Engines TWR and ISP

[kerbalfreak]

I'm making some very efficient rocket SSTOs (much faster and easy to launch and land than planes - and maybe more realistic, because the game ignores the heat damage of the spaceplanes). And I am looking in more detail at the engines specs.

Seems that the Vector, the Mammoth and the Aerospike can beat any other engine specs (TWR, Isp, Gimbal) for sea level, with very good efficiency at vacuum for the Aerospike.

Well, this is balanced in a Sandbox game? Despite matching the size of a certain payload, these 3 engines are better than all the rest for sea level? If I use Tweakscale, this will make all other stock engines obsolete?

Thanks!

 

[Rocket In My Pocket]

Keep in mind engines have their own weight, a powerful engine is often very heavy as well, this can have a huge impact on the final performance of the design.

Very generally speaking; high TWR=low Dv. You'll often want to use the weakest, lightest engines you can get away with for efficiency.

Can't really speak to Tweakscale as I don't use it but yeah; balance goes out the window when you start adding mods in to the mix obviously.

[Gaarst]

KSP is not balanced. The parts values are placeholders that were never replaced (word of a developer).

That being said, the Vector, Mammoth and Aerospike all have their flaws.
First they are all very late-tech and expensive so not the first choice if money is a concern (obviously this doesn't apply to sandbox).
Second, the Vector and Mammoth are extremely heavy and powerful, too much for smaller payload; you'll find that a smaller engine with a lower Isp might be more efficient than a larger (heavier) engine with a higher Isp for some situations. When you're in space, TWR doesn't matter that much and a large engine is just dead mass. For an SSTO, mass is a major concern so you might want to try using lighter engines to see if you can get better performance (though it's not as important in rocket SSTOs).
Third, the Aerospike has no gimbal and the Vector has too much. KSP's SAS is terrible and most of the time is can't handle the full gimbal of the Vector so you end up limiting it. Having zero gimbal is not too much of a concern in stock since reaction wheels are stupidly OP, but it's still more pleasant to fly a craft that can stabilise itself.

The rest is mostly a matter of personal preference: if you want absolute best performance then you will end up using the same lifter engines (mostly because we barely have any alternatives for each size), but some people prefer to use other engines for better looks when efficiency is not the only objective.

I have limited experience with Tweakscale, so I can't really tell how it impacts "balance".

[OhioBob]

14 hours ago, Plusck said:

All things considered, I'd expect most players will find the Poodle to be the "best" engine... if the mass and form of the craft allow it to be used. Or maybe the Rhino for bigger ships, it really is a beauty. Or maybe the Terrier because it is just so damned useful.

The Rhino is both expensive and heavy for what it provides.  I consider it one of the worst engines in the game in terms of both mass and cost efficiency.  That doesn't stop me from using it, however.

Although it rarely gets any love, one of the best engines is the Twin-Boar.  I think it tends to get overlooked because its specific impulse doesn't knock your socks off, and just how good it is in terms of mass and cost efficiency isn't immediately apparent because its numbers include the fuel tank.  If we subtract out the fuel tank (which is effectively a Jumbo-64), we begin to see just how good it really is.  Compared to other engines in its size class, here's how it stacks up:

	Twin-Boar	Mainsail	Rhino
Cost	11,250	13,000	25,000
Mass	6.5	6	9
Thrust, SL	1,867	1,379	1,500
Thrust, vac	2,000	1,500	2,000
Isp, SL	280	285	255
Isp, vac	300	310	340

[kerbalfreak]

I'm making some very efficient rocket SSTOs (much faster and easy to launch and land than planes - and maybe more realistic, because the game ignores the heat damage of the spaceplanes). And I am looking in more detail at the engines specs.

Seems that the Vector, the Mammoth and the Aerospike can beat any other engine specs (TWR, Isp, Gimbal) for sea level, with very good efficiency at vacuum for the Aerospike.

Well, this is balanced in a Sandbox game? Despite matching the size of a certain payload, these 3 engines are better than all the rest for sea level? If I use Tweakscale, this will make all other stock engines obsolete?

Thanks!

 

[Aegolius13]

The Vector is really overpowered; the attempt to "balance" it in Career was to give it an absurdly high price tag.   But that doesn't matter much if you're playing  sandbox,  if you plan to recover the craft and get a refund, or if you just have a ton on money banked.  The Mammoth is even more overpowered, and a lot more cost efficient,  but bigger than you need for a lot of jobs. 

Anyhow, yeah, if you have a reasonably large ship,  these are likely the best engines to use on an SSTO.   The biggest bottleneck is thrust at launch, and the Vector and Mammoth can't be beat in that regard. 

[regex]

27 minutes ago, Gaarst said:

KSP is not balanced. The parts values are placeholders that were never replaced (word of a developer).

IIRC the engines were rebalanced around 1.0 by @NathanKell (?) to better match the theoretical maximums of Aerozine50/NTO engines, with the fuel now defined as having the specific heat properties of said chemicals and generally in line with how engines work in KSP (deeply throttleable, infinitely restartable, with fuels that are infinitely storable). The jet engines and nuclear engine are outliers but the lack of detail in the game makes that a fairly moot point.

As far as "sandbox balance" KSP never had that (which might be the most realistic thing about it, lol), there have always been engines and parts that overshadow others. In true sandbox fashion you should use the engines and parts that best fit your play style.

Edited November 20, 2017 by regex

[kerbalfreak]

1 hour ago, Gaarst said:

KSP is not balanced. The parts values are placeholders that were never replaced (word of a developer).

That being said, the Vector, Mammoth and Aerospike all have their flaws.
First they are all very late-tech and expensive so not the first choice if money is a concern (obviously this doesn't apply to sandbox).
Second, the Vector and Mammoth are extremely heavy and powerful, too much for smaller payload; you'll find that a smaller engine with a lower Isp might be more efficient than a larger (heavier) engine with a higher Isp for some situations. When you're in space, TWR doesn't matter that much and a large engine is just dead mass. For an SSTO, mass is a major concern so you might want to try using lighter engines to see if you can get better performance (though it's not as important in rocket SSTOs).
Third, the Aerospike has no gimbal and the Vector has too much. KSP's SAS is terrible and most of the time is can't handle the full gimbal of the Vector so you end up limiting it. Having zero gimbal is not too much of a concern in stock since reaction wheels are stupidly OP, but it's still more pleasant to fly a craft that can stabilise itself.

The rest is mostly a matter of personal preference: if you want absolute best performance then you will end up using the same lifter engines (mostly because we barely have any alternatives for each size), but some people prefer to use other engines for better looks when efficiency is not the only objective.

I have limited experience with Tweakscale, so I can't really tell how it impacts "balance".

Yes, the Mammoth has the best TWR, Vector a huge gimbal and Aerospike no gimbal but great Isp at vacuum. The 3 can beat any other engine specs.

If placing 3 or more aerospikes, that are not linear, a mod called TCA can control the craft by controlling independently the thrust of each engine. I simple love it.

I play sandbox with mods, mostly to build more good looking and maybe realistic ships, but I really hate when something is simple better than all the other alternatives, without any drawback. For me tweakscale is a must, and I find sad that some engines turn complete useless. Will see if I can balance it somehow, or I will don't use it to scale engines down.

Thanks!

Edited November 20, 2017 by kerbalfreak

[Gaarst]

5 minutes ago, regex said:

IIRC the engines were rebalanced around 1.0 by @NathanKell (?) to better match the theoretical maximums of Aerozine50/NTO engines, with the fuel now defined as having the specific heat properties of said chemicals and generally in line with how engines work in KSP (deeply throttleable, infinitely restartable, with fuels that are infinitely storable). The jet engines and nuclear engine are outliers but the lack of detail in the game makes that a fairly moot point.

What I remember is that a developer (might have been NathanKell as well) mentioned that these 1.0 tweaks were to be temporary and eventually replaced by a more thorough engines rework. I believe this was at the time of Porkjet's departure and cancellation of his overhaul, so things were a bit messy but it was made clear that we wouldn't get an engines rework and so that the placeholder values would remain.

[RX2000]

6 minutes ago, kerbalfreak said:

I play sandbox with mods, mostly to build more good looking and maybe realistic ships, but I really hate when something is simple better than all the other alternatives, without any drawback. For me tweakscale is a must, and I find sad that some engines turn complete useless. Will see if I can balance it somehow, or I will don't use it to scale engines down.

You should try career. The high price of some engines can really be a drawback. "Do I want to use this huge insanely powerful engine to complete this contract that doesnt give me much money?"

[regex]

4 minutes ago, Gaarst said:

What I remember is that a developer (might have been NathanKell as well) mentioned that these 1.0 tweaks were to be temporary and eventually replaced by a more thorough engines rework. I believe this was at the time of Porkjet's departure and cancellation of his overhaul, so things were a bit messy but it was made clear that we wouldn't get an engines rework and so that the placeholder values would remain.

Huh, well, the so-called "placeholder" values are much better than the state of affairs we had before 1.0 so I tend to think of them as a decent balance pass.

The problem is that you simply can't have engines that fill every "niche" in a space game because the number of "niches" is fairly limited. Instead you end up with overlap and engines that are more suited to early tech tree than late, etc... which remove any sort of "balance" from sandbox. That's fine, that's just like how real life works, where you upgrade from the MB-3 to an RS-27 because it's clearly better and you stop making the MB-3 because there's no point to it.

Edited November 20, 2017 by regex

[Helmetman]

Haven't used tweakscale much myself. I don't think it will matter much. Tweakscale does exactly what the name says, it scales things.

So part values should be maintained I think, but I'm guessing here as I've only used TS a couple of times. My take on this so far...

But if those values are maintained you won't get any benefit other then the benefit of the square cube law. I haven't magnified how numbers of wet and dry mass scale up with TS. But tweakscaling something should change the square cube values even in KSP, right? 

Also, SLT matters (sea level thrust), but not much. The difference between vacuum thrust and atmospheric thrust parameters shown on the part information list is the ISP difference between 1 atmosphere (the Launchpad) and space (70 km up)

The ISP gains immediately as you gain altitude, and you only need a few kilometers altitude to be halfway to your vacuum ISP. 
It's better to have a vacuum optimized rocket rather then a atmospheric optimized rocket.
Unless you takeoff from Eve, have a very high twr rocket with a very early gravity turn I wouldn't even waste my time optimizing it for optimal atmospheric efficiency. You often do better giving your rocket more thrust and making it more aerodynamic then to nitpick a engine for the 1st minute of flight, but you can obviously.

I wouldn't use the Aerospike ever, but I did a few times on Eve. It has good atmospheric isp, yes! But they're quite heavy for the thrust they provide.
I wouldn't necessarily advice against non gimballing rocket engines like the aerospike. Some good placed fins far enough down and some adequate reaction wheels and a well timed gravity turn will get you to orbit trust me. But I simply wouldn't because the aerospike just isn't the best choice imho.

 

[Aegolius13]

7 minutes ago, RX2000 said:

You should try career. The high price of some engines can really be a drawback. "Do I want to use this huge insanely powerful engine to complete this contract that doesnt give me much money?"

True for an expendable stage.    But if you go SSTO (OP's intended purpose) and are able to land near KSC, even the cost doesn't really mean much since you'll get almost everything back.  

4 minutes ago, Helmetman said:

The ISP gains immediately as you gain altitude, and you only need a few kilometers altitude to be halfway to your vacuum ISP. 
It's better to have a vacuum optimized rocket rather then a atmospheric optimized rocket.
Unless you takeoff from Eve, have a very high twr rocket with a very early gravity turn I wouldn't even waste my time optimizing it for optimal atmospheric efficiency. You often do better giving your rocket more thrust and making it more aerodynamic then to nitpick a engine for the 1st minute of flight, but you can obviously.

I disagree with this to some extent.  Yes,  you spend more time in thinner air.   But in an SSTO, your thrust is most constrained at launch.   In any craft,  TWR will improve as you climb and burn fuel.   This means if you go for a vacuum oriented launch engine,  you'll need a disproportionately big one you get off the ground, and will have tons of unneeded thrust later.   And that mass will hurt your efficiency even if UPS is good. 

Of course, if there was a huge difference in vacuum ISP, that could play a role.   But the Vector, Mammoth etc. are not that much less efficient than even the vacuum oriented engines - like 10%.

[NSEP]

Yes, maybe. The thrust might go up through the roof, if you scale them up, and since the ISP (likely) wont change, you can probably make true monsters. I think the best way to answering your question is to try it out!

[GoSlash27]

Kerbalfreak,

 I'm not sure I understand the concept of "game balance" where sandbox is concerned. If you are free to build whatever you want with no financial or practical limitations, what does it matter which parts you use, or which parts you don't?

The Aerospike outperforms comparable engines if you ignore the cost, but the Vector and Mammoth don't. The Vector is deceptive in that it's small form factor belies it's mass. It is really comparable in mass and thrust to the Skipper and Mainsail.
 When you factor in the cost (career mode) and mass (career and science), these engines get used a lot less frequently.
*edit* There are a lot of different ways to define "efficient", especially when comparing career designs to sandbox designs.
Best,
-Slashy

Edited November 20, 2017 by GoSlash27

[Plusck]

As GoSlashy says, there are a lot of factors to consider.

All things considered, I'd expect most players will find the Poodle to be the "best" engine... if the mass and form of the craft allow it to be used. Or maybe the Rhino for bigger ships, it really is a beauty. Or maybe the Terrier because it is just so damned useful.

Or maybe the Nerv. There is no substitute if you need to propel Kerbal-sized objects to extreme ranges, like a 180° inclined orbit somewhere near Moho. Only the LV-N can give your command pod the requisite 15km/s dv with acceptable burn times and launchable masses.

Personally, I love the Aerospike. It's a shame that it has no gimbal, meaning that it really needs a perfectly balanced craft, but that's part of the joy of designing the ship it will power. But if you're trying to return from the surface of Eve you need to mix Aerospikes with Vectors, there is no other reasonable option. I really never touch the Vector for anything else.

And when it comes to launching craft from Kerbin, I have a very clear preference for the Skipper. It's so much cheaper than the alternatives for that size of vessel, and I know it loses some of its power and ISP at low altitudes but it very quickly recovers and it is far, far cheaper than the far heavier alternatives (i.e. Mainsail... never liked that one at all)...

[Spricigo]

3 hours ago, Plusck said:

As GoSlashy says, there are a lot of factors to consider.

All things considered, I'd expect most players will find the Poodle to be the "best" engine... ...Or maybe the Rhino... ....Or maybe the Terrier... ....Or maybe the Nerv.

I can't say what other players think. But for me Kickback, Reliant, Spark and Spider are excellent engines, often much better than anything in your list. 

[GoSlash27]

As usual, I concur with OhioBob. The Twin-Boar is the most economical lf&o booster in the game. When combined in parallel with Kickbacks, it can lift huge payloads extremely cheaply.