1.02 LF&O engine comparo

[GoSlash27]

Now that 1.0 has dropped and rebalanced all of the engines, I thought it would be helpful to analyze the capabilities of the engines in their new form and see how they stack up against each other in various uses. A lot has changed and we have some surprising new "best" engines.

They are placed in a standard situation with a given DV budget, and evaluated by their payload fraction. The lighter an entire stage is in comparison to it's payload, the less mass the preceding stages will have to lift. Generally, this results in small and inexpensive rockets on the pad and maximizes your capabilities.

I start with the values in the config files for the engine's mass, thrust, and Isp. Based on the desired Isp and payload, I calculate how much fuel and tankage is required. I then adjust the payload and reiterate until it shows that one engine is required to accelerate the total mass at the desired G. Finally, I record the payload and total mass, and divide the payload into the mass to give payload fraction.

"Best" engines exhibit payload fractions of >95% of the single best performer.

"Good" are 90-95%

"Fair" are 85-90%.

All are shown with their payload per engine in tonnes and payload fraction in percent.

Scenario #1: Single main core stage with SRBs.

The DV budget 3,600 m/sec, vacuum Isp and 1.0 t/w ratio.

"Excellent" engines

[TABLE]

[TR]

[TD]KR-2L Rhino[/TD]

[TD]47[/TD]

[TD]23.1[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-L10 Poodle[/TD]

[TD]5.7[/TD]

[TD]22.4[/TD]

[/TR]

[/TABLE]

"Good" engines

[TABLE]

[TR]

[TD]LV-909 Terrier[/TD]

[TD]1.28[/TD]

[TD]21[/TD]

[/TR]

[/TABLE]

"Fair" engines

[TABLE]

[TR]

[TD]RE-I5 Skipper[/TD]

[TD]13.6[/TD]

[TD]20.6[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]KS-25x4 Mammoth[/TD]

[TD]83.5[/TD]

[TD]20.5[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]T-1 Aerospike[/TD]

[TD]3.76[/TD]

[TD]20.5[/TD]

[/TR]

[/TABLE]

Scenario #2: Two stage upper stage

DV Budget 1,800 m/sec, vacuum Isp and 1.0 t/w ratio.

"Excellent" engines

[TABLE]

[TR]

[TD]KR-2L Rhino[/TD]

[TD]98[/TD]

[TD]48.2[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]KS-25x4 Mammoth[/TD]

[TD]188[/TD]

[TD]46.2[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-L10 Poodle[/TD]

[TD]11.6[/TD]

[TD]45.8[/TD]

[/TR]

[/TABLE]

"Good" engines

[TABLE]

[TR]

[TD]RE-M3 Mainsail[/TD]

[TD]69.3[/TD]

[TD]45.4[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-909 Terrier[/TD]

[TD]2.67[/TD]

[TD]43.8[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]T-1 Aerospike[/TD]

[TD]7.95[/TD]

[TD]43.4[/TD]

[/TR]

[/TABLE]

"Fair" engines

[TABLE]

[TR]

[TD]KR-1x2 Twin-Boar[/TD]

[TD]87.9[/TD]

[TD]43.2[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]48-7S Spark[/TD]

[TD]0.78[/TD]

[TD]42.6[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-T45 Swivel[/TD]

[TD]8.66[/TD]

[TD]42.5[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-T30 Reliant[/TD]

[TD]9.26[/TD]

[TD]42.3[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]24-77 Twitch[/TD]

[TD]0.674[/TD]

[TD]41.4[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]Mk-55 Thud[/TD]

[TD]5.01[/TD]

[TD]41[/TD]

[/TR]

[/TABLE]

Scenario #3: Three stage mid or upper stage

DV Budget 1,200 m/sec, vacuum Isp, and 1.0 t/w ratio

"Excellent" engines

[TABLE]

[TR]

[TD]LV-909 Terrier[/TD]

[TD]9.93[/TD]

[TD]61.9[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]KR-2L Rhino[/TD]

[TD]122[/TD]

[TD]60[/TD]

[/TR]

[/TABLE]

"Good" engines

[TABLE]

[TR]

[TD]KS-25x4 Mammoth[/TD]

[TD]238[/TD]

[TD]58.6[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-M3 Mainsail[/TD]

[TD]88.4[/TD]

[TD]57.9[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-I5 Skipper[/TD]

[TD]38.3[/TD]

[TD]57.9[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-L10 Poodle[/TD]

[TD]14.4[/TD]

[TD]56.8[/TD]

[/TR]

[/TABLE]

"Fair" engines

[TABLE]

[TR]

[TD]KR-1x2 Twin-Boar[/TD]

[TD]113[/TD]

[TD]55.6[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]48-7S Spark[/TD]

[TD]1[/TD]

[TD]54.8[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-T30 Reliant[/TD]

[TD]11.9[/TD]

[TD]54.5[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]T-1 Aerospike[/TD]

[TD]9.93[/TD]

[TD]54.2[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]24-77 Twitch[/TD]

[TD]0.877[/TD]

[TD]53.9[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-T45 Swivel[/TD]

[TD]10.9[/TD]

[TD]53.8[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]Mk-55 Thud[/TD]

[TD]6.43[/TD]

[TD]52.7[/TD]

[/TR]

[/TABLE]

Scenario #4: Long- distance interplanetary mass- mover

DV budget 6,000 m/sec, vacuum Isp, 0.5 t/w ratio

"Excellent" engines

[TABLE]

[TR]

[TD]RE-L10 Poodle[/TD]

[TD]3.16[/TD]

[TD]6.2[/TD]

[/TR]

[/TABLE]

"Good" engines

[TABLE]

[TR]

[TD]KR-2L Rhino[/TD]

[TD]22.9[/TD]

[TD]5.64[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-909 Terrier[/TD]

[TD]0.683[/TD]

[TD]5.59[/TD]

[/TR]

[/TABLE]

"Fair" engines

Scenario #5: Short range spaceflight (within the Kerbin system)

1500 m/sec DV, vacuum Isp, 0.5G t/w

"Excellent" engines

[TABLE]

[TR]

[TD]KR-2L Rhino[/TD]

[TD]230[/TD]

[TD]56.5[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-L10 Poodle[/TD]

[TD]28.3[/TD]

[TD]55.6[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]KS-25x4 Mammoth[/TD]

[TD]443[/TD]

[TD]54.5[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-909 Terrier[/TD]

[TD]6.64[/TD]

[TD]54.4[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-I5 Skipper[/TD]

[TD]72[/TD]

[TD]54.4[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]T-1 Aerospike[/TD]

[TD]19.7[/TD]

[TD]53.9[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-M3 Mainsail[/TD]

[TD]164[/TD]

[TD]53.8[/TD]

[/TR]

[/TABLE]

"Good" engines

[TABLE]

[TR]

[TD]LV-T45 Swivel[/TD]

[TD]21.4[/TD]

[TD]52.6[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]KR-1x2 Twin-Boar[/TD]

[TD]212[/TD]

[TD]52.1[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]48-7S Spark[/TD]

[TD]1.89[/TD]

[TD]51.7[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-T30 Reliant[/TD]

[TD]22.6[/TD]

[TD]51.6[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]Mk-55 Thud[/TD]

[TD]12.4[/TD]

[TD]51.1[/TD]

[/TR]

[/TABLE]

"Fair" engines

[TABLE]

[TR]

[TD]24-77 Twitch[/TD]

[TD]1.64[/TD]

[TD]50.6[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-1 Ant[/TD]

[TD]0.206[/TD]

[TD]50.6[/TD]

[/TR]

[/TABLE]

Scenario #6: Munar ascent or descent stage

800 m/sec DV, 1.5G Munar t/w, vacuum Isp

"Excellent" engines

[TABLE]

[TR]

[TD]KR-2L Rhino[/TD]

[TD]607[/TD]

[TD=width: 86, bgcolor: #7030A0]74.3[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-L10 Poodle[/TD]

[TD]75.4[/TD]

[TD=width: 86, bgcolor: #7030A0]73.8[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-909 Terrier[/TD]

[TD]17.9[/TD]

[TD=width: 86, bgcolor: #7030A0]73[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]KS-25x4 Mammoth[/TD]

[TD]1190[/TD]

[TD=width: 86, bgcolor: #7030A0]72.9[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-I5 Skipper[/TD]

[TD]193[/TD]

[TD=width: 86, bgcolor: #7030A0]72.9[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]T-1 Aerospike[/TD]

[TD]53.5[/TD]

[TD=width: 86, bgcolor: #7030A0]72.7[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-M3 Mainsail[/TD]

[TD]444[/TD]

[TD=width: 86, bgcolor: #7030A0]72.4[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-T45 Swivel[/TD]

[TD]58.6[/TD]

[TD=width: 86, bgcolor: #7030A0]71.7[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]KR-1x2 Twin-Boar[/TD]

[TD]583[/TD]

[TD=width: 86, bgcolor: #7030A0]71.3[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]48-7S Spark[/TD]

[TD]5.23[/TD]

[TD=width: 86, bgcolor: #7030A0]71.1[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-T30 Reliant[/TD]

[TD]62.4[/TD]

[TD=width: 86, bgcolor: #7030A0]71[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]Mk-55 Thud[/TD]

[TD]34.7[/TD]

[TD=width: 86, bgcolor: #7030A0]70.7[/TD]

[/TR]

[/TABLE]

"Good" engines

[TABLE]

[TR]

[TD]LV-1 Ant[/TD]

[TD]0.576[/TD]

[TD]70.4[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]24-77 Twitch[/TD]

[TD]4.59[/TD]

[TD]70.3[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]LV-1R Spider[/TD]

[TD]0.561[/TD]

[TD]68.6[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]CR-7 RAPIER[/TD]

[TD]38.9[/TD]

[TD]68[/TD]

[/TR]

[/TABLE]

"Fair" engines

Scenario #7: Tylo Ascent or descent stage

2400 m/sec DV, vacuum Isp, 1.5G acceleration on Tylo

"Excellent" engines

[TABLE]

[TR]

[TD]KR-2L Rhino[/TD]

[TD]63.6[/TD]

[TD]37.5[/TD]

[/TR]

[/TABLE]

"Good" engines

[TABLE]

[TR]

[TD]RE-L10 Poodle[/TD]

[TD]7.48[/TD]

[TD]35.3[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]KS-25x4 Mammoth[/TD]

[TD]119[/TD]

[TD]35.2[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-I5 Skipper[/TD]

[TD]19.2[/TD]

[TD]34.9[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]RE-M3 Mainsail[/TD]

[TD]43.8[/TD]

[TD]34.4[/TD]

[/TR]

[/TABLE]

"Fair" engines

[TABLE]

[TR]

[TD]LV-909 Terrier[/TD]

[TD]1.69[/TD]

[TD]33.2[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]T-1 Aerospike[/TD]

[TD]5.01[/TD]

[TD]32.8[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]KR-1x2 Twin-Boar[/TD]

[TD]54.6[/TD]

[TD]32.2[/TD]

[/TR]

[/TABLE]

[TABLE]

[TR]

[TD]24-77 Twitch[/TD]

[TD]0.412[/TD]

[TD]30.3[/TD]

[/TR]

[/TABLE]

Best,

-Slashy

Edited May 7, 2015 by GoSlash27

[Moss]

The KSP Wiki seems somewhat neglected at present.

[Argon]

Thanks, this is useful since the neutering of the nuke.

-Argon

[LiquidFenrir]

The KSP Wiki seems somewhat neglected at present.

Well, a lot of things have changed with 1.0 so it might take a lot of time to update everything

[UmbralRaptor]

The LV-1 Ant seems like it would excel for a minimal mass (0.005 mass instruments, FL-T100 or 200) probe. TWR is unimpressive, but mass and cost might be competitive with an ion-based craft for some roles. No idea what to do with the LV-1R, though.

[Starwhip]

UR,

The LV-1R looks decent for a lander engine. For a munar ascent, a pair of them will lift a tonne of payload with a total stage mass of 1.5 tonnes.

The LV-1, OTOH... same mass as the LV-1R and 1/3 the thrust. Or even worse compared to the 48-7S. You've got to have a need for a seriously tiny engine before the ant becomes an attractive option.

Best,

-Slashy

Gilly would like to have a word with you...

[Francesco]

I don't get why the 909 gets outclassed, given its high ISP... it just needs a first stage to push it past 10km or so... at least that's how I've used it for a small Kerbin orbiter, in career mode.

What am I missing?

[NathanKell]

Slashy, what do you mean 1/3 the thrust? They have the same thrust in vacuum. I think you got confused by looking at the SL thrust numbers.

I also question your reliance on high-TWR metrics. Of course vacuum engines optimized for long burn times (but high Isps) are gonna suck by that metric. A 909 will beat the pants off a T30 as an upper stage, though.

[Radam]

In comparison, the LV-N at 0.5 twr and 6000ms dv, has a payload fraction of about 16%

[impyre]

I take offense, good sir, at this affront upon my favorite little engine that could (almost). I agree with NathanKell, your results are so one-sided because your methods (probably unintentionally) bias the results in favor of those engines.

[Commissioner Tadpole]

I don't know why, but I find Skippers to be awful in vaccuum and Poodles to still be much better.

[Francesco]

Something that could be useful - also, this is in defense of the 909

This small chart tells you the altitude needed on Kerbin to get a certain amount of delta-v out of a 909-powered rocket stage. (Roughly.)

It associates an altitude with a multiplier, which goes from ~0.246 (sea level delta-v, 85s/345s) to 1.0 (max vacuum delta-v).

Done with Kerbal Engineer.

multiplier    km

0.246         0.0
0.4           2.0
0.5           3.3
0.6           4.6
0.7           6.2
0.8           8.3
0.85          9.7
0.9           11.7
0.95          15.2
1.0           ~50

So at ~12km, a 909 runs at 90% of its max ISP.

[impyre]

I would like to submit, for your perusal, my own analyses as a response to the OP.

First I'll explain what I did and why I did it that way.

To the rocket equation! dV=ISP*g*ln(massfull/massempty) this equation shows that ISP and dV are linearly proportional given a constant mass ratio. That's an easy enough relationship to grasp... ceteris paribus (all else being constant)... twice the ISP = twice the dV. If we solve for mass ratio we get e^(dV/(isp*g))=mass ratio. This tells us that for decent values of ISP, changing mass ratio by even a tiny amount can result in increasing dV by an order of magnitude. When you put these two things together, you have a basic understanding of the tradeoff between ISP and mass. For example, if engine A has an ISP of 400 and a mass of 1, and engine B has an ISP of 800 and a mass of 1.2 then you'll likely get better results (at least in terms of dV) with engine B. It is only slightly heavier but gives twice the ISP. However, if engine A weighed only .5, you'd have to look more closely to determine which would be better for your particular vessel. IE: shaving .1t off a 30 ton monster won't do much to change dV... but increasing isp by even 50% will make a big difference. On the other hand, sticking a nuke on a 2t lander won't do you any favors since it's mass will easily more than make up any savings you get from higher ISP... it weighs more than the craft itself! Using a 48-7s or an lv-909 would probably be better choices.

tl;dr Here's the graphs

The red is my pure dV potential function, essentially ISP(vac) per mass(tons). The best ratio will provide better ISP for any given payload, though the size of the payload will change the relative differences between the engines. Some payloads may notice little difference from one engine to the next, while others may notice significant differences.

We all know that putting a microscopic engine on a giant fuel tank would create large dV... but is it really useful? I think not... most of us care about thrust (some more than others), and thrust is needed to execute maneuvers in a clean fashion. Thus the blue line represents my arbitrary "utility" function. It's ISP*(thrust/mass). Thus, it accounts for dV potential (to some extent) and for thrust capability (again, to some extent).

The interesting thing to note is the relationship between the raw dV potential and the "utility" function. There seems to be an almost "inverse" relationship between the two... higher dV = harder to use... but this graph illustrates two interesting points. Firstly, the lv-909 represents a unique compromise between dV and utility... as such it is uniquely suited to smaller-scale repetitive ops (like landers that go to the surface to get science and then make orbit to refuel... rinse/repeat); the second interesting point is that the 48-7s (one of my personal favorites, and the one I use on most of my single-seater landers) has the highest combined dV potential *and* utility function of any other engine... this little guy is pure workhorse (if limited in the types of vessels). All of these use an arbitrary coefficient of 1 on all terms, so it's not fair to suggest that it's "accurate" or represents anything other than a "notion" of usefulness. This is just a good representation of the "usefulness" of various engines in *my* opinion.

Also, I'd love some feedback on whether you agree or disagree (and whether you find it useful or accurate or not).

EDIT: Correcting some verbiage. I keep saying "weight" where I *should* be saying "mass".

- - - Updated - - -

Also, it's worth noting that I didn't raise any mass terms to exponentials (though that might be more accurate)... So this representation actually favors ISP more heavily than it should. (And is more forgiving of "heavy" engines)

Edited May 3, 2015 by impyre

[Alshain]

There really should be absolutely 0 engines in this game that do not stack (not counting radials). Even Jet engines should stack! There is absolutely no reason to not allow an engine to stack.

Anyway, rant asside. This looks like a pretty good list, thanks for doing this. I disagree on the assessment of the LV-909. It makes a great upper stage.

Edited May 3, 2015 by Alshain

[GoSlash27]

Slashy, what do you mean 1/3 the thrust? They have the same thrust in vacuum. I think you got confused by looking at the SL thrust numbers.

I also question your reliance on high-TWR metrics. Of course vacuum engines optimized for long burn times (but high Isps) are gonna suck by that metric. A 909 will beat the pants off a T30 as an upper stage, though.

NathanKell,

After seeing this, I went back and re-checked the config files.

They are indeed the same, and that calls into question this entire analysis. I don't know if they got re-juggled during 1.01/ 1.02 or if I just plain read them wrong, but they are not correct, so my results are invalid.

Gotta start over...

thankful and grumbling all at the same time...

-Slashy

- - - Updated - - -

I would like to submit, for your perusal, my own analyses as a response to the OP.

First I'll explain what I did and why I did it that way.

To the rocket equation! dV=ISP*g*ln(massfull/massempty) this equation shows that ISP and dV are linearly proportional given a constant mass ratio. That's an easy enough relationship to grasp... ceteris paribus (all else being constant)... twice the ISP = twice the dV. If we solve for mass ratio we get e^(dV/(isp*g))=mass ratio. This tells us that for decent values of ISP, changing mass ratio by even a tiny amount can result in increasing dV by an order of magnitude. When you put these two things together, you have a basic understanding of the tradeoff between ISP and mass. For example, if engine A has an ISP of 400 and a mass of 1, and engine B has an ISP of 800 and a mass of 1.2 then you'll likely get better results (at least in terms of dV) with engine B. It is only slightly heavier but gives twice the ISP. However, if engine A weighed only .5, you'd have to look more closely to determine which would be better for your particular vessel. IE: shaving .1t off a 30 ton monster won't do much to change dV... but increasing isp by even 50% will make a big difference. On the other hand, sticking a nuke on a 2t lander won't do you any favors since it's mass will easily more than make up any savings you get from higher ISP... it weighs more than the craft itself! Using a 48-7s or an lv-909 would probably be better choices.

tl;dr Here's the graphs

The red is my pure dV potential function, essentially ISP(vac) per mass(tons). The best ratio will provide better ISP for any given payload, though the size of the payload will change the relative differences between the engines. Some payloads may notice little difference from one engine to the next, while others may notice significant differences.

We all know that putting a microscopic engine on a giant fuel tank would create large dV... but is it really useful? I think not... most of us care about thrust (some more than others), and thrust is needed to execute maneuvers in a clean fashion. Thus the blue line represents my arbitrary "utility" function. It's ISP*(thrust/mass). Thus, it accounts for dV potential (to some extent) and for thrust capability (again, to some extent).

http://i.imgur.com/bQroQ7M.png

The interesting thing to note is the relationship between the raw dV potential and the "utility" function. There seems to be an almost "inverse" relationship between the two... higher dV = harder to use... but this graph illustrates two interesting points. Firstly, the lv-909 represents a unique compromise between dV and utility... as such it is uniquely suited to smaller-scale repetitive ops (like landers that go to the surface to get science and then make orbit to refuel... rinse/repeat); the second interesting point is that the 48-7s (one of my personal favorites, and the one I use on most of my single-seater landers) has the highest combined dV potential *and* utility function of any other engine... this little guy is pure workhorse (if limited in the types of vessels). All of these use an arbitrary coefficient of 1 on all terms, so it's not fair to suggest that it's "accurate" or represents anything other than a "notion" of usefulness. This is just a good representation of the "usefulness" of various engines in *my* opinion.

Also, I'd love some feedback on whether you agree or disagree (and whether you find it useful or accurate or not).

EDIT: Correcting some verbiage. I keep saying "weight" where I *should* be saying "mass".

- - - Updated - - -

Also, it's worth noting that I didn't raise any mass terms to exponentials (though that might be more accurate)... So this representation actually favors ISP more heavily than it should. (And is more forgiving of "heavy" engines)

Impyre,

My analysis is purely mathematical. Basically, it's taking the rocket equation and flipping it backwards. Given a desired acceleration and DV budget, you can derive from the available specs which engines can push the most payload and still meet the goal. It's counter- intuitive, but very often the most efficient engine doesn't come out on top.

Say you have two engines, one heavier and more efficient, the other lighter and less efficient. Both produce the same thrust.

Which one will deliver the most DV for the payload will depend on how much DV is required and how much acceleration you need. It's often not the more efficient engine that does this, but the lighter engine. The difference in engine mass is replaced with additional fuel and tankage and thus the "inefficient" engine winds up getting more done with less total mass.

Having said that, my initial numbers are all suspect, so I'm going to have to redo all of this.

Best,

-Slashy

[impyre]

NathanKell,

Impyre,

My analysis is purely mathematical. Basically, it's taking the rocket equation and flipping it backwards. Given a desired acceleration and DV budget, you can derive from the available specs which engines can push the most payload and still meet the goal. It's counter- intuitive, but very often the most efficient engine doesn't come out on top.

Say you have two engines, one heavier and more efficient, the other lighter and less efficient. Both produce the same thrust.

Which one will deliver the most DV for the payload will depend on how much DV is required and how much acceleration you need. It's often not the more efficient engine that does this, but the lighter engine. The difference in engine mass is replaced with additional fuel and tankage and thus the "inefficient" engine winds up getting more done with less total mass.

Having said that, my initial numbers are all suspect, so I'm going to have to redo all of this.

Best,

-Slashy

I agree with you up to a point; however, given your speculative statement regarding two engines, one lighter and less efficient and the other heavier and more efficient, the one which delivers the most dV for the payload (or for itself for that matter) does *not* depend on dV *or* acceleration. That statement is self-contradictory. "The one that delivers the most dV depends on how much dV and acceleration you need." is the essence of the statement. This is contradictory and circular... further, there is no part of the rocket equation that makes any consideration for thrust or acceleration (though we have to consider those things when *spending* our dV to get the most out of it... which is where my "utility" function comes into play). Given these two engines... which have given ISP's and masses, the engine that will give the most possible dV for a given fuel/payload configuration depends solely on the fraction of mass that said engine contributes to the overall mass. Larger payload/fuel configurations will likely benefit more from ISP than weight savings while smaller payload/fuel configurations will likely benefit more from reduced weight than ISP. Of course this is all relative to the disparity in ISP and mass between the two engines... and it is this disparity which will set the dividing line along the equipotential surface that separates the region where one engine dominates from the region where the other dominates. Of course this also depends on how you define "efficient", I assume for this discussion when referencing an engine's "efficiency" you are in fact talking about ISP.

http://forum.kerbalspaceprogram.com/threads/113476-ISP-Weight-tradeoff-on-engines-(graph)

Above is a post I made previously for any arbitrary combination of dV, ISP, and engine mass (as percentage of total vessel mass). The upper graph shows the relationship for a given payload mass, while the lower graph is animated to show the effect of reducing the payload mass (and displays engine mass in units instead of percentage)... this shows how the upper graph changes shape as payload size changes. As you can see, as payload size becomes smaller... engine weight becomes more important. I considered placing several of KSP's engines on this graph to show how they relate to each other, but it's redundant information if you understand the basic relationships.

EDIT: also, I'm not trying to hijack your thread. I promise

Edited May 3, 2015 by impyre

[GoSlash27]

I agree with you up to a point; however, given your speculative statement regarding two engines, one lighter and less efficient and the other heavier and more efficient, the one which delivers the most dV for the payload (or for itself for that matter) does *not* depend on dV *or* acceleration. That statement is self-contradictory.

Impyre,

When I say "efficient" in this case I'm referring to specific impulse.

And yes, they both do definitely matter, since Isp and ln(Rwd) both factor into the Dv.

Ultimately, the thrust does as well, since that sets the maximum mass of the stage for a desired acceleration.

Unfortunately, I've got to go back and review all of my initial values since I know that at least some of them are incorrect.

*edit*

Yeah, they readjusted the thrust values sometime after 1.00. I went back and compared my stored copies of the .cfg files from 1.00 to the current 1.02. They changed them and didn't mention it.

Almost all engines were buffed about 15% from the previous thrust values. The thrust values for the LV-1, LV-909, and Poodle were quadrupled (which explains why they were so outclassed before). Thrust was reduced slightly for the RAPIER in closed cycle and for some reason the thrust was doubled for the Rhino.

I'm not sure why they would do this, since it was already pretty dominant. My quick back-of-the-envelope shows it capable of achieving 5.6% payload fraction in the mass-mover role and an almost-insane 60% payload fraction in the 3 stage upper stage role.

Back to the drawing board...

-Slashy

Edited May 3, 2015 by GoSlash27

[NathanKell]

Thrust values did not change at all. You're just not accounting for the difference in thrust between sea level and vacuum.

Except the aerospike (which got 5 more seconds vac Isp), the only thing that changed 1.0 to 1.0.1 was that maxThrust is now given in terms of vacuum thrust instead of sea level thrust. If you actually look at the ratings ingame, where it states vacuum and sea level thrust, you'll see that nothing changed. The only change was behind the scenes.

Since you think the thrusts did change, I'm guessing you did your original calculations based off the cfg, and thus missed changes in thrust as air pressure changes. I would beg you to redo your calculations, this time keeping in mind that Isp at a given pressure matters for thrust.

[GoSlash27]

Thrust values did not change at all. You're just not accounting for the difference in thrust between sea level and vacuum.

Except the aerospike (which got 5 more seconds vac Isp), the only thing that changed 1.0 to 1.0.1 was that maxThrust is now given in terms of vacuum thrust instead of sea level thrust. If you actually look at the ratings ingame, where it states vacuum and sea level thrust, you'll see that nothing changed. The only change was behind the scenes.

Since you think the thrusts did change, I'm guessing you did your original calculations based off the cfg, and thus missed changes in thrust as air pressure changes. I would beg you to redo your calculations, this time keeping in mind that Isp at a given pressure matters for thrust.

^ Close, but not quite. They were recorded in the .cfg files as "max thrust" in both cases. If they were being employed differently, it wasn't self- evident.

Example:

LV-909 cfg file from 1.00

MODULE
    {
        name = ModuleEngines
        thrustVectorTransformName = thrustTransform
        exhaustDamage = True
        ignitionThreshold = 0.1
        minThrust = 0
       [B] maxThrust = 14.7826087[/B]
        heatProduction = 20
        fxOffset = 0, 0, 0.21
        EngineType = LiquidFuel
        PROPELLANT
        {
            name = LiquidFuel
            ratio = 0.9
            DrawGauge = True
        }

LV-909 cfg file from 1.02

MODULE
    {
        name = ModuleEngines
        thrustVectorTransformName = thrustTransform
        exhaustDamage = True
        ignitionThreshold = 0.1
        minThrust = 0
        [B]maxThrust = 60[/B]
        heatProduction = 80
        fxOffset = 0, 0, 0.21
        EngineType = LiquidFuel
        PROPELLANT
        {
            name = LiquidFuel
            ratio = 0.9
            DrawGauge = True
        }

If you happen to have a copy of the 1.00 cfg files, you can verify it yourself.

Best,

-Slashy

[NathanKell]

I'm well aware of that. What I am saying however is that you appear to have missed one of the big changes in KSP 1.0--namely, that thrust changes with Isp, not fuel flow--and thus if you were going by the cfg files alone in either case you are missing significant details. Which you would have seen, had you examined the parts ingame, where the editor info window clearly shows both the sea level thrust and the vacuum thrust (neither of which changed going 1.0 to 1.0.x).

You need to consider the engine's performance throughout its used range when determining delta V--that is, for lifters, probably getting weighted-average Isp from sea level though about 0.05atm tops, and for vacuum engines only vacuum isp. If TWR figures into your usefulness calculations, then you need to calculate TWRs the same way (weighted average of sea level and vacuum thrust).

That said, given those results I'm really glad the change in 1.0.1 was made, to make maxThrust again the true max thrust, not the sea level thrust, because it will hopefully reduce similar confusion in future.

[GoSlash27]

I'm well aware of that. What I am saying however is that you appear to have missed one of the big changes in KSP 1.0--namely, that thrust changes with Isp, not fuel flow--and thus if you were going by the cfg files alone in either case you are missing significant details.

Likewise, I'm well aware of *that*. I think most of us were. We were aware of it because it was in the dev notes.

What I was *not* aware of was the "BTW, the 'max thrust' values are no longer max thrust" part. Specifically, because it was not mentioned.

And then, of course, the little part about "OHAI, max thrust is max thrust again. BTW, we changed them all and some of the Isps". This is the kind of stuff you're supposed to point out when you do patches. Just sayin'...

That said, given those results I'm really glad the change in 1.0.1 was made, to make maxThrust again the true max thrust, not the sea level thrust, because it will hopefully reduce similar confusion in future.

I'm also hopeful that this is the case. It's not cricket to change how config files are interpreted by the program without mentioning it in the dev notes.

It's also not cricket to change the config files themselves without mentioning it in the dev notes. A couple things here and there are understandable. The core specs for every rocket engine in a *rocket building game*... not so much.

Not blaming you, mind you. Unless this shoddy recordkeeping happens to be your fault.. in which case, I am.

Edited May 4, 2015 by GoSlash27

[GoSlash27]

Finally!

New corrected data is up!

Also added a new scenario today for Tylo ascent/ descent stages.

Best,

-Slashy

[Maukse]

So LV-n is totaly worthless now ?

[GoSlash27]

So LV-n is totaly worthless now ?

Not at all.

It's just not a LF&O engine, so it's not included here.

Best,

-Slashy

[gmiddlemass]

Thanks Slashy, very informative... will have a proper look at it later.