Developer Insights #17 - Engines Archetypes

[Intercept Games]

Engines: How to Avoid Shipping a Rocket Scientist

By Chris Adderley

We’ve mentioned approachability as a core pillar of our KSP2 design, and I’m here today to talk about one of the less-obvious ways we are focusing on helping players reach the stars.

An area we’ve noticed players struggling with in testing is making sense of the dizzying array of engines you’re presented with in the VAB. KSP1 had 35 engines for you to choose from (more if the Making History DLC is installed), spread across Liquid Fuel/Oxidizer, Liquid Fuel, Monopropellant, Xenon and Solid fuel types. This leads to a good deal of player confusion when starting out – what engine should I use? What engine is best for what I want to do? Why isn’t this rocket lifting off the pad even though I put 20 Terriers on it? There’s a lot of trial-and-error gameplay before you learn the hard-won lessons about specific impulse, thrust to weight ratio, and fuel density that can rocket you to success in KSP. Hah.

It unfortunately gets a bit worse. When you're looking for an engine, all of your important details are buried deep. You're searching for specific impulse, thrust, mass, heat production, and how the engine performs in multiple situations (sea level, orbit, other planets). It's a lot of work when you're learning!

When we look at our plans for KSP2, we’re only making this problem worse. We’re adding more engines, more fuel types and more engine sizes. Ouch. Clearly, we need to find good ways to teach new and returning players how to select an engine and teach players at the very least which engines are better at which missions they want to accomplish. I’m going to go into some detail on how we’re going to work towards addressing this, focusing in on the most common type of engines in KSP – the venerable liquid fuel engine category, which boasts such illustrious names as the Mainsail, Rhino and… Ant.

 

Liquid Fuel -> Methane

Before we get into this, a bit of terminology. Let’s start with talking about… methane and methane accessories. KSP1 gave us an abstracted resource to run our most common workhorse engines: the well-regarded Liquid Fuel . For KSP2, we’ve decided to take this resource and… name it. It’s methane. For their space program, Kerbals have passed over the brutish kerosene, toxic hypergolics and seductive lure of liquid hydrogen to settle on this nice middle ground fuel. It’s a good choice – a number of commercial companies are currently moving engines using methane and oxygen propellants to operational readiness.

When we talk about engines you might recall from KSP1 that sported the Liquid Fuel/Oxidizer moniker, we’re always talking about methalox engines. Yes, this nomenclature change applies to jet engines as well for simplicity, so jet engines are now methane engines.

 

Engine Archetypes

So, looking in detail at the methalox engines we have inherited from KSP1, we can see that we’ve got an interesting challenge on our hands. More than half of those 35 engines are methalox, and they’re practically the first engines a player gets introduced to. If we’ve done our job right, they’ll continue to be useful engines in some niche even after you have access to objectively more power engines, so they’ll stick around for a while. So, how to sort and help players determine how best to use them? I’ll present the concept of Engine Archetypes.

Rocketry fans will be familiar with three high-level types of liquid fuel engines. Firstly we have the high-thrust, high power engine which we can call the booster engine. These engines are great for getting a ship out of the atmosphere and pushing really heavy payloads, but don’t [SG16] have the efficiency to make them great deep space engines. Examples of this could be the Saturn V’s F-1 engines, or the Falcon 9’s Merlin engines.

 

Secondly, we have the sustainer type engine. This is typically a more efficient engine that burns for a longer duration, but doesn’t really have the oomph needed to throw heavy payloads into orbit without a little help. This type of engine is often paired with extra boosters of some type to get a kick up into orbit. Good examples of this include the Space Shuttle’s RS-25 engines and the Ariane series of rockets’ Vulcain engines.

 

Thirdly we have pure vacuum, orbit-only engines, best for operating in the cold depths of space and really, really efficient, but it will be lucky to push an overstuffed Kerbal though even thin atmospheres. A shining example here is the Aerojet RL-10 engine, which has existed for so long (early versions flew in the early 1960s, and the current version is used on the SLS rocket’s Interim Cryogenic Propulsion Stage) that it is basically the kitchen appliance of rocketry.

 

We can map these engine archetypes to KSP engines fairly well – see the following table.

Archetype	KSP Examples
Booster	Reliant, Mainsail, Mammoth
Sustainer	Swivel, Skipper, Rhino
Vacuum	Ant, Terrier, Poodle

 

This provides a good starting point for laying out KSP2’s methalox engine lineup.

 

Vacuum Engines – an aside

We’re always looking for opportunities to improve teaching about real rocketry concepts. One of the places KSP1 hasn’t quite lined up with the literature is the nature of the vacuum engines it uses. In reality, the shape and size of the nozzle attached to a rocket engine makes a big difference in terms of its performance at different atmospheric pressures. A good way of looking at this is to compare something called expansion ratio – which is a measure of the difference between the area of the engine’

s throat and its area of the nozzle exit. In vacuum, the ideal expansion ratio is extremely large – a good vacuum engine has a very narrow throat compared to its exit. To make a given engine work better in vacuum, we use a really big nozzle (though there’s obviously a lot more to it that just making your booster engine’s nozzle bigger).

Simplified rocket engines with small and high nozzle expansion ratios

 

Of course, reality sets in here because you can’t just add moar expansion ratio (a multi-kilometer wide nozzle might be a bit heavy) .Rocket scientists have tested novel concepts like the inflatable nozzle (look this one up), the hinged nozzle, and other creative ways of compressing nozzles so they become really big in orbit but can be launched with a smaller footprint. A working example of that is the RL-10B-2 engine that uses an extending lower nozzle cone that deploys once the rocket’s upper stage separates. You’ll see something like that in KSP2 with our NERV-US engine.

Unfortunately, KSP1’s vacuum engines are actually smaller than their atmospheric counterparts, which causes no end of consternation among the more technically minded of KSP players. This is a bit of dichotomy, because we all love using the Terrier and Poodle as lander engines due to their small footprint and suitability for landing legs. For KSP2, we will we be looking at moving towards a model that keeps these heritage KSP1 engines around as a subclass of engines that we’ll define as the Orbital class. These will maintain some level of excellence in space, get a bump to their atmospheric stats and leave the door wide open to the long, efficient Deep Space class of engine that lines up more with idealized vacuum engines – a new set we’ll be introducing through Early Access.

 

KSP2 Methalox Engine Archetypes

So, given all the above we have defined four engine archetypes: Booster, Sustainer, Orbital, and Deep Space.

 

With these archetypes in mind, we can design for them and use them to teach players. Players who know how to use Thrust and ISP to find the engines they want still have that information. More novice players can build to that point by first learning archetypes.

 

How do we teach archetypes? Well, here’s what we’re working on:

 

Terminology: We have aligned ingame terminology, like subtitles and descriptions, to specifically work on teaching player that any given engine belongs to specific archetypes. At Early Access you’ll for example see the Mainsail comes with a tagline of ‘Methalox Booster Engine’ that helps players situate it in the hierarchy of engines.

Archetype subtitle for the Terrier

Visuals: We have created specific design languages for each engine type, so picking up an engine and looking at it will be a good way to think about how it performs. Building these languages into our engine models is going to be an ongoing process through Early Access.

Balance and Tuning: We have mapped broad bands of engine characteristics to types, and then aligned many engines to better tell their stories. There are always strange engines, but they get to be strange because standard engines exist (like the Dart, that weird little aerospike guy).

Visual Language

Having good visual language for concepts is one of my passions. We want KSP’s rocket engines to be similar to, but not be real life engines. Reality is full of cool engines, and some of our engines hew very close to existing or conceptual designs. It's tempting to do that all the time, but the closer we lean to reality, the more the engines must skew to reality in all regards. I call this the "Why can't I build a space shuttle with three Vectors" problem.

In addition, we’re unlikely to have anything close to the great variety of fuels and tanks that reality has, so being very high fidelity with designs for engines creates disconnects for a detail-oriented realism players (this terrain is great for modders). Instead, when we’re looking at our archetype language for KSP, we will try to be a bit more general and inspired by real engines, rather than creating exact copies.

I’ve put together some sketches of these four archetypes to guide our artistic design going forward. The goal is for each of them to have a distinct visual look that is preserved through all size classes, and is versatile enough that, for example, a Mainsail doesn’t just look like a smaller Mammoth. We can pick and choose from a number of reality-alike design elements to create cool, Kerbal-native engines.

Booster engine features and possible design variations

 

Sustainer engine features and possible design variations

Orbital engine features and possible design variations

Deep Space engine features and possible design variations

 

The first place you’ll see this visual language in Early Access is the 3.75m engine lineup featuring the Labradoodle, Mammoth-II, and Rhino.

 

Applied design! From left to right, the Labradoodle, Mammoth-II and Rhino engines sporting, respectively, Orbital, Booster and Sustainer visual queues, courtesy of artists Jonathan Cooper and Pablo Ollervides.

 

Balance and Tuning

As we get to the end of this article, I wanted to touch on balance and tuning. Our guiding principles in tuning engines can be summed up with 3 points:

1. Don't deviate from KSP1 for the sake of it. A methalox rocket in KSP2 should perform similarly to a similar looking Liquid Fuel/Oxidizer rocket from KSP1
2. Engines of an archetype have similar characteristics.
3. Engines within a fuel type exist in a similar band of power, so newer or larger engines should not make older engines obsolete.

These rules still give us a lot of room for play while letting us increase approachability. Some engines, like the Vector, needed a hard look under these guidelines.

We’re basically trying to follow this chart, which I find a useful way of looking at the overall capabilities of engines. If an engine is a Methalox Sustainer, it should fall in the blue region, as an example – and we are really trying to keep things out of the Useless and Way Too Useful regions . The Way Too Useful region is a story for later in Early Access with more exotic engines , which have their own, unique challenges for building and flying.

 

Taken together, this means that outside of some specific areas, you won’t see massive statistical changes to most engines in KSP2 from KSP1, despite the naming change from Liquid Fuel and Oxidizer to Methalox. Places to watch out for are:

• KSP2’s 3.75m engines have had some overhauls to account for the addition of an Orbital engine in this size class (say hello to the Labradoodle, as named by Scott Manley!)
• KSP2’s Orbital engines have better atmospheric performance than their KSP1 counterparts.
• The relationship between the Mammoth (now Mammoth-II) and the Vector has been adjusted for KSP2, as they no longer need to match visually.

 

Putting it all Together

I can sum everything up using a table. Tables are almost my favorite things, narrowly being edged out by graphs.

Other Fuels

“But Chris!”, you say, “I thought KSP2 was about MORE than just Methalox?”. That’s absolutely true, and we’ll be looking to follow the same general rules when creating archetypes through other fuel types as we reveal things through Early Access.

[Vl3d]

I think everyone completely missed this post... Amazing!

PS1: Somewhat related to engines - a quick question. Because we have a few preset diameters and length steps for fuel tanks - can the length be adjustable from the tank properties instead of each fuel tank with the same diameter being a separate part? Not asking for completely procedural tanks, just step-wise for length. It would declutter the part picker list. Same for solid fuel boosters..

PS2: "For their space program, Kerbals have passed over the brutish kerosene, toxic hypergolics and seductive lure of liquid hydrogen..." - does this mean the mentioned fuel types are not going to be in the game? I'm surprised, because Nate mentioned liquid hydrogen as being a fuel type used for the nuclear thermal engines. A little confused..

PS3: Will there be fuel switching for engines, will we have engines that can use multiple fuel types? Is the Vector a methalox engine?

Edited December 16, 2022 by Vl3d

[Luriss]

Wait, it's been 19 hours and no one's posted? Huh

 

This post is really cool too. Gives a lot of insight into the thinking behind the engines that really shines through in the finished models. Great job!
Furthermore, communicating to the player what each engine does really is important even if it doesn't seem big deal at first (if you've ever tried playing with the KSP Interstellar mod you'll know what I mean).

[Domonian]

Those engines with that lighting look fantastic. I cannot wait to get my hands on that stuff! And as people have mentioned, why has it been 20 hours with 2 comments? I swear I didn't see this post when I checked yesterday, so maybe it was hidden?

[The Aziz]

The Labradoodle

Because, you know, it's the Poodle's cousin

Look at him go!

[MechBFP]

The topic was created yesterday and likely was just "un-hidden" shortly ago. Just an FYI that when this is done it looks like it tends to break the automatic notifications that occur when you are configured to follow a sub-forum. I am following the Dev Diaries sub forum for example but no notification got to me for this post.

I noticed a couple errors sneak through the editing process:
"but don’t [SG16] have the efficiency " <- editor note?

"between the area of the engine’

s throat and its area of the nozzle exit"  <- incorrect carriage return.

 

Anyway as to the topic of the post. I really liked this dev diary. As I go back and remember my first time playing KSP, I think that adding engine archtypes would have definitely helped me out without me needing to immediately dive into the specific stats like ISP to better understand what engines to use.

It is also good to see that attention is being provided to game play balance.

 

[机械主教71号]

20 hours ago, Intercept Games said:

Liquid Fuel -> Methane

Before we get into this, a bit of terminology. Let’s start with talking about… methane and methane accessories. KSP1 gave us an abstracted resource to run our most common workhorse engines: the well-regarded Liquid Fuel . For KSP2, we’ve decided to take this resource and… name it. It’s methane. For their space program, Kerbals have passed over the brutish kerosene, toxic hypergolics and seductive lure of liquid hydrogen to settle on this nice middle ground fuel. It’s a good choice – a number of commercial companies are currently moving engines using methane and oxygen propellants to operational readiness.

When we talk about engines you might recall from KSP1 that sported the Liquid Fuel/Oxidizer moniker, we’re always talking about methalox engines. Yes, this nomenclature change applies to jet engines as well for simplicity, so jet engines are now methane engines.

To be honest, I am quite disappointed to hear that there's only methane for the first day. I'd have been looking forward to diversifying liquid fuels for different tasks. Well, I'll wait.

Edited December 16, 2022 by 机械主教71号

[The Aziz]

I'm just laughing how mobile forum layout emphasized a detail in one paragraph

Did you know that was about engines? 

 

Just now, 机械主教71号 said:

To be honest, I am quite disappointed to hear that there's only methane. I'd have been looking forward to diversifying liquid fuels for different tasks. 

 

20 hours ago, Intercept Games said:

Other Fuels

“But Chris!”, you say, “I thought KSP2 was about MORE than just Methalox?”. That’s absolutely true, and we’ll be looking to follow the same general rules when creating archetypes through other fuel types as we reveal things through Early Access.

Some of you missed the last sentence.

[TheTennesseeFireman]

This seems like a really sensible layout. As a layman who was starting out (and is still a relative scrub who hasn’t played in a while), sensible rocket construction was just as big of a hurdle for me as orbital mechanics. Starting with archetypes designed around a basic, balanced fuel type that can get you to space will really help newcomers make sense of what a basic rocket should actually look like.

[机械主教71号]

37 minutes ago, The Aziz said:

I'm just laughing how mobile forum layout emphasized a detail in one paragraph

Did you know that was about engines? 

 

 

Some of you missed the last sentence.

I think that means xenon, solid fuel, monopropellant, hydrogenium, nuclear and other engines. What I want are kerosene, toxic hypergolics and liquid hydrogen engines, and I make it quite clear, 'diversifying liquid fuels'.

Edited December 16, 2022 by 机械主教71号

[JoeSchmuckatelli]

21 hours ago, Intercept Games said:

Players who know how to use Thrust and ISP to find the engines they want still have that information. More novice players can build to that point by first learning archetypes

Neanderthals of the Kerbolar universe rejoice! 

Now when someone on the forums tells me I'm doing it wrong... I might just understand why! 

[Ahres]

5 minutes ago, The Aziz said:

I'm just laughing how mobile forum layout emphasized a detail in one paragraph

Did you know that was about engines? 

The funny thing is that it took me a couple additional readings to understand what you were even talking about. Ha!

[JoeSchmuckatelli]

2 minutes ago, 机械主教71号 said:

I think that means hydrogenium, nuclear and other engines. What I want are kerosene, toxic hypergolics and liquid hydrogen engines.

He did say this was an area ripe for the modding community to address 

[机械主教71号]

11 minutes ago, JoeSchmuckatelli said:

He did say this was an area ripe for the modding community to address 

Well, wouldn't it be better to have it in the vanilla game? 

[Strawberry]

1 hour ago, Vl3d said:

PS2: "For their space program, Kerbals have passed over the brutish kerosene, toxic hypergolics and seductive lure of liquid hydrogen..." - does this mean the mentioned fuel types are not going to be in the game? I'm surprised, because Nate mentioned liquid hydrogen as being a fuel type used for the nuclear thermal engines. A little confused..

I'm pretty sure this is for chemical engines in specific. We'll be getting hydrogen propelled nuclear engines, just not hydrogen fueled chemical ones (except metallic hydrogen of course).

 

2 minutes ago, 机械主教71号 said:

Well, wouldn't it be better to have it in the vanilla game? 

The devs have commented on this in the past, and basically they just dont want to overwhelm players more, ksp1 is hard to learn as is. Its something thats easy to mod in and implement though, so its good mod territory as those who do want to add that extra level can.

[The Aziz]

We will have at the very least 5 types of fuel. Already more than KSP1 had. For a veteran player that may not be a problem <to sit swapping tank types to fit certain engine, having two or more of them on the same stage because the spacecraft happens to have more than one type of engine active> but for a newbie? It's a good window to thanks too complex goodbye.

[机械主教71号]

50 minutes ago, The Aziz said:

We will have at the very least 5 types of fuel. Already more than KSP1 had. For a veteran player that may not be a problem <to sit swapping tank types to fit certain engine, having two or more of them on the same stage because the spacecraft happens to have more than one type of engine active> but for a newbie? It's a good window to thanks too complex goodbye.

 You are right. Now I believe it will be better to keep 'Liquid Fuel' in KSP2, simple and clear, meanwhile, low Isp stands for toxic hypergolics, medium Isp stands for kerosene and methane, high Isp stands for liquid hydrogen. 

Wait a minute... If the liquid fuel become methane, do the jets also use methane as fuel? That's ridiculous...

Edited December 16, 2022 by 机械主教71号

[MARL_Mk1]

Can't wait for KSP2's Real Fuels mod then, I refuse to have a Methalox Vector

[Sennalen]

I'm glad they are making it something specific rather than mystery "liquid fuel". I would have rather it have been UDMH/NTO as the KSP1 equivalent, seeing that matches the properties of being dense, non-cryogenic, and ignites reliably. Also hypergols are still the dominant choice for orbital engines IRL. As long as its moddable, I guess that's okay.

Speaking of which, I don't love the statements about keeping the balance similar to KSP1. This would be the time to make a clean break and go for realistic physical values for the engines, then use the size of Kerbin as the difficulty tuning knob.

[TheOtherDave]

1 hour ago, 机械主教71号 said:

Well, wouldn't it be better to have it in the vanilla game? 

I could see an argument for having an “officially recognized” realism overhaul/realistic progression plugin, and I can even see an argument for that being a 1st party thing, but the vanilla game needs to be fun and have as wide an appeal as they can get without losing sight of the core vision, and this compromise seems reasonable to me.

Edited December 16, 2022 by TheOtherDave

[Pthigrivi]

Damn these look amazing.
 

23 hours ago, Intercept Games said:

...and fuel density that can rocket you to success in KSP. Hah.

Sir...
 

23 hours ago, Intercept Games said:

So, given all the above we have defined four engine archetypes: Booster, Sustainer, Orbital, and Deep Space.

Man I really appreciate this approach and the balance pass. I would say aside from the fine motor control needed for precision landing the biggest learning curve hurdle for me was learning to efficiently lift heavy things. Its all fine when you can more or less copy a Scott Manley tutorial build out of 1.25m parts but when you've got combos of different engines on different boosters and knowing when to drop them and how fast you can go before you lose stability... its a lot in the beginning. I think this plus letting players know early about the ~1.5 TWR  rule of thumb will make things much easier. I also really appreciate keeping the lower profile orbital class engines for landing. Can you tell us if they'll get a commensurate ISP nerf in exchange for the increased atmo performance? 

 

23 hours ago, Intercept Games said:

Having good visual language for concepts is one of my passions. We want KSP’s rocket engines to be similar to, but not be real life engines. Reality is full of cool engines, and some of our engines hew very close to existing or conceptual designs. It's tempting to do that all the time, but the closer we lean to reality, the more the engines must skew to reality in all regards. I call this the "Why can't I build a space shuttle with three Vectors" problem.

I also really like this and the way that the various styles 'feel' like what they're meant to do. I think 1:1 historical analogs probably would have been weird anyway if we're consolidating our LF rockets into Methalox. What did the solution for the Vector end up being? Looks like maybe its in the sustainer category?


 

[Vl3d]

2 minutes ago, Pthigrivi said:

What did the solution for the Vector end up being? Looks like maybe its in the sustainer category?

[Pthigrivi]

Just now, Vl3d said:

Ahh there it is

[Tangle]

Oh G-d, only methalox? Guess there aren't gonna be any engines older than 2010 in KSP2...

 

The making of the 4x RS-25 into a really big F-1B is a... really weird decision.

[ChubbyCat]

So is the Mammoth II replacing the original mammoth, or will they exist alongside each other? It would be only slightly disappointing to not be able to create accurate SLS recreations anymore.