Atomic Motors Not Worth It

[CaptainArchmage]

With the stock atomic motor, you have a large specific impulse (exhaust velocity) but high mass (2.25 tons) and low thrust (60kn). This means your final delta-v is very high, but you have to run the engine for a long time. For efficiency with the fuel, you should go for fewer (1-2) LV-Ns, which means you have to run the engine when you have reached orbital or near-orbital speeds because of the low acceleration.

A bad effect of the low thrust is you may lose on gains from the Oberth effect, since most of your burn is going to be "out of" the planet's gravity well. To get a proper idea of the comparison between an LV-N and a (high-thrust) LV-30 or LV-45 or Aerospike, you probably need to do the mathematical calculations on a computer, and the results will vary depending on your destination.

A good effect of low thrust is you do not need excessive strutting, which increases part count. High acceleration can cause a spacecraft to break apart. If you watch Scott Manley's videos, there is one where he uses a mainsail on an interplanetary injection burn, and the maximum safe thrust is about 1/3rd of the throttle.

[Monkeh]

Until you're in orbit you're pushing against gravity accelerating you downwards. The low thrust of the nukes struggle to overcome it. Try the same test from an orbital starting point and the nukes will win hands down.

I've recently been using the long slow burn principal with 2 nukes pushing a fat lot of fuel. Might take a 15 minute burn at 0.2 TWR, but when this stage alone has 7 or 8 thousand delta v it's all good.

[magnemoe]

One interesting fact is that an atomic engine has better twr than other engines just 2 kilometers up at 0.8-0.7 bar.

So if they are free to burn and not on top of tanks you can just as well ignite them well before gravity turn, this has the added benefit that you can use the last two boosters as drop tanks, kill the chemical engines and do the last part of gravity turn and circulate with the nuclear ones.

Obviously work better with ships with an decent TWR with the LV-N.

[Spartwo]

Once you go above 10 tons atomic motors become the most efficient option(from what I've tested)

[JUDUFU]

Seems to me that atomic motors are not the best thing for last stage voyages through space. Despite having an 800 isp and 60 thrust, the weigh 2.25, as much as small diameter half fuel tank. In comparison, the LV-909 weighs .5 and has 50 thrust. I just performed a test with a small parachute on top of a small capsule on top of a small half fuel tank powered by an LV-909. After a vertical decoupler, the lower stage was a small diameter quarter tank on top of a full tank on top of an LVT-45. I launched completely vertically and made it to 1,275,000 meters. (The second stage ignited at 11k feet, almost in a vacuum, so the isp was already quite near optimal.) I performed the same test, except I swapped out the LV-909 for an atomic motor. It made it 525,000 meters.

Conclusion: Although the atomic motor is just over twice as efficient as the LV-909, it's thrust-weight ratio is worse due to its much larger mass. Therefore, it is actually more efficient to use an LV-909. If you have enough fuel, the atomic motor's efficiency may eventually outweigh its thrust disadvantage, but at that point, you'd probably be better off using a more powerful LVT-45, so your burns don't take forever.

Maybe I'm insane, but this makes sense to me. Thoughts, feedback, and your own testing results would be much appreciated.

Assuming one engine and stable orbit in vacuum the total mass (engine + fuel) of your upper stage has to be over about 6 tons for Delta-V of the LV-N to become greater than the LV-909. There is usually a waste in Delta-V if the upper stage is not in orbit because of the large angles off of the prograde vector that may be needed to maintain altitude. Vacuum is at 70km so the ISP of the LV-N is likely much lower when you activate it inside the atmosphere.

TL;DR

For spaceships smaller than about 6 tons the LV-909 is more efficient than the LV-N

For spaceships heavier than about 6 tons the LV-N is more efficient than the LV-909

I would get your LV-N into orbit before using it.

[M5000]

Ahem. Basically, what everyone else in this thread just said.

If your ship is small enough to the point where an LV-909 gets you more DV than an LV-N, your ship isn't massive enough for it.

You see, due to the diminishing returns of sticking a massively heavy engine onto a tiny craft, the LV-909 still has some relevance on smaller vehicles. Now, the larger your ship gets, the less and less the LV-N's mass matters in the grand scheme of things, so you're able to take advantage of the higher Isp with some meaningful outcome.

That's what I see as the beauty of this game. Every ship has an optimal engine you're going to want to use, and that's all going to depend on how big/massive it is, and where it will be doing the majority of its operations, read: how much atmosphere will it encounter most often? You have to actually design your ship.

Generally, yes, the LV-N will have more applications than the LV-909. That is, unless you're Giggleplex and all your stuff is teeny tiny. But, if you happen to be designing a 1.25m or smaller 2.5m craft, the LV-909 may be more useful for you. Really, try it out. Experiment with a Delta-V measuring tool of your choice, and see what the results are as you change the mass of your ship, amount of fuel, and engines for propulsion.

[AlamoVampire]

yup, all the posts here.

[Whackjob]

If you do not want your atomic engines, send them all to me.

[loknar]

lol

Did someone calculate the dV threshold for when Atomic is starting to pay off (considering weight of engine + fuel used).

I brought a 8-way atomic engine shuttle/tugger and it was quite heavy to bring that to orbit (did not use the atomics at all until above 60km). After a many Duna trips and a few more trips to Eve, I'm sure it started to be efficient a long time ago.

[Hodo]

So to sum it up,

Depends on the situation which is useful. Like almost ALL things in KSP.

[arq]

lol

Did someone calculate the dV threshold for when Atomic is starting to pay off (considering weight of engine + fuel used).

I brought a 8-way atomic engine shuttle/tugger and it was quite heavy to bring that to orbit (did not use the atomics at all until above 60km). After a many Duna trips and a few more trips to Eve, I'm sure it started to be efficient a long time ago.

<dV> = <ISP>*log(<WetMass>/<DryMass>)

The wet/dry ratio is always less than 9 (because of the parameters of fuel tanks). For example, if we have a wet/dry mass of 8 (which is quite high) with a LV-909 (or any other 390-ISP engine), we would be able to achieve the same results with a

2.76=exp(390/800*log(8))

wet/dry mass ratio with an LV-N. However, note that the LV-N is heavier, which increases dry mass and thus hurts the ratio.

Another interesting statistic is asymptotic single-stage delta-V, which just assumes a wet/dry ratio of 9 (so we have one engine and a near-infinite fuel stock. In this case, range is directly proportional to ISP, so for very heavy loads the LV-N can achieve double the dV of any other (rocket) engine.

[Awass]

Assuming one engine and stable orbit in vacuum the total mass (engine + fuel) of your upper stage has to be over about 6 tons for Delta-V of the LV-N to become greater than the LV-909. There is usually a waste in Delta-V if the upper stage is not in orbit because of the large angles off of the prograde vector that may be needed to maintain altitude. Vacuum is at 70km so the ISP of the LV-N is likely much lower when you activate it inside the atmosphere.

TL;DR

For spaceships smaller than about 6 tons the LV-909 is more efficient than the LV-N

For spaceships heavier than about 6 tons the LV-N is more efficient than the LV-909

I would get your LV-N into orbit before using it.

Thank you. That is incredibly helpful. You should go get it pinned it up somewhere on the forums. But could you add in the skipper and poodle engines to the graph? I use those a lot.

[Awass]

Vacuum is at 70km so the ISP of the LV-N is likely much lower when you activate it inside the atmosphere.

Well actually, the isp at 11k feet was already 760 out of the optimal 800, but yes, I was not using the atomic motor as it really should be.

[arq]

LV-N ISP is fine on Kerbin anywhere except the bottom few km. The low TWR is more the issue. Still, if I'm in an asparagus vessel I'll try to get the LV-N's firing fairly early for a slight boost to thrust and ISP (though it's pretty insignificant next to a few mainsails).

[loknar]

Thank you. That is incredibly helpful. You should go get it pinned it up somewhere on the forums. But could you add in the skipper and poodle engines to the graph? I use those a lot.

That graph shows increasing weight I'm guessing to be fuel only (no tanks?). That's not "total ship weight" just to be clear. Adding more weight from other parts, or stages/landers with fuel that is not used by the engine under the test, further advantages the high ISP of the LV-N.

So typically you don't need to burn 6t of fuel and/or 6000dV to take advantage of it.

I'd like to see another graph with a scenario. eg: engine+fuel+40% (? how much is it usually ?) dead weight cargo. Hehe I feel bad not bothering making it myself.. Usually when we plan a mission there is a purpose and equipment to bring there. The weight of the tanks is important to consider.

But this is the most helpful graph because it is easy to analyse. Adding more variables will be confusing.

Edited September 30, 2013 by loknar

[AlamoVampire]

Okay, I am absolutely FLOORED by the geniuses here. no sarcasm at all! I know this is a game using real world rocket science, but to see so many of you who have such a spectacular grasp of the math involved is awe inspiring!! I am truly a small fish among giants! Heck, I would not be surprised if Michio Kaku was wandering around here somewhere OR Professor Brian Cox, both super smart men! <bows down to his fellow players>

[loknar]

Okay, I am absolutely FLOORED by the geniuses here. no sarcasm at all! I know this is a game using real world rocket science, but to see so many of you who have such a spectacular grasp of the math involved is awe inspiring!! I am truly a small fish among giants! Heck, I would not be surprised if Michio Kaku was wandering around here somewhere OR Professor Brian Cox, both super smart men! <bows down to his fellow players>

+1. and we can pat ourselves on the back. I mean; try to get the same average IQ on Grand Theft Auto forums

[UmbralRaptor]

I'd like to see another graph with a scenario. eg: engine+fuel+40% (? how much is it usually ?) dead weight cargo. Hehe I feel bad not bothering making it myself.. Usually when we plan a mission there is a purpose and equipment to bring there. The weight of the tanks is important to consider

I tried a table that actually did that upthread, but it only covers mass breakpoints. Maybe I should add craft ÃŽâ€V at those points?

[Kerbart]

I've started to use these a lot: http://kerbalspaceprogram.com/radial-nuclear-engines/

Near-stock balanced, well-textured and far easier to attach to your designs.

The largest is 2.55 mass and 830 dV.

The smallest one looks like it could be a great hand dryer in Kerbal bath rooms. It'll get glowing reviews!!

[FlamedSteak]

+1. and we can pat ourselves on the back. I mean; try to get the same average IQ on Grand Theft Auto forums

+2. I grasp the concepts of what people talk about on here, but the mathematics behind it are way above my head.

[CalculusWarrior]

Okay, I am absolutely FLOORED by the geniuses here. no sarcasm at all! I know this is a game using real world rocket science, but to see so many of you who have such a spectacular grasp of the math involved is awe inspiring!! I am truly a small fish among giants! Heck, I would not be surprised if Michio Kaku was wandering around here somewhere OR Professor Brian Cox, both super smart men! <bows down to his fellow players>

If you think so, check out the KSP Interstellar mod discussion, quite frequently there will be discussions on advanced particle physics that go on for several pages.

[magnemoe]

That graph shows increasing weight I'm guessing to be fuel only (no tanks?). That's not "total ship weight" just to be clear. Adding more weight from other parts, or stages/landers with fuel that is not used by the engine under the test, further advantages the high ISP of the LV-N.

So typically you don't need to burn 6t of fuel and/or 6000dV to take advantage of it.

I'd like to see another graph with a scenario. eg: engine+fuel+40% (? how much is it usually ?) dead weight cargo. Hehe I feel bad not bothering making it myself.. Usually when we plan a mission there is a purpose and equipment to bring there. The weight of the tanks is important to consider.

But this is the most helpful graph because it is easy to analyse. Adding more variables will be confusing.

I think you are pretty right here, I did some calculations based on various payloads and found that the 48-7S wins for small probes, yes it you have an 50 to 200 kg payload the 48-7S is most efficient at least any other places than Moho.

For larger weights the LV-N takes over, the 909 is not well suited for interplanetary missions, however its an very good lander engines for lander can and larger landers.

[JUDUFU]

Too much to reply to right now. The total weight is total weight including fuel container weight. I should have been more clear. Adding more weight will have the effect of decreasing the Delta-V.

I may need to make a blog post about this.

Here's another graph somewhat self explanatory. The more kerbals you pile into your space craft (dead mass) the more fuel you will have to carry to increase your Delta-V.

[Pixel of Life]

TWR doesn't really matter for interplanetary trips. Delta-V is more important.

This is true, however, if you're departing from Kerbin (or any other planet or moon with an atmosphere) and using the maneuver nodes to plan your burns, the lower your TWR the higher your parking orbit needs to be. This is because with a lower TWR you must start your burn earlier and the maneuver nodes act as if the planet/moon is not there, thus guiding you straight into the atmosphere. With a higher parking orbit you won't be able to utilize the oberth effect quite as well. I don't know how much dV you lose if you burn from, say, 200km Kerbin orbit as opposed to 100km - it's probably not enough to cause any problems.

So TWR does sort of matter, depending on where you are. I prefer using really powerful engines to get my stuff from LKO to interplanetary space (screw efficiency), and then getting rid of those and using more efficient ones for fine tuning my trajectory.

[Brotoro]

This is true, however, if you're departing from Kerbin (or any other planet or moon with an atmosphere) and using the maneuver nodes to plan your burns, the lower your TWR the higher your parking orbit needs to be. This is because with a lower TWR you must start your burn earlier and the maneuver nodes act as if the planet/moon is not there, thus guiding you straight into the atmosphere. With a higher parking orbit you won't be able to utilize the oberth effect quite as well. I don't know how much dV you lose if you burn from, say, 200km Kerbin orbit as opposed to 100km - it's probably not enough to cause any problems.

So TWR does sort of matter, depending on where you are. I prefer using really powerful engines to get my stuff from LKO to interplanetary space (screw efficiency), and then getting rid of those and using more efficient ones for fine tuning my trajectory.

You just need to do your burn during multiple passes near periapsis.