Please explain to me, what all the LV-N hype is about?

[Laie]

Sort of off topic, IRL, didn't they use superheated water as the propellant? Or am I mis-remembering?

I've only ever heard of that in SciFi novels, most recently Seveneves. The ISP should be lower than with Methane, where 600+ seems feasible with just a little bit of future tech (keep your salt cellar handy). Pouring water into a Nerva-like device (that is, something we could build right now) would give you a worse ISP than an LF+O engine. The only benefit is that water is comparatively easy to come by.

[Red Iron Crown]

Sure: Pre-1.0, LV-Ns handily beat basically every other other engine for efficiency.

They still do. (Excluding air breathers and ions.)

This changed in 1.0, and now some players are disappointed that "Use nukes" isn't the answer to maximize dV in every case.

It never was, though the number of scenarios where other engines are better has increased.

The game state now is better for the change, and people will get over it in time. 1.0 also nerfed some other engines like the 48-7S that ruined similar "use this for everything" situations.

Agreed, there's a niche for almost every engine now.

[Nich]

I don't know, skippers and mainsails just don't do anything for me. Below 23 tons (second stage) I use SRBs to launch. Above I use twin boars because there shorter which makes things more stable for me when I get up to that size then the mainsail equivalent.

[Guest]

Do I have to do this every time?

Yup, I think I do.

[tutike2000]

Quick fix: edit your Squad folder's nuke engine part.

Give it 4x the thrust, 4x the cost, 4x the mass, 4x the heat generation, etc etc

Now nukes no longer suck. Also, you'll need some proper heat management now.

I love the Nuclear Age engines, but the large one's ISP is way too large (1400)

[Xyphos]

I must refer to UmbralRaptor's entirely accurate "Stages of Rocket Design":

Sounds like the OP is at step 1 or 2. When you get to steps 3 and 4, suddenly the LV-N becomes awesome.

Actually, I've been using poodles or skippers for my upper stages, the only time I attempted to actually use LV-N was in a spaceplane design due to it's fuel requirement I figured it would be better to use than aerospikes, but I found the engine was too heavy for that, which prompted me to post the OP and try to figure out their purpose.

[Red Iron Crown]

Actually, I've been using poodles or skippers for my upper stages, the only time I attempted to actually use LV-N was in a spaceplane design due to it's fuel requirement I figured it would be better to use than aerospikes, but I found the engine was too heavy for that, which prompted me to post the OP and try to figure out their purpose.

They're not great for ascents, where TWR is often a significant factor and their lack of thrust can really hurt. They are best for maneuvers after orbit is established, where TWR matters far less and Isp becomes much more important.

[Terwin]

Do I have to do this every time?

http://youtu.be/WoiVej1rccs

Yup, I think I do.

The self-starting and started/stopped 10 times is really impressive, from what I understand that is one of the more technically challenging issues with using the same engine for multiple parts of a mission.

I wonder how well and how fast it could be throttled...

It's a shame it is no longer politically feasible to use nuclear rocket engines(Even RTGs are a political challenge, apart from not having much fuel left)

[Orbital Vagabond]

It never was, though the number of scenarios where other engines are better has increased.

Yes, you can nit-pick that my statement wasn't perfectly accurate, but the old nuke design was so overwhelming superior the statement does stand.

Agreed, there's a niche for almost every engine now.

It's objectively a better game state and this is largely due to the beating the crap out of the LV-N with a nerf bat, though other changes helped.

[Red Iron Crown]

It's objectively a better game state and this is largely due to the beating the crap out of the LV-N with a nerf bat, though other changes helped.

It is. I'd almost have preferred a different sort of nerf for it, like requiring heavy shielding to prevent a lethal dose of radiation. Though maybe that is what the extra mass represents.

[Violent Jeb]

I must agree that the LV-N does not excel in many areas. As pointed out on the first page, because of their 4.5t weight, they are generally not effective for payloads below 10t.

For comparison, I typically send a 14t lander to the mun or duna, which can seat two kerbals, and 1 of every experiment, with sufficient power. As you expand operations, your payload can quickly exceed 30t.

As a general rule, I refuse to burn (even interplanetary) with less than 0.4 TWR (my optimal is 0.6), because I see diminished results from long and inaccurate burns. I like to be able to do any burn in one shot, and usually less than 120s - so my kerbals always get where they're going.

As your weight slowly increases beyond 30t payload, the number of LV-Ns (and associated parts) will proportionally increase. As a rough benchmark, if your craft requires 10 LV-N engines, it will need 10 connecting tanks, 10 fuel lines and maybe even 10 struts. This means that using LV-Ns in even with a remotely reasonably sized configuration you have an interplanetary stage that is easily 50 parts once you include telemetry, docking ports, sas, rcs fuel/thrusters. Once you have a launcher and a payload, you can easily clear 200 parts, which is too many for many people.

So far, we have a very part-heavy craft with a low TWR. Consider that in career mode this craft will also be easily 50% more expensive than a comparable LFO/OX rocket. The ONLY drawback to having to use LFO/OX rockets is you need to bring more fuel. One part usually accomplishes this need.

As i will demonstrate with an image from http://meithan.x10.mx/KSP/engines/, using 10 engines as a reasonable limit for most users, it becomes clear that as your payload exceeds a meager 30t, you WILL NOT get more than 6k ÃŽâ€V out of your LV-N, without substantially increasing the number of engines and parts.

Alternatively, we have the Rhino, with a respectable 345 isp in vaccuum. This engine WILL get more than 6000 ÃŽâ€V, and usually with only 1-3 engines. as your payload increases beyond 30t, the payload fraction actually grows with the rhino. When fuel is less expensive and uses less CPU resources than efficiency, the payload fraction is largely moot anyway.

It seems to me there is a very narrow band where the LV-N is superior. It appears to be best (and indeed only useful) in the range of 10-30t crafts if you don't mind longer burns and smaller TWR.

[Red Iron Crown]

That "narrow band" appears to be most of the chart, a chart which ignores staging.

Great analysis though, and your points about part count and burn times are right on. Not everyone wants to split up burns or sit through long ones to save a bit of mass.

[Laie]

Great analysis though, and your points about part count and burn times are right on. Not everyone wants to split up burns or sit through long ones to save a bit of mass.

As far as I'm oncerned, it's not the mass holding me back but the part count. Violent Jeb alluded to it above: the mass you save usually is one or two large tanks, but you do that at the price of adding numerous 60kN engines and their supporting bits. It's entirely possible to have 0.4g from Nukes (and 6km/s goes a loooong way), but what good is a decent TWR if your game plays at 12FPS?

[NecroBones]

I must agree that the LV-N does not excel in many areas. As pointed out on the first page, because of their 4.5t weight

Correction: 3 tons.

[Jouni]

As a general rule, I refuse to burn (even interplanetary) with less than 0.4 TWR (my optimal is 0.6), because I see diminished results from long and inaccurate burns. I like to be able to do any burn in one shot, and usually less than 120s - so my kerbals always get where they're going.

My typical design goal is 6000 m/s of delta-v and initial TWR 0.2, which corresponds to about 9 tonnes of payload per LV-N. If you do a transfer burn from LKO to Jool with that, you'll probably spend 10-15% extra delta-v. That's much more efficient than carrying around unnecessary dry mass in the form of engines.

(Of course, it's usually more convenient to launch the interplanetary ship on an oversized rocket, and use the upper stage for the initial ~1000 m/s of the transfer burn.)

As your weight slowly increases beyond 30t payload, the number of LV-Ns (and associated parts) will proportionally increase. As a rough benchmark, if your craft requires 10 LV-N engines, it will need 10 connecting tanks, 10 fuel lines and maybe even 10 struts. This means that using LV-Ns in even with a remotely reasonably sized configuration you have an interplanetary stage that is easily 50 parts once you include telemetry, docking ports, sas, rcs fuel/thrusters. Once you have a launcher and a payload, you can easily clear 200 parts, which is too many for many people.

My typical pre-1.0 transfer stages had 2-12 engine nacelles (LV-N, fuel tank, fuel line, reaction wheels, and a nose cone) around a large central tank. Add a probe core, a docking port, and some utility parts, and we're talking about 0.6-0.8 parts per tonne of payload.

These days the main limiting factor for using the LV-N are the strange shapes of the LF tanks. Radial symmetry doesn't work too well with them, so you can no longer build simple transfer stages. As a result, I haven't completed that many interplanetary missions in 1.0.

So far, we have a very part-heavy craft with a low TWR. Consider that in career mode this craft will also be easily 50% more expensive than a comparable LFO/OX rocket. The ONLY drawback to having to use LFO/OX rockets is you need to bring more fuel. One part usually accomplishes this need.

If you play on moderate, typical contract rewards are 10-20x higher than the price of the big and expensive rocket you built. Efficiency and cost-effectiveness don't really matter in the game, as long as you generally know what you're doing.

As i will demonstrate with an image from http://meithan.x10.mx/KSP/engines/, using 10 engines as a reasonable limit for most users, it becomes clear that as your payload exceeds a meager 30t, you WILL NOT get more than 6k ÃŽâ€V out of your LV-N, without substantially increasing the number of engines and parts.

12 engines is a practical upper limit for simple designs, because you can easily add 12 engine nacelles around a 3.75 m fuel tank. That means 108 tonnes of payload with around 6000 m/s of delta-v, around 70 parts for the transfer stage, and total ship mass around 360 tonnes. With 20% payload fraction for the launch vehicle, we're talking about a 1800-tonne rocket, which might mean 6 Mammoth-based boosters and a central stack with a Rhino, or 65-70 parts.

The Rhino-based interplanetary ship would weigh over 2000 tonnes, so you would need a gigantic 10000-tonne rocket with hundreds of parts to launch it.

[Violent Jeb]

That "narrow band" appears to be most of the chart, a chart which ignores staging.

Great analysis though, and your points about part count and burn times are right on. Not everyone wants to split up burns or sit through long ones to save a bit of mass.

Does the "narrow band" still appear to be "most of the chart" when you look at all of the potential payloads, and not just the one the LV-N is good at? This image shows conclusively that anything above 50t is not going to be reasonable using the LVN.

Additionally, you mentioned staging. How many stages should an interplanetary stage have? if you're dropping engines mid-burn that messes up your maneuvers even more. I prefer my interplanetary burns in one stage, so the charts are well suited to that analysis.

If you play on moderate, typical contract rewards are 10-20x higher than the price of the big and expensive rocket you built. Efficiency and cost-effectiveness don't really matter in the game, as long as you generally know what you're doing.

12 engines is a practical upper limit for simple designs, because you can easily add 12 engine nacelles around a 3.75 m fuel tank. That means 108 tonnes of payload with around 6000 m/s of delta-v, around 70 parts for the transfer stage, and total ship mass around 360 tonnes. With 20% payload fraction for the launch vehicle, we're talking about a 1800-tonne rocket, which might mean 6 Mammoth-based boosters and a central stack with a Rhino, or 65-70 parts.

The Rhino-based interplanetary ship would weigh over 2000 tonnes, so you would need a gigantic 10000-tonne rocket with hundreds of parts to launch it.

I don't care how you slice it, you will not move 100+ t interplanetary payloads quickly (accurately) or with low part counts, using the LV-N.

I've been launching 800 t crafts with 50-80t payloads using mammoths and rhinos and my part counts typically do not go above 150. The same simply cannot be done with any sort of effectiveness using the Lv-N at that scale. The LV-N certainly needs more benefits to outweigh the considerable drawbacks.

Edited June 19, 2015 by Violent Jeb

[Guest]

I wonder how well and how fast it could be throttled...

I'm a pilot, not a nuclear physicist, but the rate of throttle control could be calculated by taking the time it takes to mechanically turn the control rods, added to the time of sufficient neutron absorption to stop the fission reaction, added to the time the reactor cools to a temperature safe enough to shut off the fuel flow.

Add all of those variables together, and you get the amount of time it takes to go from 100% power to 0% power.

[Jouni]

I don't care how you slice it, you will not move 100+ t interplanetary payloads quickly (accurately) or with low part counts, using the LV-N.

I've been launching 800 t crafts with 50-80t payloads using mammoths and rhinos and my part counts typically do not go above 150. The same simply cannot be done with any sort of effectiveness using the Lv-N at that scale. The LV-N certainly needs more benefits to outweigh the considerable drawbacks.

Just look at the Meithan's engine charts carefully. If we take the 6000 m/s / TWR 0.4 / max 10 engines as our goal, the transition from nuclear engines to Rhinos happens when the payload exceeds 29.5 tonnes. Below the line, we can achieve the goal with a 153-tonne ship using nuclear engines. If we increase the payload marginally (but still keep it at 29.5 tonnes), we need a 5x bigger 762-tonne ship to reach the goal with Rhinos.

That 5x gap basically tells one thing: the 6000 m/s / TWR 0.4 / max 10 engines goal makes no sense. If you need a lot of delta-v from a single stage, and you can't have high enough TWR with a reasonable number of engines, there's no reasonable way to reach the goal with any other kind of engines. If you need a lot of delta-v, you either have to stage or to accept a low TWR.

[NecroBones]

Additionally, you mentioned staging. How many stages should an interplanetary stage have? if you're dropping engines mid-burn that messes up your maneuvers even more. I prefer my interplanetary burns in one stage, so the charts are well suited to that analysis.

Ideally you wouldn't be staging engines often. More likely, you'd be wanting to use drop-tanks.

[Laie]

I wonder how well and how fast it could be throttled...

Brotoro has brought up actual numbers in one of the other LV-N threads since 1.0: the official documents of the time assume 30 seconds for shutdown. That includes dealing with the worst post-shutdown decay heat. After that, ordinary cooling systems were supposed to suffice.

[Red Iron Crown]

Does the "narrow band" still appear to be "most of the chart" when you look at all of the potential payloads, and not just the one the LV-N is good at? This image shows conclusively that anything above 50t is not going to be reasonable using the LVN.

It does, with three large caveats:

- The player is pushing large payloads (mine rarely exceed 50t)

- The player demands a 0.4 initial TWR (I kick periapses senseless, so much so that I wrote a tool for it, I tolerate TWRs as low as 0.25)

- The player demands 10 engines or less (This is very much dependent on PC specs and personal lag tolerance, I've made ships with over 30 LV-Ns that worked well enough)

These are playstyle choices, and of course they may be what works best for you, but that doesn't make it universal. Here's what my chart for the same dV/payload range looks like:

Very different, because my performance parameters are different.

Additionally, you mentioned staging. How many stages should an interplanetary stage have? if you're dropping engines mid-burn that messes up your maneuvers even more. I prefer my interplanetary burns in one stage, so the charts are well suited to that analysis.

I dislike staging events during burns, too, it's just simpler to do them well without them. But I almost never make an ejection burn that is more than 3km/s, so I don't really need more than that in a single stage. I find it is almost always worthwhile to stage away empty tanks (though I keep my nukes), especially right after the ejection burn when TWR is low and the dry mass dropped is high. Again, playstyle.

I don't care how you slice it, you will not move 100+ t interplanetary payloads quickly (accurately) or with low part counts, using the LV-N.

Here's the thing: My split burn strategy with a 0.25 TWR will be more accurate than a single burn with a 0.4 TWR for any ejection burn other than Moho. Why? I know how to calculate when to start periapsis kicking, and my final burn will be about 950m/s less than a single burn strategy. This means my last burn will almost always be shorter than a single burn done with a 0.4 TWR, and thus more accurate. Part count is subjective, adding a bunch of LV-Ns doesn't make performance intolerable for me (and makes for a smaller overall vessel, requiring fewer struts, a smaller lifter, is easier to turn etc).

Many players don't want to bother with such a strategy for ejection (it certainly takes more game time), and that is fair game. But for the way I play, LV-Ns are an excellent engine once the destinations are outside Kerbin's SoI.

I've been launching 800 t crafts with 50-80t payloads using mammoths and rhinos and my part counts typically do not go above 150. The same simply cannot be done with any sort of effectiveness using the Lv-N at that scale. The LV-N certainly needs more benefits to outweigh the considerable drawbacks.

It's certainly fair to say that you don't find them useful, but I'm not sure it's fair to say that that makes them useless for everyone.

[Violent Jeb]

You're right, I don't believe in doing periapsis kick burns. I've been under the guise that each 20 min orbit is going to sufficiently adjust my ejection angle such that a single well timed launch is more ideal. I would be happy to update that viewpoint. How do you go about determining which burn will be the final burn, and where will your ejection be with relation to the target? If you're saving 900 m/s on a burn i'd like to know how to keep the ejection angle in check. (i'm aware we're slightly off topic, but i'd be happy to incorporate a .25 TWR design parameter using LV-Ns if I can still get around the solar system with minimal fuss

Generally, I can't speed up time to a suitable rate for these burns, unless my orbit is 120km+ (and IRL time is by-and-large the most valuable resource for me). In that case, would the gains from multiple periapse burns offset the ÃŽâ€V losses from utilizing the oberth effect at a higher altitude AND the adjusted ejection angle? As of now i settle at 70-75km orbit.

I'm sure I will get around to testing it myself but it's easier to start doing things "the right way". Plus there's the task of an interplanetary with 10+ engines..

Edit: While i'm here and editing. The hypothetical burns I talked about are almost always greater than 0.6. How long (or how far) does the burn ETA have to be from periapsis to make you decide to do multiple burns (burn ETA > 1min, 3min, 5min. etc)?

Edited June 19, 2015 by Violent Jeb

[Laie]

You're right, I don't believe in doing periapsis kick burns. I've been under the guise that each 20 min orbit is going to sufficiently adjust my ejection angle such that a single well timed launch is more ideal. I would be happy to update that viewpoint. How do you go about determining which burn will be the final burn, and where will your ejection be with relation to the target?

Red has a link in his signature. You may want to check it out.

But for what it's worth, here's how I do it:

• launch well ahead of time (at least three days, and depending on need and tolerance, up to 40)
  
• place maneuver node for a single-burn transfer in n days
  
• set the alarm clock to the time of burn, +-0 seconds
  
• plot a prograde burn real soon now at the position of the node
  
• now my periapsis is where I need to do the final burn
  
• do repeated burns every time I pass the periapsis
  
• finally:
  
• adjust the magnitude of the last burn so that I will reach periapsis when the timer on the clock runs out
  
• optional: adjust inclination on the last apoapsis (only worthwhile if your apo is well beyond the Mun, otherwise it's usually better to add some normal to every periapsis burn)
  

It is fully sufficient if you get the placement of the periapsis "about right". You shouldn't just slap down the nodes any old how, but extreme precision doesn't pay off. Just remember that you don't have to slavishly place the transfer burn at your periapsis. The eventual maneuver node may be a whole minute ahead or behind of your periapsis without any noticable extra cost on the burn.

Edited June 20, 2015 by Laie

[Red Iron Crown]

You're right, I don't believe in doing periapsis kick burns. I've been under the guise that each 20 min orbit is going to sufficiently adjust my ejection angle such that a single well timed launch is more ideal. I would be happy to update that viewpoint. How do you go about determining which burn will be the final burn, and where will your ejection be with relation to the target? If you're saving 900 m/s on a burn i'd like to know how to keep the ejection angle in check. (i'm aware we're slightly off topic, but i'd be happy to incorporate a .25 TWR design parameter using LV-Ns if I can still get around the solar system with minimal fuss

It was exactly those concerns that motivated me to find a more accurate way to kick periapses, so I did a bit of math to come up with the precomputed burns here.

Generally, I can't speed up time to a suitable rate for these burns, unless my orbit is 120km+ (and IRL time is by-and-large the most valuable resource for me). In that case, would the gains from multiple periapse burns offset the ÃŽâ€V losses from utilizing the oberth effect at a higher altitude AND the adjusted ejection angle? As of now i settle at 70-75km orbit.

I use 100km as my parking orbit to make rendezvous easier during assembly and give a bit of leeway for changes in periapsis when kicking, I don't think the difference in Oberth effect is overly significant between 100km and 75km (though I'll admit I haven't run the numbers), if it was a big concern I guess I could recalculate for 75km. I'll be honest though, if IRL time is most important to you then you'll be better served with what you are doing now, the same dV is being expended at lower TWR so the total burn time is longer and there's those interim orbits to warp through.

I'm sure I will get around to testing it myself but it's easier to start doing things "the right way". Plus there's the task of an interplanetary with 10+ engines..

Edit: While i'm here and editing. The hypothetical burns I talked about are almost always greater than 0.6. How long (or how far) does the burn ETA have to be from periapsis to make you decide to do multiple burns (burn ETA > 1min, 3min, 5min. etc)?

I try to keep burns about 2 minutes for kicks and about 5 minutes for the ejection. For closer transfers like Eve and Duna I can usually just set the TWR just a bit higher and do a single 5 minute ejection, for further away I often kick instead.

Edited June 20, 2015 by Red Iron Crown

[PB666]

Cuz I just don't get it. they're small underpowered engines that don't need oxidizer but we don't have any suitable LF-only tanks outside of spaceplane parts.

I attempted to use them but I find them to be quite lacking.

Plus

-800 dV in space

Minuses

1.-Do not work well in vacuum

2.-They have low thrust given their massive weight, they can only lift at 1 g 2wice their weight.

3.-No dedicated LF only tanks

4.-They get really really hot.

5. Because of the added weight, and because these are final stage engines, they can create alot of launch wobble

Solutions

1&2. Don't use the as a launch engine except from low-g planets with no vacuum. Use good landing stuts, LV-N do not like to be used as landing struts.

3. Duplicate and rename the cfg of any tank you want to fuel the LV-N then edit the new cfg.

- Change the name parameter to [tankname] & "LFonly".

- Change the name of the title to [title] & "Lf only version".

- Remove the "Oxidizer" resource

- Double the amount and maximum values for the "Liquid fuel" resource

-Or (if your a purist or lazy)-

-after you add the tank in VAB right click the part and remove the green in the Oxidizer GUI. You will be carrying around on average 10% more weight in fuel facilities than you needed.

4. Place a heat shield between the LV-N and the tanks and run a fuel line from outmounts on the tank to the LV-N

5. In VAB LV-N should set on 18A decoupler which then goes on a F1-2 converter. Connect a strut from the side of the converter nearest the F2 to the outmounts on the fuel tanks above the LV-N.

Note that many use LV-N on F2 tanks like the big orange fuel tank. In these cases you place the converter on top of the big orange, you might even mod the converter to hold fuel and place the big orange upside down on top of the converter and place the 18A and a nose cone (or just a nose cone and burn it off) on top of the upside down engine. The assembly will need a controller like MechJeb mounted somewhere. LV-N on a big orange is a very good way to carry loads between worlds that have a high dV relative to kerbin (such as Moho and Jool). LV-N used on fuel carriers may not need the protection of a heat shield.

- - - Updated - - -

I use 100km as my parking orbit to make rendezvous easier during assembly and give a bit of leeway for changes in periapsis when kicking, I don't think the difference in Oberth effect is overly significant between 100km and 75km (though I'll admit I haven't run the numbers), if it was a big concern I guess I could recalculate for 75km. I'll be honest though, if IRL time is most important to you then you'll be better served with what you are doing now, the same dV is being expended at lower TWR so the total burn time is longer and there's those interim orbits to warp through.

75 km you should be going for 70km.

That level of planning is too much for me.

One thing I did not like about MechJeb, is it would waste alot of fuel trying to create phasing. Like you I keep my targets at 100 km. When I launch I try to get Apo at 100 and bring Pe to 70. Then I just let the whole stuff orbit until the target is close to my ship Apo, then I increase the Pe to make the next orbit cross exactly. I waste no fuel to phase. Since the targets period is longer than the vessel, its best to launch as the target is crossing over, then let the vessel catch the target.

The only thing I let MechJeb do is Smart SAS. To much wasted fuel, to many target crashes and too many time trying to literally cross a planet in order to reach a desired orbit.