what is the best engine for a 15.4 ton spacecraft?

[GeneralVeers]

The mothership with which I made the trip had a total loaded weight of more than 300 tons. She just casually glanced at the 4-6 extra tons of weight on the lander, and yawned at me.

No respect, I tell ya.

[GoSlash27]

JWOC,

 The best engine will depend on how much DV you want from it, how much t/w is required, and how you define "best". Best could mean lightest overall stage, cheapest stage, or least fuel consumed.

Assuming you want the lightest stage and at least .5 t/w, it breaks down like this:

up to 2200 m/sec DV Aerospike

2300-3000 Poodle

3100-4500 Skipper

4600-6000 Rhino

^ single engine solutions.

For multiengine solutions (up to 6), the LV-N is dominant from 1900-5100 m/sec.

Best,
-Slashy

 

 

 

 

 

 

[Snark]

1 hour ago, GeneralVeers said:

Pretty long post. The design considerations took a lot less words than that: my Duna lander was around 12 tons before the engines went on. So it was either 1-2 more tons for standard engines, or 4-6 more tons (I forget which version I did the landing in) for nukes at just under twice the ISP. Either way, the nukes were the best choice.

So, let's say it's in post-1.0, which is when the OP is trying to make all this work.

You have 12 tons of craft, plus 6 tons of nukes, for an 18-ton lander with 120 kN thrust.

Takeoff acceleration will be 6.67 m/s².  Duna gravity is 2.94 m/s².  Therefore, your actual vertical acceleration on liftoff would be 3.73 m/s².  It means you'd be wasting fully 44% of your fuel with gravity loss, at least until you can build up enough horizontal speed to make a difference, which will take a while at that acceleration.

Compared with, for example, putting a single Poodle on.  Let's say we keep the total takeoff mass to 18 tons, which would be:  1.75 tons for the Poodle, 3.78 tons of extra fuel, and .47 tons of extra fuel tankage.  The acceleration will be 13.88 m/s², making the gravity losses just 21%.  Those 3.78 tons of fuel will give you 800 m/s of dV before you've even started to touch the original fuel reserve included in your 12-ton figure.

That's nearly enough to Duna orbit right there.

So really, the nuke-vs-Poodle comparison is:  take the Poodle craft, give it an 800 m/s head-start to Duna orbit, and then stack it up against the nukes that are losing 44% of their fuel to gravity loss.

 

1 hour ago, GeneralVeers said:

When you "get hammered with gravity losses", it's not the weight of the engines that counts. It's the total weight of the entire vehicle. If you get twice the ISP and you less than double the weight of the vehicle, you made a good tradeoff.

The issue is not "twice the weight" ... it's "twice the acceleration after subtracting gravity".  What you just said would be true, if what we did was to replace the heavy, weak nukes with light, weak engines that have the same total thrust.  But that's not the thing to do.  It's replacing heavy, weak nukes with light, much stronger engines that will therefore significantly reduce gravity losses.  Your effective I_sp for those nukes, if you subtract gravity loss at takeoff, is just 447 s.  By comparison, the effective Isp at takeoff for the Poodle example I give is 275 s.  Yes, it's still a higher effective I_sp for the nukes at takeoff... but on the other hand, the net vertical acceleraiton of the Poodle version would be roughly three times higher, meaning that it has to pay those gravity losses for a much shorter time than the nukes would.

 

Also, one more thing to consider,

1 hour ago, GeneralVeers said:

The mothership with which I made the trip had a total loaded weight of more than 300 tons. She just casually glanced at the 4-6 extra tons of weight on the lander, and yawned at me.

...that's an excellent point, and the mass of shipping the nukes there wouldn't make much difference in that case.  But that's the specific case of sending a massive multi-hundred-ton behemoth to Duna.  Not sure what the OP was proposing to do... but the number mentioned was 15.4 tons, which is rather smaller than 300 tons. 

Edited February 14, 2016 by Snark

[GeneralVeers]

1 hour ago, Snark said:

The issue is not "twice the weight" ... it's "twice the acceleration after subtracting gravity".

For JWOC's ship, it's the weight. It won't be landing on anything. When you're in orbit, it doesn't matter if it takes three times as long to complete a transfer burn to get your apoapsis to intersect with Duna.

For the Duna lander I was blabbering about, acceleration minus gravity turned out to be a non-issue. All-around fuel efficiency turned out to be the big win, and the nuclear-armed (snicker) Duna lander worked magnificently in actual use.

Edit: Whoops! Forgot something. Nuclear engines don't use oxidizer. Big weight savings there.

Edited February 14, 2016 by GeneralVeers

[Snark]

2 hours ago, GeneralVeers said:

For JWOC's ship, it's the weight. It won't be landing on anything. When you're in orbit, it doesn't matter if it takes three times as long to complete a transfer burn to get your apoapsis to intersect with Duna.

Yup, no argument with you there.    As long as it's in a situation where the TWR doesn't matter (such as JWOC's orbital-transfer vehicle), the nuke is the winner, hands-down.

2 hours ago, GeneralVeers said:

Edit: Whoops! Forgot something. Nuclear engines don't use oxidizer. Big weight savings there.

Not sure what you mean by that-- the fact that nukes don't use oxidizer is irrelevant to weight.  A ton of reaction mass is a ton of reaction mass-- doesn't matter whether it's LF or LF+O.  It just means that a post-1.0 nuke will burn 1 ton of LF, whereas a pre-1.0 would burn 0.45 ton of LF and 0.55 ton of O.  Same result, same mass, just a different mixture.

On the other hand, if you meant that you were looking at someone's design, and they have nukes on it, and they've mistakenly supplied it with LFO tanks instead of LF tanks so that it's hauling around a whole bunch of oxidizer which is effectively nothing but dead weight, then yeah, ditching the oxidizer would be a weight savings, just like ditching any dead weight is a weight savings.

[juanml82]

2 hours ago, GeneralVeers said:

For JWOC's ship, it's the weight. It won't be landing on anything. When you're in orbit, it doesn't matter if it takes three times as long to complete a transfer burn to get your apoapsis to intersect with Duna.

Actually, it can matter, since the longer the burn, the less accurate it is. On top, if it's too long and you're starting from a low orbit to take advantage of the Oberth effect, a long burn might lower your Pe to the atmosphere... or you have to start burning later, so it's even more inaccurate.

Case in point: I was returning from Moho, starting at a low 20km orbit. The burn required 2400 m/s, give or take and it was about 8 minutes long. Since I hadn't raised my orbit beforehand, I couldn't start the burn until a minute or so before the node or I would have conducted a lithobraking instead. Eventually, I had to cut the burn before it was over and correct the course in solar orbit.

More thrust (and starting from a higher orbit, Moho's Oberth effect is pathetic anyways) would have saved me dV, despite the additional weight of moar nukes

 

[Plusck]

FWIW, my first manned landing on Duna also used dual nukes. I had to land a lab for a contract, so I had the nukes on pylons set to detach, each with a single LF tank and Kerbin re-entry pod on top. They worked great. Slight problem was that the base wasn't stable with releasing the return pods one at a time, so I had to nurse both of them to orbit (staying within non-self-destruct distance) at the same time. But it worked fine.

Was it strictly the most efficient set-up? I don't know. But it worked, got the goods on the ground for minimal part count and smallish fuel tanks, and got the bods back home on top of a nuclear missile.

[Temstar]

21 minutes ago, juanml82 said:

Actually, it can matter, since the longer the burn, the less accurate it is. On top, if it's too long and you're starting from a low orbit to take advantage of the Oberth effect, a long burn might lower your Pe to the atmosphere... or you have to start burning later, so it's even more inaccurate.

You don't have to follow the node marker, the burn can also be completed by following the prograde marker so you never deorbit yourself.

When you follow the node marker for the entire burn you get zero cosine loss but some gravity drag, due to having some component of the burn in or against the direction of gravity.

When you follow the artificial horizon on the navball for the entire burn, you get zero gravity drag, but some cosine losses because some of your burn has radial component relative to the node. You start the burn before the node so the radial is in one direction which is cancelled out by the radial component in the opposite direction from the part of the burn that's perform after you've passed the node

When you follow the prograde market on the navball for the entire burn, you get some gravity drag since prograde starts to move away from the artificial horizon once you start your burn, you also get some cosine loss because prograde does not follow the node exactly. The sum of gravity drag and cosine loss should be lower than either the "follow the node" method or the "follow the horizon" method.

[Snark]

9 minutes ago, juanml82 said:

Actually, it can matter, since the longer the burn, the less accurate it is. On top, if it's too long and you're starting from a low orbit to take advantage of the Oberth effect, a long burn might lower your Pe to the atmosphere... or you have to start burning later, so it's even more inaccurate.

Case in point: I was returning from Moho, starting at a low 20km orbit. The burn required 2400 m/s, give or take and it was about 8 minutes long. Since I hadn't raised my orbit beforehand, I couldn't start the burn until a minute or so before the node or I would have conducted a lithobraking instead. Eventually, I had to cut the burn before it was over and correct the course in solar orbit.

Yup, point taken.  Though there are some ways to mitigate this.  First, you can (up to a point) split the burn.  Instead of doing the whole burn in one whack, you can do a "reasonable" burn that's not too long (say, a minute, or a minute and a half) to boost your Ap somewhat, then coast around and do it again.

You can't do that all the way, because there's only so much dV you can add before you're on an escape trajectory.  But it does give you a place to invest a few hundred dV, thus reducing the size of the last burn.  Also, it makes it easier to to do that last burn, because it's happening at the periapsis of a long elliptical orbit, so your orbital track is less curved, so you can do a longer burn without distortions/inaccuracies.

The other thing is, even in those cases where you do have to do a really long burn, you don't do it by pointing at the maneuver node and burning.  That would be inefficient (big cosine losses) and problematic (lower your Pe, possibly catastrophically).  Instead, you point prograde (which will be higher than the node).  It means a very dicey navigational job, since the ideal burn won't necessarily be a 50/50 split across the node, and judging that can be tricky.  But it works well enough, especially with practice.  What I usually do in such cases is do what I've just described, but stop burning after the burn's about 85% done, then kill the maneuver node and set up a new one a few minutes ahead of my craft.  The savings in accuracy is generally worth the slight loss of Oberth benefit during those few minutes of coasting.

[Temstar]

10 minutes ago, Snark said:

It means a very dicey navigational job, since the ideal burn won't necessarily be a 50/50 split across the node, and judging that can be tricky.  But it works well enough, especially with practice.

Yeah it's close to 50/50 but not exactly, usually the part before the node should be a little bit longer than after the node.

I tried to compile a list of experimental data to come up with some rules of thumbs on how to split the burn using follow the prograde but it turns out no such rule is possible - the split not only depends on TWR when you start the burn, but also on TWR at the end of the burn. Since TWR before and after depends on the size of the burn and the total delta-V of the ship / its dry mass not only will be split be different between different ships of differing TWR, it will also be different for the same ship for different sized ejection burns.

[Laie]

6 hours ago, Snark said:

Yes, nukes work just fine in Duna's atmosphere.  However, the issue is one of TWR.  Duna has fairly significant gravity, 0.3g at the surface.  Nukes are massive and have very poor TWR, which doesn't matter in orbit but really matters in a lander.

You keep saying that.... I don't like that blanket statement. Nukes are good for up to 5m/s² before their dV plummets below what you'd get out of a conventional engine.

The rightmost vessel is a late version of my venerable Labrover, from way back when you needed a lab to reset experiments. If memory serves, it was 15t dry / 32t wet for not quite 6000m/s @3.5-5m/s². This baby could thoroughly explore any body but the big three, mostly by means of suborbital hops. The killer feature you can't see is the large docking port in the bottom: the nukes not only powered the descent/ascent, but all transfers as well. Sometimes with the assistance of EVEN MOAR NUKES because I could afford it.

It wasn't build for Duna, but I found that it worked nicely. ISP in the deepest valley I could find was 760s -- or maybe even 780s, I'm not sure. It's been a while.

To recap: carrying a nuke all the way to Duna to only use it there is probably not worthwhile. If your whole dV requirement is a Duna ascent (plus a short puff for a soft landing), it's definitely not worthwhile. But if you happen to have a Nuke with you, Duna's gravity & atmosphere is no reason to stop using it.

Edited February 15, 2016 by Laie

[Snark]

12 minutes ago, Laie said:

The killer feature you can't see is the large docking port in the bottom: the nukes not only powered the descent/ascent, but all transfers as well.

...

But if you happen to have a Nuke with you, Duna's gravity & atmosphere is no reason to stop using it.

Certainly the cost is mitigated if you're using them for the interplanetary transfers as well-- any time in KSP that you can get one component to do double duty, it's a win.

14 minutes ago, Laie said:

If memory serves, it was 15t dry / 32t wet for not quite 6000m/s @3.5-5m/s².... It wasn't build for Duna, but I found that it worked nicely. ISP in the deepest valley I could find was 760s -- or maybe even 780s, I'm not sure. It's been a while.

Like I said, one can make it work.  The question is whether it works "better" than a lighter/stronger engine would.  The name of the game there is gravity losses, so how effective it will be totally depends on how heavy the thing you're lifting is.  e.g. when lifting off from Duna and going up to orbit, how heavy was it?  Did it have a mostly full load of fuel?  Or was it mostly empty?  Suppose it was mostly full, say, 30 tons mass on takeoff.  Under those circumstances, they'd only get 4 m/s² worth of thrust on takeoff, but Duna gravity is 2.94 m/s², so it would only be running at 26.5% efficiency with gravity losses, until it can really get going (which would take a while).  A pair of Poodles would get over four times the thrust; acceleration would be 16.67 m/s², for 82% efficiency with gravity losses.

Which one would be "better" overall would depend on a lot of factors (not just the takeoff weight, but also overall mission design).  Clearly this design worked for you, for the purpose you needed it for, and that's great.  I'm not trying to say "never use nukes" (I love 'em for orbital transfers, use them all the time) or even "never use them in landers" (I do that too, mainly on low gravity worlds.  They work great on Minmus, for example.)

All I'm saying is "they have low TWR and are vulnerable to taking really big gravity losses that can outweigh their benefits if you're not careful, so be sure you do the math if you're planning on using them somewhere that's not really low gravity."

 

[GeneralVeers]

2 hours ago, Snark said:

Not sure what you mean by that-- the fact that nukes don't use oxidizer is irrelevant to weight.

It means you don't have to carry oxidizer, which means less weight. What to do with the weight savings?? Well, if you design the way I do, that means "add more engines!"

[Snark]

10 minutes ago, GeneralVeers said:

It means you don't have to carry oxidizer, which means less weight. What to do with the weight savings?? Well, if you design the way I do, that means "add more engines!"

No, it doesn't mean less weight.  It doesn't mean "you don't have to carry oxidizer."  It means "you have to carry liquid fuel instead of oxidizer."

The dV that a rocket gets is dependent on two things:

• the Isp of the rocket
• the amount of mass that it expends

In the case of a conventional rocket, the mass is a mixture of LF and O.  In the case of a nuke, the mass is LF only.  But it's still the case that mass is mass.

Yes, nukes save on fuel mass, but it's because the Isp is higher, not because they're LF-only.  A pre-1.0 nuke would have gotten a certain amount of dV out of a tank containing 4 tons of LF+O mixture.  A post-1.0 nuke, running on LF only, would require 4 tons of LF to get that same dV.  It's exactly the same amount of required mass.  It's just a different mixture.

[GeneralVeers]

3 hours ago, Snark said:

No, it doesn't mean less weight.  It doesn't mean "you don't have to carry oxidizer."

And "you don't need oxidizer" means "less weight". Pre-1.0 nuclear engines, instead of being "you don't need oxidizer" were a case of "you need fuel and oxidizer, but only half as much". Which still means "less weight".

Just to make double sure, I booted up KSP and did some testing in sandbox mode. Two test rockets: the first one with a Rockomax X200-16 fuel tank (720 fuel and 880 oxidizer) and a Poodle engine; the other with a 400-unit jet fuel tank (i.e. liquid fuel only, no oxidizer) and a nuclear engine. Poodle was around 11 tons, Dr. Strangelove with the nuclear engine was 5.5 tons. Then I had an idea and doubled up Dr. Strangelove--two fuel tanks and two engines, that way both ships were about the same mass. Landed both on Duna in the same spot, and launched each into a 55k circular equatorial orbit to see which was better. So Poodle had the same mass as Dr. Strangelove, but slightly more than twice the thrust.

Despite what you've been saying, there was no significant difference in performace; each ship burned just under half its fuel load to reach stable orbit at 55k, and any "gravity losses" for Dr. Strangelove (with the heavier and weaker nuclear engines) were not measurable. Whatever your numbers say on paper, they don't stack up with an actual ship in the game. The two ships have about the same delta-V.

Myth busted.

Now, here's how the nuclear engines come out ahead: suppose you want to double the burn time of the above two ships? For Poodle, that means another Rockomax X200-16 tank. Mass, 9 tons. For Dr. Strangelove, that means two more Mk1 jet fuel tanks. Mass of those two tanks, FOUR AND A HALF TONS. Now you can stack 4.5 tons more payload on Dr. Strangelove than on Poodle and still get the same performance. Win!

[Plusck]

13 minutes ago, GeneralVeers said:

And "you don't need oxidizer" means "less weight". Pre-1.0 nuclear engines, instead of being "you don't need oxidizer" were a case of "you need fuel and oxidizer, but only half as much". Which still means "less weight".

If what you say is true, the rocket equation wouldn't apply to nuclear engines. However it does apply, as it should. Since oxidizer is part of the wet mass of a ship, and the wet mass minus dry mass is one of the two defining variables in the rocket equation, then what you are saying simply cannot be true.

The "wet" part is the propellant. The propellant can be fuel, it can be fuel+oxidizer, it can be polystyrene beads if you have some way of throwing them out the back end quickly enough. But the rocket equation doesn't care.

 

13 minutes ago, GeneralVeers said:

...Landed both on Duna in the same spot, and launched each into a 55k circular equatorial orbit to see which was better. So Poodle had the same mass as Dr. Strangelove, but slightly more than twice the thrust.

Despite what you've been saying, there was no significant difference in performace; each ship burned just under half its fuel load to reach stable orbit at 55k, and any "gravity losses" for Dr. Strangelove (with the heavier and weaker nuclear engines) were not measurable. Whatever your numbers say on paper, they don't stack up with an actual ship in the game. The two ships have about the same delta-V.

Myth busted.

[...]

You have successfully proved that doubling the ISP halves the amount of propellant needed for a given delta-v. And that over a given TWR, additional thrust doesn't make much difference.

[Laie]

3 hours ago, GeneralVeers said:

And "you don't need oxidizer" means "less weight". Pre-1.0 nuclear engines, instead of being "you don't need oxidizer" were a case of "you need fuel and oxidizer, but only half as much". Which still means "less weight".

You got that wrong. ISP basically means how much useful thrust you get out for one unit of mass you put in, or how much buck you get from your bang. The LVN has more than twice the ISP than the next best engine, hence uses less than half as much propellant for the same push. Whether that propellant is LF or LF+O or gummi bears is totally irrelevant.

[Snark]

5 hours ago, GeneralVeers said:

And "you don't need oxidizer" means "less weight". Pre-1.0 nuclear engines, instead of being "you don't need oxidizer" were a case of "you need fuel and oxidizer, but only half as much". Which still means "less weight".

You seem to be missing the point.  The type of fuel mixture is irrelevant to dV, mass, weight, or anything else.  A kilogram of reaction mass is a kilogram of reaction mass.  Switching them to be LF-only in 1.0 had zero effect on their performance, it just means you need a different kind of fuel tank.  Not a lighter one, just a different one.

Say you design a spacecraft so that it's LV-N powered, and you measure its performance (weight, dV, etc.) in some flight situation.  And then let's say you mod the engine so that it runs on LFO mixture instead (like it was pre-1.0), and you give it an LFO tank to replace the LF tank.  Guess what?  It will be exactly the same mass and have exactly the same performance with exactly the same dV.

 

5 hours ago, GeneralVeers said:

Just to make double sure, I booted up KSP and did some testing in sandbox mode. Two test rockets: the first one with a Rockomax X200-16 fuel tank (720 fuel and 880 oxidizer) and a Poodle engine; the other with a 400-unit jet fuel tank (i.e. liquid fuel only, no oxidizer) and a nuclear engine. Poodle was around 11 tons, Dr. Strangelove with the nuclear engine was 5.5 tons. Then I had an idea and doubled up Dr. Strangelove--two fuel tanks and two engines, that way both ships were about the same mass. Landed both on Duna in the same spot, and launched each into a 55k circular equatorial orbit to see which was better. So Poodle had the same mass as Dr. Strangelove, but slightly more than twice the thrust.

Despite what you've been saying, there was no significant difference in performace; each ship burned just under half its fuel load to reach stable orbit at 55k, and any "gravity losses" for Dr. Strangelove (with the heavier and weaker nuclear engines) were not measurable. Whatever your numbers say on paper, they don't stack up with an actual ship in the game. The two ships have about the same delta-V.

Myth busted.

Um, no.  You just mentioned 11 tons as the mass of a test ship consisting of, first, a 9-ton fuel tank, and second, a 1.75-ton Poodle.  Which means that your "spaceship" was essentially nothing more than engine, fuel tank, and probe core.

Ditto the case of your nuke craft, with 6 tons of engine and 4.5 tons of fuel tank.  Just engines and fuel tanks and nothing else.

What you've done is to try to optimize both of those for the highest possible TWR you can get from that engine combination, which minimizes the gravity loss which is the real killer for nukes-as-lander-engine.  Your nuke-only ship has 10.9 m/s² of acceleration, which is over 3 times Duna gravity and therefore minimizes gravity loss.

So yeah, if your message to people is that "Nukes are adequate Duna lander engines if your spaceship is nothing but engine with a small amount of fuel tank attached," then sure, I'll agree with that.  But I don't think that's what most people have in mind when they're building a  lander, so it doesn't seem particularly relevant here.

Try this test:  take both of those rockets, exactly as designed, and add a 20-ton payload on top.  Now how does it stack up?

5 hours ago, GeneralVeers said:

Now, here's how the nuclear engines come out ahead: suppose you want to double the burn time of the above two ships? For Poodle, that means another Rockomax X200-16 tank. Mass, 9 tons. For Dr. Strangelove, that means two more Mk1 jet fuel tanks. Mass of those two tanks, FOUR AND A HALF TONS. Now you can stack 4.5 tons more payload on Dr. Strangelove than on Poodle and still get the same performance. Win!

Yes, absolutely.  If what you want is to, for example, set up a high-dV interplanetary burn with that nuke-powered lander after achieving orbit, thus using LV-N's for what they're primarily good at, then yes, it would be a win.

Of course, that would be silly, because it would mean you just sent 6 tons of LV-N engines down to the surface and back which you didn't need-- could have made a far smaller and lighter lander, and left the LV-N's in orbit where they're more effective.

And would also not be relevant to the current discussion, which was about making something that goes between Duna surface and Duna orbit.  If you want much higher dV than that, then yeah, nukes are gonna win.  Put nukes in a situation where TWR doesn't matter, and of course they'll beat any other engine (other than ions) hands-down. It's why people (including me) use them a lot.

[Temstar]

1 hour ago, Snark said:

The type of fuel mixture is irrelevant to dV, mass, weight, or anything else. 

Let's not generalise too much. Whilst that's true for KSP's LV-N when it switched from LFO to LF, it's not true in general. In general, NTR's performance does depend on fuel mix - the lighter the average molar mass of the exhaust, the higher the Isp. And this does come up in KSP as well once you start to deal with mods that introduce cryogenic fuels.

[Snark]

7 minutes ago, Temstar said:

Let's not generalise too much. Whilst that's true for KSP's LV-N when it switched from LFO to LF, it's not true in general. In general, NTR's performance does depend on fuel mix - the lighter the average molar mass of the exhaust, the higher the Isp. And this does come up in KSP as well once you start to deal with mods that introduce cryogenic fuels.

Okay, fine.  FOR A GIVEN Isp, the type of fuel mixture is irrelevant to dV, mass, weight, or anything else.

There, happy? 

[GeneralVeers]

12 hours ago, Plusck said:

If what you say is true, the rocket equation wouldn't apply to nuclear engines.

Then you've got some explaining to do. Obviously the rocket equation does apply to nuclear engines, yet that is not what I saw when I performed an actual experiment. Real world trumps theory ("real world"? lol). If an experiment doesn't square with the theory and/or the equations, then it's the theory and/or the equations that must be wrong.

Truth is, I know what's "wrong". I did the whole time. Can you figure it out?

hint:

Spoiler

Inverse square

 

6 hours ago, Snark said:

Um, no.  You just mentioned 11 tons as the mass of a test ship consisting of, first, a 9-ton fuel tank, and second, a 1.75-ton Poodle.  Which means that your "spaceship" was essentially nothing more than engine, fuel tank, and probe core.

Gravity doesn't care. Gravity can't tell the difference between 11 tons of spacecraft and 11 tons of flank steak (wait--what??) Gravity sees both and says "hey, what do you think you're doing up there??" and tries to make both of them crash. The only difference here being that the second one will make a rather nasty mess when it impacts.......(either that or get flame-broiled by re-entry heat and be really tasty!)

250 Kn of thrust pushing 11 tons, versus 120 Kn of thrust, also pushing 11 tons. GRAVITY LOSSES BY THE SECOND CRAFT WERE NOT MEASURABLE. Case closed.

Bottom line: when choosing a rocket engine, there are some factors which do exist but which are in reality too small to be significant. They can be ignored in favor of other factors such as cost, convenience, and tendency to explode on a hard landing. So the first thing you do is run a few test flights to see which factors actually matter and which you can crumple up into a paper ball and attempt a three-point shot to the trash can.

[Snark]

2 minutes ago, GeneralVeers said:

Then you've got some explaining to do. Obviously the rocket equation does apply to nuclear engines, yet that is not what I saw when I performed an actual experiment. Real world trumps theory ("real world"? lol). If an experiment doesn't square with the theory and/or the equations, then it's the theory and/or the equations that must be wrong.

Truth is, I know what's "wrong". I did the whole time. Can you figure it out?

hint:

  Reveal hidden contents

Inverse square

250 Kn of thrust pushing 11 tons, versus 120 Kn of thrust, also pushing 11 tons. GRAVITY LOSSES BY THE SECOND CRAFT WERE NOT MEASURABLE. Case closed.

Exactly!  You've just proved my point, which is that you picked a poor example. 

The gravity losses aren't measurable because they're very small.  And the reason they're very small is that both craft have high local TWR.  And they're only able to have a high local TWR because they're unrealistic ships that are just stripped-down models that have nothing but a fuel tank and an engine.

The efficiency of an engine on vertical takeoff, taking gravity losses into account, is going to be 1 - 1/TWR.  Thus, in your (artificial, unrealistic) example:  the Poodle craft has a local TWR of 7.73, for an efficiency of 87%, whereas the LV-N craft has a local TWR of 73%.  In other words, very little difference between them-- yes, the LV-N ship would be slightly worse off for gravity losses than the Poodle ship would... but that would be far outweighed by the better Isp.  So the nuke wins in that case, yes.  But that's irrelevant for anybody who's launching something more... functional than a fuel tank strapped to an engine.

However, suppose we added some payload to those ships, to bring the total weight up to, say, 30 tons instead of 11.  (This is the example I gave earlier-- the one you're ignoring and didn't bother to do the math for.)  In this case, the Poodle craft would have an efficiency of 65%, whereas the LV-N ship has an efficiency of just 26%.  In that case, the Poodle has the edge.

[Plusck]

47 minutes ago, GeneralVeers said:

Then you've got some explaining to do. Obviously the rocket equation does apply to nuclear engines, yet that is not what I saw when I performed an actual experiment. Real world trumps theory ("real world"? lol). If an experiment doesn't square with the theory and/or the equations, then it's the theory and/or the equations that must be wrong.

Truth is, I know what's "wrong". I did the whole time. Can you figure it out?

hint:

  Hide contents

Inverse square

 

I was careful to separate the two things. (1) Your blanket statement about there being less weight due to there being no oxidiser (which is simply not true - the rocket equation forbids it and doesn't care what propellant you use), and (2) your real-world experience which just shows that between a high TWR and a very high TWR, gains are minimal.

I believe you when you say you knew what was wrong the whole time. However, I'm not sure of the benefits of writing things that are patently false in a subforum that is intended to help people understand and play the game.

[GeneralVeers]

1 hour ago, Snark said:

The gravity losses aren't measurable because they're very small.  And the reason they're very small is that both craft have high local TWR.

Bingo. Anywhere except low atmosphere on Kerbin (and Eve.....) the gravity losses you're complaining about simply don't matter. On the Mun, two Spark engines will do it for a ten-ton lander. The premier design consideration is to get enough TWR to get you out of the high-grav zone quickly, and you don't need a Chuck Norris engine for that on Duna.

 

1 hour ago, Snark said:

And they're only able to have a high local TWR because they're unrealistic ships that are just stripped-down models that have nothing but a fuel tank and an engine.

 Unrealistic ships? If only you'd known--Dr. Strangelove is based on the Duna lander I used for my first-ever Duna landing. Two nuclear engines, 720 units of liquid fuel, a tiny amount of oxidizer for the RCS system (Vernier jets), 3-Kerbal cockpit, solar panels, docking port, parachutes, landing gear, and science gear. Total mass, about 15 tons. A very realistic ship indeed--I actually used it.

 

1 hour ago, Snark said:

However, suppose we added some payload to those ships, to bring the total weight up to, say, 30 tons instead of 11.

Suppose we don't.

You cherry-picked exactly the right payload that your ship can lift but mine can't. You're trying to change my good design into a bad one and then force me to compare a bad nuclear design against a good non-nuclear one. Not happening. So yes, I'm ignoring your "proposal", and I'm going to keep ignoring it.

 

1 hour ago, Plusck said:

However, I'm not sure of the benefits of writing things that are patently false in a subforum that is intended to help people understand and play the game.

Then we're at a bit of an impasse, because we seem to disagree on who's the one saying things that are patently false.

 

5 hours ago, Plusck said:

and (2) your real-world experience which just shows that between a high TWR and a very high TWR, gains are minimal.

Exactly. Notice that the word "nuclear" was nowhere in what you just wrote. ALL engines perform badly with low TWR in a gravity well. This is not a problem with LV-N's. It's a problem with every engine in the game. With nuclear engines, there's more mass in the engine itself, but you save mass by getting the same burn time for slightly less than half the fuel. It's entirely a question of which factor is bigger.

[Snark]

1 hour ago, GeneralVeers said:

Unrealistic ships? If only you'd known--Dr. Strangelove is based on the Duna lander I used for my first-ever Duna landing. Two nuclear engines, 720 units of liquid fuel, a tiny amount of oxidizer for the RCS system (Vernier jets), 3-Kerbal cockpit, solar panels, docking port, parachutes, landing gear, and science gear. Total mass, about 15 tons. A very realistic ship indeed--I actually used it.

Yup.  Unrealistic.  Because people don't land ships like that, generally.  You didn't.  Your craft was 15 tons, not 11.  If that's the example you mean, then use that example to prove your point, not a stripped-down version.

15 tons vs 11 is a significant difference in efficiency.  It lowers the takeoff efficiency from 73% to 63% (the equivalent Poodle craft would be lowered less, from 87% to 82%).  The equivalent Poodle craft would have an extra 4.25 tons of fuel, enough to give it over 1100 m/s of dV... so basically it can get to orbit without even touching the remainder of the fuel that the nukes start burning at takeoff.

The 30-ton figure wasn't any more ridiculous than your 11-ton figure.  That's an actual ship, too-- it's the design that an earlier poster in this thread specifically mentioned.

Look, nobody's attacking your ship or your decisions.  And nobody is saying "you couldn't fly that"; after all, you did.  If it works for you, great.

But the fact is that massive, low-TWR engines have a constrained "usability envelope" when it comes to powering landers.  Unless it's a really low-gravity world, you have to know what you're doing to pull it off.  There are lots of things that could trip up a newbie there.  Lack of understanding of the efficiency effects of gravity losses would do it-- sure, you know not to put a lot more than 15 tons on that nuke powered lander, but a newbie might not. A non-nuke lander can be much smaller and lighter; the 15 ton lander you describe could be built for less than half the mass, carrying effectively the same payload and with the same dV. As it happened, that didn't matter much to you, because you were shipping it with a 300-ton behemoth, but it's likely that the newbie's not doing that-- the lander is very often the primary payload, and mass really matters. The earlier poster with the 30-ton rover pulled off a design coup by letting his nukes do double-duty as both the lander lifting engine and the interplanetary transfer mover-- newbie might not think of all that.

So giving advice to a newbie, I'm not going to recommend "use nukes".  Or, at the very least, if he's considering options, I'll point out that "nukes are tricky to use there, be careful, there are pitfalls."  Which is really all I've been trying to point out.

In any case, this topic has long since veered hopelessly away from the OP's question, since he's specifically indicated that he's not going to be landing this 15.4-ton ship on Duna anyay.

Edited February 16, 2016 by Snark