KSP Stock Engines

[something]

For some reason, I felt like plotting the Thrust of KSP engines vs their specific impuls. It is basically a table of "How efficient is the boom you selected?" Nothing surprising, but I figured, somebody might be interested in it, nonetheless...  Except for Jet engines all values are vacuum data. Jet engines feature highest thrust possible. RAPIER and Panther come with one entry for each mode. Pay attention for the double logarithmic scale, though.

The plot can be split in four regions. In the top left you do have the Liquid and Solid fuel engines. They come with various amounts of thrust at Impulses up to 340s. In the bottom left, you have the RCS thrusters - low thrust and low impulse. The top right, that's the Jet engines - unfortunately you cannot use them in vacuum. In the bottom right you do have the Ion engine. I am not really sure weather it is a Jet engine without thrust or an efficient RCS thruster. Reading stories of hour long injection burn times, I come to believe it was intended as an RCS thruster...

And a zoom into the liquid fuel engines in the top left corner.

 

The conclusion you ask? I guess, I should implement price and weight in a future plot. But unless I figure how to present 5-dimensional data in a meaningful way, I am kind of stuck, I guess...

 

[steuben]

5 dimensional data can be represented as a moving 3d plot with changing colour

 

[AeroGav]

As a spaceplane person,  I find it hard to get worked up about the (seemingly minor) differences between different liquid fuel rocket engines.    

Genuine question,  would there be much difference between a large rocket launched with only Terriers and Reliants vs one with the best Liquid Fuel / Oxidizer engines available ?  Obviously,  using mammoths and Rhinos would reduce the part count drastically, but in terms of payload fraction how much difference would it make?

[KerikBalm]

plot TWR instead of thrust, its still 2 axis

[mk1980]

14 hours ago, AeroGav said:

As a spaceplane person,  I find it hard to get worked up about the (seemingly minor) differences between different liquid fuel rocket engines.    

Genuine question,  would there be much difference between a large rocket launched with only Terriers and Reliants vs one with the best Liquid Fuel / Oxidizer engines available ?  Obviously,  using mammoths and Rhinos would reduce the part count drastically, but in terms of payload fraction how much difference would it make?

imo payload fraction isn't such an important metric when designing a rocket. when i bother to optimize a disposable lifter, i optimize it for cost.  if i can get a payload to orbit using a stack of SRBs and that solution turns out cheaper than an alternate design with LFO engines, i couldn't care less which of the solutions has a lower weight

[something]

Rocket design needs to take into account two stages of usage: Launch to orbit and travel in vacuum.

During your interplanetary travels, you want to have a high Isp with low engine mass to increase your dV. During launch to orbit, you can sacrifice a some m/s of dV as you can stage away the unused mass during ascent, but you want to have high thrust with low engine weight to get of the pad.

The mammoth weighs 15t and gives you 4MN of thrust, the Isp is 315s. That's roughly 266 N/kg and 21 s/t.

The Reliant comes at 192 N/kg and 248s/t.  The Terrier gets 240N/kg and 690s/t.

So comparing those three engines, the Mammoth would be a good lifter to orbit but a horrible engine in vacuum. The Terrier is nice for travel in space,  however.

When doing spaceplanes, you can't just throw stuff away as you like, so you need engine combinations that work well. That's one of the reasons for the popularity of the RAPIER since it is basically two engines combined to one part, meaning you save some engine weight.

[KerikBalm]

On 9/14/2017 at 3:31 PM, mk1980 said:

imo payload fraction isn't such an important metric when designing a rocket. when i bother to optimize a disposable lifter, i optimize it for cost.  if i can get a payload to orbit using a stack of SRBs and that solution turns out cheaper than an alternate design with LFO engines, i couldn't care less which of the solutions has a lower weight

Unless of course its a SSTO rocket that you can recover

[Snark]

On 9/14/2017 at 5:01 AM, KerikBalm said:

plot TWR instead of thrust, its still 2 axis

^ This.  The thrust of an engine isn't all that interesting a measure of its oomph-- after all, if an engine has a tiny thrust, you can just spam multiple of it, right?

Much more interesting is TWR, which is a better measure.  Plot your graph with engine TWR rather than engine thrust on the Y axis, it'll be a considerably more interesting plot, IMO. 

Another interesting plot might be Isp versus cost-per-kN (i.e. TCR instead of TWR, C for "cost").

 

[A_name]

On 13/9/2017 at 5:49 PM, AeroGav said:

As a spaceplane person,  I find it hard to get worked up about the (seemingly minor) differences between different liquid fuel rocket engines.    

Genuine question,  would there be much difference between a large rocket launched with only Terriers and Reliants vs one with the best Liquid Fuel / Oxidizer engines available ?  Obviously,  using mammoths and Rhinos would reduce the part count drastically, but in terms of payload fraction how much difference would it make?

Depends on what you mean by "best" engine. Best for what? Using your example, the Terrier would be a great choice for space maneuvers once you're out of the atmo, provided your plane is not too heavy. As to the Reliant, there are very few scenarios where this engine is actually the best option, the obvious one being when you’re building a 1.25 m lifter need just a bit of extra thrust on liftoff. The swivel is better in most all other applications, due to its gimbaling capacity and its higher Isp.

 

As to how much difference it makes, I’m sure it wouldn’t be so hard to do the math and find out exactly how much Dv you waste by going with a lower efficiency or heavier set of engines, but in terms of gameplay feel, I can tell you from my own experience it does make a difference. That said, for spaceplanes you’re almost always better off just going with RAPIERS as your only LFO engine, because any benefit of including a more efficient engine will likely be offset by the Dv losses of carrying around the extra mass. The exception being the Nerv for long-haul flights.

 

Another thing to note is that engine performance rises to near-maximum values at around 8 km altitude, so ASL values are only really relevant for the first stage of a rocket lifter.

 

On a personal note, my favorite engine is the Rhino, extremely efficient and a very good thrust. Only disadvantage of course is its size. IMO the 2.5 m engines are mostly useless (i.e. you’re better off going with another size option altogether), except for the Twin Boar on lifter first stages. In the 1.25 m size my favorite is the Aerospike due to its good efficiency across all flight conditions.

[AeroGav]

53 minutes ago, A_name said:

As to the Reliant, there are very few scenarios where this engine is actually the best option, the obvious one being when you’re building a 1.25 m lifter need just a bit of extra thrust on liftoff. The swivel is better in most all other applications, due to its gimbaling capacity and its higher Isp.

Reliant has higher sea level ISP than the swivel, higher thrust and lower weight.    As an aircraft nut,  when i do build rockets i'm not shy about putting control surfaces on the bottom so the lack of gimbal is not a problem.   Oh and it's also cheaper than a Swivel.

From the perspective of someone who mostly designs spaceplanes, the differences in various chemical rocket engines seem small and unimportant.

For example,  Panther vs Rapier at 100 m/s,   Panther is far far superior,  more than double TWR.     OTOH at 1000 m/s you could definitely say the tables have turned, because the Panther will have flamed out and won't make any thrust at all,  while by that point the RAPIER will be nearly 8x the thrust it was making at 100 m/s.    Similarly, comparing RAPIER (airbreathing)  with NERV,  the RAPIER has 4x the ISP, and much higher TWR - until you get over 25km,  at which point it's impersonating a candle and makes the NERV look like some kind of powerhouse.   Now when you've deorbited on Laythe and are cruising around looking for somewhere to land -   RAPIER 3200 ISP, Panther (dry) 9000 ISP.

As you can see the differences between the right and the wrong engine for a given situation are more than a single digit percentage increase in ISP.   For the record,  I tend to use a mix of panthers, rapiers and nervs on my current designs.  Panther (wet) to get us supersonic.   Rapier really gets going just as the panther is quitting.   NERV takes care of things when airbreathing is no longer possible.   Dry panther for atmospheric cruising flight.

[sdj64]

52 minutes ago, AeroGav said:

From the perspective of someone who mostly designs spaceplanes, the differences in various chemical rocket engines seem small and unimportant.

I think this is closer to true with staged rockets for getting off of Kerbin than for other applications.  There are many engines that can lift a payload to orbit, and not a huge difference whether you use a big cluster of Reliants, a Delta 4 heavy with mainsails, or a single Mammoth, unless you're in career mode and you're worrying about cost.

Building a reusable lander for Tylo, a single stage rocket to Minmus, or an Eve ascent vehicle simply cannot be done with some (most) engines.  For landers especially, choosing the wrong engines could force it to be much bigger than it could be with the right engines.  The more you push the limits, the more a "single percentage point in ISP" becomes the deciding factor whether you can or can't do something.  For spaceplanes, the choice in engines is usually clear and the deciding factor is minimizing drag.  And of course, reducing dry mass is good for every type of craft.

[something]

5 hours ago, Snark said:

Much more interesting is TWR, which is a better measure.  Plot your graph with engine TWR rather than engine thrust on the Y axis, it'll be a considerably more interesting plot, IMO. 

It basically says, the RAPIER is an efficient jet engine with a much too high TWR...also the RV-105 RCS block is horrible...

 

 

 

5 hours ago, Snark said:

Another interesting plot might be Isp versus cost-per-kN (i.e. TCR instead of TWR, C for "cost").

Yes, sure you can have that....here we go. Don't use Ions. They have a high cost to thrust (CTR) ratio. The Ion engine gives you 4200s for 3 funds/Newton. The Rapier gives you 305s for just 0.03 funds per Newton. Also, SRBs are cheap in terms of CTR.

 

All plots - including zooms into the LF area - can be found here:

https://postimg.org/gallery/2jtaoyvl2/

[Pand5461]

On 19.09.2017 at 8:10 PM, AeroGav said:

Reliant has higher sea level ISP than the swivel, higher thrust and lower weight.    As an aircraft nut,  when i do build rockets i'm not shy about putting control surfaces on the bottom so the lack of gimbal is not a problem.   Oh and it's also cheaper than a Swivel.

With all that in mind, I actually switched to SRBs for first stages. Cheap as dirt, horrible Isp, great thrust - just the right thing to haul a Terrier or Poodle upper stage to 10-15 km and take full advantage of high Isp engine from there.

[Rune]

Yup, you guys are basically re-deriving the classical rocket science mantras:

-The most important figure of merit for a first stage engine is TWR.

-The only important figure of merit for an in-space engine is Isp.

-SRBs are the thing to get cheap thrust off the pad.

The second one is kind of obvious (more Isp=more dV= you get further), but the first one is not any less true. A higher TWR can do absolute wonders to your total takeoff mass, not least because it will minimize gravity losses, which are the ones that dominate in pretty much all but the tiniest rocket designs.

 

Rune. Bottom line: for pure LFO performance, and screw cost and scale, nothing beats a Mammoth (first Eve SSTO, first true SSTA). The Vector is the close second that is actually easy to configure for a given mission.

Edited September 22, 2017 by Rune