Atomic Motors Not Worth It

[Temstar]

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

Or follow the prograde marker and start the burn earlier, instead of following the maneuvering node

[Vlk]

Or follow the prograde marker and start the burn earlier, instead of following the maneuvering node

Thats inaccurate, plus i think its a waste of fuel - and this is the point of this topic: efficiency. But the disadvantage of bad twr is the loss of the ability to take a good advantage on the oberth effect, i think we can all agree on this.

[Temstar]

Thats inaccurate, plus i think its a waste of fuel - and this is the point of this topic: efficiency. But the disadvantage of bad twr is the loss of the ability to take a good advantage on the oberth effect, i think we can all agree on this.

No it's perfectly legitimately, in fact real life missions have been planned using this type of "spiral orbit":

http://www.astronautix.com/craft/stus1957.htm

Ernst Stuhlinger came up with a mission planned to Mars known as Stuhlinger Mars 1957 which called for nuclear reactor powered, ion thrusters propelled spacecrafts to accelerate prograde for some 124 days in a spiral orbit close to 400 revolutions around Earth before building up enough velocity to reach escape velocity. The ship continues to fire the engines on the way to Mars before turning around half way and firing retrograde. When it gets to the edge of Mars's SOI the ships then perform a 126 day deceleration inwards in a reverse spiral orbit until it reaches low circular Martian orbit.

Yes, you do lose out on Oberth effect, but the idea is that if your propulsion system is efficient enough then the loss is more than offset by the increase in specific impulse.

Firing prograde in an arc is the same idea only not taken that that extreme.

[TMS]

I must admit that I tend to just place low-thrust ships in a slightly higher parking orbit (100-120k) and then do the transfer burn in one pass.

Typically, I'll graze the atmosphere anywhere between 70-80k peri.

[Kerbart]

No it's perfectly legitimately, in fact real life missions have been planned using this type of "spiral orbit"

(...)

Yes, you do lose out on Oberth effect, but the idea is that if your propulsion system is efficient enough then the loss is more than offset by the increase in specific impulse.

Firing prograde in an arc is the same idea only not taken that that extreme.

Those IRL missions have been planned with ion engines who have (a) insane ISP's and ( even less thrust than the Kerbal ion engine. Yes, you're losing out on the Oberth effect but in this particular case the virtues of using the ion engine (and the necessity of the spiral burn) negate that.

The main point of the discussion here will be "how far up the arc are you going to fire". A 180° burn (or worse) will obviously negate the advantage of the nuclear engine. On the other hand, a 30° burn probably will not (I'm guessing here). As with everything, not taking things up to the extreme will probably do the trick.

In some cases you will not have the luxury of burning in multiple stages (when being captured by a planet for instance). That's where you'll need a regular engine. In all other cases it comes down to trading convenience (single burn) for fuel efficiency. And given that I don't have to swipe my card to send up yet another set of 4 orange tanks to put an XL-400 filled with fuel in orbit, I can't blame people for preferring convenience over tedium, sometimes (I use nukes wherever I can).

[magnemoe]

I must admit that I tend to just place low-thrust ships in a slightly higher parking orbit (100-120k) and then do the transfer burn in one pass.

Typically, I'll graze the atmosphere anywhere between 70-80k peri.

Me to, however then using ion engines doing two burns might be smart more so as you often get shadow problems.

Another effect of the long burn who drop your Pe is that you tend to earn some on the Oberth effect as you go low and fast and your estimated burn tend too long, found it most effective to cut long burns a bit short anyway and correct after main burn

[Temstar]

A 180° burn (or worse) will obviously negate the advantage of the nuclear engine.

Actually no, why would it? It's not like in a 180° burn you're burning retrograde at the start or something. All that burn is prograde so you're still converting 100% of your delta-V to prograde acceleration that you need, gravity is just doing the work for you to bend your path so as you reach escape velocity just as you happen to hit the correct ejection angle. The only difference is that other than the very start of the burn you're losing out progressively more and more on Oberth effect as your orbit widens. Although I don't have the calculation I don't think that loss in Oberth effect is going to cancel out more than double the Isp compared to chemical engines.

[Vlk]

Hmm, sorry Temstar, you are right I had a slight doubt that the thrusting to different directions would work against itself. I guess that the gravity of the planet makes it right then? I remembered something from the ingame kerbal tutorial that if you point the wrong direction, you can burn a lot of fuel in an orbit while only lightly changing the orbit path. And that confused me. Is it when you point your nose 90° from the forward vector (pointing away from the planet)? Or is it incorrect too?

[lazarus1024]

As a number have pointed out, it is when the mass percentage of the motor becomes a relatively trivial portion of the total mass that the LV-N comes in to its own.

I could do the math, but I am lazy. Needless to say, it is when you are getting up around the 15-25% range, so basically total craft weight needs to be in excess of about 10-15t for the LV-N to make sense. That includes fuel, tankage, the engine itseld, capsule/probe core, etc.

Lower than that and the mass penalty paid for the engine is not worth it.

[Runboy]

*Its engine,not motor.Motors are electrical,engines are rockets and such.

[lincourtl]

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.

Holy Jool! So that's why that was happening. Thanks. I didn't know the maneuver nodes didn't take the planet/moon into account. I thought it was just a bug with the maneuver node system, and had just gotten used to eyeballing my low-TWR escape burns.

[lincourtl]

*Its engine,not motor.Motors are electrical,engines are rockets and such.

Unless you're talking about a solid propellant rocket -- then it's a motor. At least that's what I remember from model rocketry.

On edit: Technically, a motor is a sub-category of engines which impart motion, so all rockets are motors, but not all engines are motors. An engine is any device which converts energy into work. Although it looks like the two terms usage had begun to converge as early as the 19th century so whatever we use now for any given device is merely a matter of convention.

Edited September 30, 2013 by lincourtl

[magnemoe]

As a number have pointed out, it is when the mass percentage of the motor becomes a relatively trivial portion of the total mass that the LV-N comes in to its own.

I could do the math, but I am lazy. Needless to say, it is when you are getting up around the 15-25% range, so basically total craft weight needs to be in excess of about 10-15t for the LV-N to make sense. That includes fuel, tankage, the engine itseld, capsule/probe core, etc.

Lower than that and the mass penalty paid for the engine is not worth it.

A bit lower, main benefit is that you can you can bring 1,75 ton / 145 liter more fuel if you use the 909, if your total fuel requirements is more than twice of that the LV-N wins. Payload does not matter, however if its very light it might be better using an 48-S7.

[Meithan]

EDIT: I re-posted below.

Edited September 30, 2013 by Meithan

[Dispatcher]

It just depends on your needs and circumstances (and budget). Each engine has its own pros and cons. Very few would be considered "all in (w)one-ders", if any. But that's the fun of KSP, selecting the right tools for the right job or mission.

[Meithan]

Expanding on JUDUFU' analysis, I think it's important to include the payload mass, since in the end the point of a spaceship is pushing something other than itself and a bunch of fuel.

The following plots show the total ship mass (including a fixed amount of payload) as a function of desired delta-v. Right now I compared only the LV-N and the LV-909, since they have comparable thrusts. It's also just one engine at the moment. I'm considering the mass of the empty fuel tanks in the calculations, taken as 1/8th of the fuel mass they carry (true for all rocket fuel tanks except the tiny ones).

For comparison, in the first plot I set the payload to zero, so this one is an engine pushing itself and its own fuel:

One sees from this plot that the LV-909 is a better choice if the desired total delta-v is below about 6400 m/s, as found by JUDUFU. If you want more delta-v, you need to take more fuel, and the LV-N becomes a better choice. As many have pointed out, this is because the larger engine mass has less of an impact since it represents a smaller fraction of the total ship mass. Let's see what happens when we start adding useful payload.

With a 1 tonne payload, this is the result:

The delta-v at which the LV-N becomes a better choice is now smaller, at around 4200 m/s. The exponential nature of the rocket equation really shows in the curve for the LV-909.

If we increase the payload to 2.5 tonnes, we get this:

The LV-N is a better choice at around 2700 m/s, which starts to become too low for some interplanetary missions. If you need to take 2.5 tonnes of payload to one of Jool's moons, the LV-N will definitely be a better choice.

Cranking the payload up to 5 tonnes, we obtain:

Now the LV-N is a better choice at around 1700 m/s. You might pull off a no-return mission to Duna with this much delta-v. The high Isp of the LV-N definitely starts to pay if you need to go farther. If you need 5000 m/s for your mission with this payload, you'd need a spacecraft twice as heavy if you used the LV-909 instead.

Of course, what I'm not considering is that as your spacecraft starts becoming more massive, you'll want more thrust too. So the LV-N will be adding its mass multiple times, which will make the graphs worse for it.

Edited September 30, 2013 by Meithan

[Sma]

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.

As others have noted by now, starting the LV-N's in atmosphere isn't really where they should be used. At 11k, you've "barely" (well, in my mind anyway) started getting out of the atmosphere and are still fighting gravity and air friction. I don't use the LV-N's generally until after I'm in orbit. Also, from what I remember when Scott Manley was talking about them, much more than somewhere around 2-4 LVN's and you start loosing out due to the mass of the engine itself. I've used 2 LVN's when I was testing a kethane miner/lander. It needed very little fuel left in them to land, and I doubt I was using the best landing technique. Not like that is a conclusive test, as I haven't tried other engines.

I may have to try a test with one of my rockets I have with kOS on it. This way I can be sure to get relatively the same exact launch profile (other than straight up) and compare amounts of fuel left over once in orbit. Currently it uses 6 asparagus stages aerospikes, and one of the engines that has thrust vectoring. This is using all "normal" sized tanks...what ever they're called.

[lazarus1024]

Mmmmm just depends. Most interplanetary stuff can get by with very low mass if you don't mind long burn times and possibly multiple orbits to achieve escape velocity.

That said, sure, chain a bunch of LV-Ns together and their mass penalty increases.

However, if you are sending a big payload, you are going to need them too. Your posited Duna mission would be able to manage possibly a one-way trip single kerbal landing can for a 5 ton mass. If you wanted a Duna trip with return-to-orbit capability, a seperate orbital module AND the whole kit-n-kaboodle (or at least the orbital module) to return to Kerbal you are probably talking 20-40t payload, not including the drive engines and fuel for them.

My typical designs have 2-4 LV-N engines for something like a 20-40t payload. The whole thing is generally given a nice gentle shove from kerbin with a Skipper engine and a big rock-o-max fuel tank. If I need Jool or Moho (especially Moho and the higher TWR it often needs for braking) and a return mission I'll usually have at least a couple of strap on rock-o-max orange tanks with 1-2 LV-N engines per tank and jettison the engines and the tanks once they are empty.

If you need temporary high TWR and lower ISP is okay, you could also have several supplementary 48-7s engines. Their mass is almost too low to be a consideration and their TWR is extremely high. Thrust is low overall, but they produce 1/3rd the thrust of an LV-N at 4% of the mass. They also have double the TWR of the L-909.

I have used that combination a few times in the past where I have LV-N as the primary engines but I have a couple of 48-7s on there for phases where I need higher TWR of the overall assembly (for example, breaking orbit with Kerbin when I don't want to spend several orbits raising my AP, or Moho capture when you have so much velocity that one or more LV-Ns just might not be able to produce the TWR needed to get capture before you tear right through the SOI of Moho).

It is a method of producing an overall very lightweight rocket. You can have even very large ships with just a single LV-N but 4+ 48-7s supplementary engines and use the LV-N for most of the mission and the 48-7s just for the times you need extra TWR. It might be the way to best maximize dV with a mission that has varying requirements for TWR.

[John FX]

Don't consider the LV-N to be the best "in a vacuum," consider it to be the best "in orbit."

It usually is. Even then, with a small fuel tank, a lighter lower efficient engine can have higher Dv.

I am learning the value of oscars and the LV-1R for certain situations...

[Meithan]

Mmmmm just depends. Most interplanetary stuff can get by with very low mass if you don't mind long burn times and possibly multiple orbits to achieve escape velocity.

That said, sure, chain a bunch of LV-Ns together and their mass penalty increases.

However, if you are sending a big payload, you are going to need them too. Your posited Duna mission would be able to manage possibly a one-way trip single kerbal landing can for a 5 ton mass. If you wanted a Duna trip with return-to-orbit capability, a seperate orbital module AND the whole kit-n-kaboodle (or at least the orbital module) to return to Kerbal you are probably talking 20-40t payload, not including the drive engines and fuel for them.

My typical designs have 2-4 LV-N engines for something like a 20-40t payload. The whole thing is generally given a nice gentle shove from kerbin with a Skipper engine and a big rock-o-max fuel tank. If I need Jool or Moho (especially Moho and the higher TWR it often needs for braking) and a return mission I'll usually have at least a couple of strap on rock-o-max orange tanks with 1-2 LV-N engines per tank and jettison the engines and the tanks once they are empty.

If you need temporary high TWR and lower ISP is okay, you could also have several supplementary 48-7s engines. Their mass is almost too low to be a consideration and their TWR is extremely high. Thrust is low overall, but they produce 1/3rd the thrust of an LV-N at 4% of the mass. They also have double the TWR of the L-909.

I have used that combination a few times in the past where I have LV-N as the primary engines but I have a couple of 48-7s on there for phases where I need higher TWR of the overall assembly (for example, breaking orbit with Kerbin when I don't want to spend several orbits raising my AP, or Moho capture when you have so much velocity that one or more LV-Ns just might not be able to produce the TWR needed to get capture before you tear right through the SOI of Moho).

It is a method of producing an overall very lightweight rocket. You can have even very large ships with just a single LV-N but 4+ 48-7s supplementary engines and use the LV-N for most of the mission and the 48-7s just for the times you need extra TWR. It might be the way to best maximize dV with a mission that has varying requirements for TWR.

Very interesting comments. Yes, TWR considerations is something I haven't included in the previous analysis, and as you say it's something that can be an important aspect of mission design.

I've been designing a Duna manned mission (which must return of course!) and the lander mass is about 20 tons. Wanting something around 200 kN thrust so the burns don't take forever, I ran the numbers comparing 3-4 LV-N engines vs. a single LV-T40 or LV-T45. If I recall correctly, I think the LV-Ns were the better choice.

[tavert]

Expanding on JUDUFU' analysis, I think it's important to include the payload mass, since in the end the point of a spaceship is pushing something other than itself and a bunch of fuel.

<snip>

Yep. Taking this analysis to its conclusion: http://forum.kerbalspaceprogram.com/threads/45155-Mass-optimal-engine-type-vs-delta-V-payload-and-min-TWR

[Meithan]

To complement the single-engine comparison of the LV-N and LV-909, I've calculated the delta-v at which the LV-N becomes a better choice than the LV-909, as a function of payload:

The red curve is the delta-v (with its scale on the left) while the blue curve is the corresponding total ship mass (scale on the right).

[Meithan]

Yep. Taking this analysis to its conclusion: http://forum.kerbalspaceprogram.com/threads/45155-Mass-optimal-engine-type-vs-delta-V-payload-and-min-TWR

Wow. Great stuff!

It seems that above 10 tons payload and 2000 m/s delta-v, the LV-N is the king pretty consistently if the required minimum TWR doesn't exceed 0.7 (Kerbin-relative), which is plenty for an interplanetary stage.

[Blackstar]

The thing about the LV-N is that its high mass make it not worthwhile to launch for short-term missions. However, the more fuel you feed through one before deorbiting it, the better the LV-N is by comparison.

I often use them in my orbital tugs, or their KSPX cousin the LV-NB (balanced so it's exactly the same as sticking several LV-Ns on your craft, but lower part count). Those craft get launched once and stay in orbit pretty much forever, getting refueled in orbit as necessary. So they end up burning a LOT more fuel than could feasibly be launched with them, offsetting the low TWR. Subsequent fuel can launches don't have to send the heavy engine up again, which makes the whole process more efficient. The relatively low thrust does mean longer burns, but on the other hand, less wobble when pulling awkward payloads around.

[magnemoe]

The thing about the LV-N is that its high mass make it not worthwhile to launch for short-term missions. However, the more fuel you feed through one before deorbiting it, the better the LV-N is by comparison.

I often use them in my orbital tugs, or their KSPX cousin the LV-NB (balanced so it's exactly the same as sticking several LV-Ns on your craft, but lower part count). Those craft get launched once and stay in orbit pretty much forever, getting refueled in orbit as necessary. So they end up burning a LOT more fuel than could feasibly be launched with them, offsetting the low TWR. Subsequent fuel can launches don't have to send the heavy engine up again, which makes the whole process more efficient. The relatively low thrust does mean longer burns, but on the other hand, less wobble when pulling awkward payloads around.

Yes if you refuel them many times the efficiency will go up.

However for pure reusable landers they are usually an bad choice even then landing on low gravity world like Mun, again an miner who do the Minmus-LKO-> Minmus run will benefit from LV-N as most of its dV use is not takeoff, however you might want to add some chemical helper engines for takeoff, you just need to burn a 2-300 m/s then from low minmus orbit to LKO docking and return pretty dry.