the blast, riding the shockwave at least a little. True, that was primarily against concussive force in the atmosphere, but that actually makes ablation worse I believe.

[RabidMonkey]

Yes, i know. The problem is, the reaction mass can't catch all your produced energy. (To be honest, it even can't get more than a tiny fraction of the energy). So we have to put the explosion far behind the vessel to avoid to much varporization from the bumperplate.

That's why the formula for calculating Isp from an Orion drive is different - it takes the wasted energy into account. The detonation doesn't have to be as far away as you might think - with no atmosphere, there's no overpressure wave or compression heating. The Project Orion designs called for sub-kiloton yield pulse units, detonating only 25m from the plate. The short duration of the thermal pulse at that range could be handled with a couple millimeters of copper backed with carbon.

This may be true, BUT there will be a lot of internal collisions in the plasmacloud with produces energy (x-rays?) again and _lower_ this speed. So the interersting value is the overall expansion-speed of the plasmacloud in the moment it hits the bumperplate. And i think, this speed will a little lower, but i havn't found any examples for it.

Plasma doesn't work that way. Unless there is a force causing collisions (like gravity or magnetism), it doesn't collide with itself. For the equation calculating Isp of a pulsed drive, the important value is the initial wavefront velocity of the plasma.

I thought, a bhangometer is a non-imaging device? anyway....

My mistake. I was thinking rapatronic camera, but for some reason bhangmeter came off of my keyboard.

I don't think, that we can get a much better fraction. As i said: the excess-energy dictates a minimal distance to the bumperplate, and the weight & strengthness of steel dictates a maximal diameter of this plate. In this design, -given a thickness of 2meters- our stellplate will weight 123464 tons. This will accelerat the plate at 0.01 m/s. (Given, the mass of the other ship + bombs is zero)

Like I said, that was an idiot-level design. But think about that - this horrible design can accelerate 123,464 TONS at .01m/s^2. Want to add crew and life support? Fine, add another 500 tons. Acceleration is still .01m/s^2. Want to carry 1000 cows along just for the hell of it? Let's add in another 10,000 tons so we can keep them alive for a while. Plus 10,000 tons of soda because we're thirsty. Acceleration is STILL .01m/s^2.

The solid steel pusher plate 2m thick is way overengineered for this. All that's really needed for the actual blast surface is a sheet of copper about 5mm thick that's coated in oil. With that, no plate ablation.

This means, our bumperplate will have now a spherical shape with a 2km diameter now. The thinkness has also to increased. I don't think, this constuct will perform better due the really HUGE mass now.

The plate is still either flat or slightly concave. I was talking about the shape of the fireball itself when I said "spherical". The best possible fraction with a spherical fireball is .5, and that's if the plate is the same diameter as the fireball itself will be at the detonation distance. Using the copper pusher plate as specified above, we get an acceleration of about 1.4m/s^2. With shaped charges, the plasma can be concentrated into a cigar shape, dumping more energy into the plate as opposed to empty space, giving us fractions greater than .5. Now we're looking at accelerations of over 1g.

The beauty of nuclear pulse propulsion is that there really is no practical limit to vessel size. Freeman Dyson was looking at multi-megaton vessels - basically flying cities. There was no technical reason why they would not have worked.

[Fractal_UK]

Are we just gonna get teasers, or we going to see some new stuff available for playtime soon?

I'm hopeful it won't be long now but the update is massive, so even though development has almost finished, tying up loose ends and testing is a massive job compared to previous releases.

[LORDPrometheus]

Remember Fractal_UK there are Plenty of people here who whould be happy to help beta-test release an alpha build with a built bug report line so you can discover major issues much faster

[RussPixie]

I'll just leave this here...

http://www.projectrho.com/public_html/rocket/enginelist.php#orion

http://forum.kerbalspaceprogram.com/threads/28428-Orion-aka-Ol-Boom-boom

[Patupi]

Just a query Fractal_UK , given I'm using 0.7.4 in my AAR at present. How likely is it that 0.8 will break saves that use KSPI and which parts are the most likely culprits? I haven't got to the really advanced stuff yet, but have just played through to the basic nuclear tech. Any of that likely to be a problem with upgrading to 0.8 Fractal?

[Donziboy2]

For those of you having issues with figuring out say the DT Vista's dV here is a little chart I slapped together. Not sure of a good place to upload the excel but the math is there so you can just grab a calculator:)

As you can see from the example a few tons of LF goes a long way, 15 units of Deut/Trit will last 3 hours. So you can set up for high thrust burns or carry more and do lower thrust burns with less LF need, im sure there is a point that it actually adds more mass then its worth.

IMAGE BURNED FOR POSTERITY, look 2 posts down for update....

Edited November 18, 2013 by Donziboy2

[RabidMonkey]

For those of you having issues with figuring out say the DT Vista's dV here is a little chart I slapped together. Not sure of a good place to upload the excel but the math is there so you can just grab a calculator:)

As you can see from the example a few tons of LF goes a long way, 15 units of Deut/Trit will last 3 hours. So you can set up for high thrust burns or carry more and do lower thrust burns with less LF need, im sure there is a point that it actually adds more mass then its worth.

Units? Labels? Something to make sense of what could just as easily be a bunch of random numbers?

[Donziboy2]

Bah, thought I forgot something....

Working on a burn duration calculator also.

Consumption rate formula is ((Mstart-Mdry) / 0.005) / Consumption Rate / 60 = Minutes

Mileage may vary 10% +-

Does github let us post excels?

Edited November 18, 2013 by Donziboy2

[Hremsfeld]

I was never sure how you 'lobe' a nuclear charge. It's hardly the same as chemical shaped charge after all which is vastly different in purpose and (I assume) mechanism.

Up until very recently, the fact that shaping a nuclear charge was possible at all was considered critical nuclear surety information and incredibly classified. It'd have pretty much the same purpose though, namely blowing something up a bit more efficiently with respect to how much of the blast is used.

Donziboy: I haven't really used the Vista yet since v0.6; does it consume Deuterium and Tritium at a constant rate, or do those fluctuate, too?

[AdmiralTigerclaw]

I've got a question.

I've finally picked up that KSPI operates on a power system disassociated from stock electrical energy.

As such, anything requiring MJ/MW/Thermal energy does not interact with existing forms of power.

For example, I can't power the brayton cycle engine with thermal energy accumulated in the solar panels.

So I raise a question. Considering that in reality (that scary place we don't like) energy converts to all forms, can we expect some form of 'converter' components to say, convert electricity into MJs or thermal power?

It seems kind of silly I can't get thermal power or MJs (even in low amounts) from a configuration like a 'solar ship'. I don't care if it's low-power, inefficient and wasteful, but I like to have self-sustaining orbital infrastructure that doesn't require me to ship up a new nuclear reactor for every little vessel design.

[Eadrom]

I also run into the issue regarding MechJeb firing early when using the plasma engine. I want to say one of the most recent updates to KSPI helped with the Vista engine, but I haven't been using it too much lately because of the radiation hazard issues are not compatible with my current infrastructure, so I can't say for sure. But yeah, the plasma thruster is totally borked when it comes to MechJeb. My workaround with the plasma thruster has been use the autopilots to create manuever nodes and then turn them off (keeps the node), set Smart A.S.S to orient to node, then use the Warp Helper to quickly warp to x seconds before node. At that point, you can tell MechJeb to execute the node and everything works perfectly fine from there.

It's a lot of extra clicks, but it's a good workaround for me.

[Tigermisu]

Does the science lab continue to generate science even when it isnt in the active craft? Or do I have to have it as the active craft?

[Fractal_UK]

I also run into the issue regarding MechJeb firing early when using the plasma engine. I want to say one of the most recent updates to KSPI helped with the Vista engine, but I haven't been using it too much lately because of the radiation hazard issues are not compatible with my current infrastructure, so I can't say for sure. But yeah, the plasma thruster is totally borked when it comes to MechJeb. My workaround with the plasma thruster has been use the autopilots to create manuever nodes and then turn them off (keeps the node), set Smart A.S.S to orient to node, then use the Warp Helper to quickly warp to x seconds before node. At that point, you can tell MechJeb to execute the node and everything works perfectly fine from there.

It's a lot of extra clicks, but it's a good workaround for me.

MechJeb and the plasma engine don't get on at the moment due to an extremely subtle bug, it's extremely hard to spot as a player but it throws off MechJeb completely. Currently the plasma engine doesn't throttle linearly, it throttles to something approximating the square of the throttle - so 10% throttle is actually calculated as ~1%. At 100% throttle though, there is no difference, which means the ships are easily flyable by Humans but not by MechJeb.

It will be fixed in 0.8 anyway.

Does the science lab continue to generate science even when it isnt in the active craft? Or do I have to have it as the active craft?

Yes, it will work in the background so long as the ship has power. You'll have to go back to the ship before the science is added to the R&D centre though.

I've got a question.

I've finally picked up that KSPI operates on a power system disassociated from stock electrical energy.

As such, anything requiring MJ/MW/Thermal energy does not interact with existing forms of power.

For example, I can't power the brayton cycle engine with thermal energy accumulated in the solar panels.

So I raise a question. Considering that in reality (that scary place we don't like) energy converts to all forms, can we expect some form of 'converter' components to say, convert electricity into MJs or thermal power?

It seems kind of silly I can't get thermal power or MJs (even in low amounts) from a configuration like a 'solar ship'. I don't care if it's low-power, inefficient and wasteful, but I like to have self-sustaining orbital infrastructure that doesn't require me to ship up a new nuclear reactor for every little vessel design.

It's not disassociated from stock electrical energy - everything that generates Megajoules also generates ElectricCharge seemlessly.

No, you can't power a Brayton Cycle turbine with a solar panel because solar panels produce electric current directly, while Brayton cycle turbines produce electricity from heat. A part that generated heat from solar energy would be quite different to a solar panel (i.e. a different part).

A conversion part has very niche uses given that at Kerbin it would take 333 stock solar panels to power a science lab - the only way that is really practical is to used solar satellites with microwave beamed power and that already works just fine because they generate megajoules as it is. A converter is only useful with either absolutely massive modded solar panels or at the orbit of Moho or closer.

Such a part might get added at some point but trust me when I say that it isn't as useful as people expect it will be.

If you want self-sustaining orbital infrastructure, a network of Low Kerbol solar satellites with microwave transmitters is the way to go.

[XanderTek]

So I raise a question. Considering that in reality (that scary place we don't like) energy converts to all forms, can we expect some form of 'converter' components to say, convert electricity into MJs or thermal power?

It seems kind of silly I can't get thermal power or MJs (even in low amounts) from a configuration like a 'solar ship'. I don't care if it's low-power, inefficient and wasteful, but I like to have self-sustaining orbital infrastructure that doesn't require me to ship up a new nuclear reactor for every little vessel design.

Well, parts that create MJ also create ElectricCharge. However, 1MJ=1000Ec so its not very practical to power the KSPI engines off of solar most of the time. The exception would probably be a satellite orbiting very close to the sun, and in that case you can use the microwave transmitter to turn Ec into MJ on the other end. Waste heat is generally heat that you need to get rid of in order for the system to work. Sometimes (in real life) you can feed it back into a recovery system to get a bit more out of it, but you can never use all of that heat. In this case, I suspect waste heat wouldn't be hot enough to power the engines or generators.

[RadHazard]

I'm having a minor bug with the TWR limiter for thermal engines. The limiter works just fine and scales thrust to the proper limit, but it seems to be running the reactor as if the engine were using full power.

Some screenshots:

Javascript is disabled. View full album

I would have had more but I was in the middle of a transfer burn when I noticed and I didn't want to screw it up.

Edited November 19, 2013 by RadHazard
Spelling

[AdmiralTigerclaw]

Oh, I'm aware of losses and diminishing returns. I'm just kind of annoyed you can't power something like a plasma engine, however patheticly, with a solar array.

I created a design using the S2 Wide bodies from B9, looks like a small supertanker. Has a TWR of .19, and a delta of 120,000 with the plasma engines. I want to use it for a fuel hauler/mobile refueling vessel. But I need a reasonable way to power it that doesn't require me to decommission it every time I run a reactor out... or have to lift a whole 40 ton reactor up every time I want to change it out. Can we haul UF6 supplies up? I'm not sure.

[RadHazard]

Oh, I'm aware of losses and diminishing returns. I'm just kind of annoyed you can't power something like a plasma engine, however patheticly, with a solar array.

I created a design using the S2 Wide bodies from B9, looks like a small supertanker. Has a TWR of .19, and a delta of 120,000 with the plasma engines. I want to use it for a fuel hauler/mobile refueling vessel. But I need a reasonable way to power it that doesn't require me to decommission it every time I run a reactor out... or have to lift a whole 40 ton reactor up every time I want to change it out. Can we haul UF6 supplies up? I'm not sure.

To power a plasma engine with the same level of thrust that an upgraded 1.25m reactor/generator combo gives (which is a piddling 0.9 kN on liquid fuel), you'd need 480 tons of solar panels (using the OX-STAT panels, the most mass efficient of them, and assuming you're in Kerbin orbit). It's really not practical with anything other than microwave beamed solar satellites, which absolutely are possible in the current version.

Other options in the current version are using a docked science lab to reprocess the fuel (vastly increases the lifetime of the reactor, but it will still run out eventually), or using antimatter reactors instead of nuclear (which requires high-level research).

The next version is going to implement an automated reprocessing part, and will allow you to refuel reactors completely using new UF4 (reactors will run off of UF4 in the update) tanks.

[Hremsfeld]

Oh, I'm aware of losses and diminishing returns. I'm just kind of annoyed you can't power something like a plasma engine, however patheticly, with a solar array.

I created a design using the S2 Wide bodies from B9, looks like a small supertanker. Has a TWR of .19, and a delta of 120,000 with the plasma engines. I want to use it for a fuel hauler/mobile refueling vessel. But I need a reasonable way to power it that doesn't require me to decommission it every time I run a reactor out... or have to lift a whole 40 ton reactor up every time I want to change it out. Can we haul UF6 supplies up? I'm not sure.

You can always change the resources.cfg file to make the UF6 and DUF6 pumpable, but something similar will be added in 0.8.

[Sternface]

Is there anything in this mod that would affect the solar panel power curves? I took a look and couldn't see anything obvious.

I use RSS, and all of a sudden I get no power from Squad's panels when out around Jool.

Please don't take this as blame, I'm just trying to narrow it down. I absolutely love this mod, and just downloaded it for the first time this week. So thank you.