Isp vs Thrust?

[Dangerous_Beans]

Hi

Something that has been confusing me; Isp independent of thrust (unless i missed something), so why do the more powerful engines in KSP have lower Isps? (or is it just a game balance thing )

Shouldn't you be able to build more powerful versions of a rocket engine, while keeping the Isp the same, by building a larger engine that uses more fuel but has the same exhaust velocity?

Or is it just an engineering issue; for a given Isp it's easier to build a 1kN engine than a 1MN engine, and conversely for a given thrust, it's easier to build a less efficient engine? AKA, add more boosters .

I just have the sneaking suspicion that I have missed something here.

Thanks

[Diche Bach]

Isp means specific impulse, basically the efficiency of an engine. Thrust is the force the engine produces.

Pretty sure, the more powerful engines tend to have lower Isp because they achieve greater thrust (assuming the same chemical fuel) largely by burning the fuel more rapidly compared to less powerful engines. There are probably clever mechanical ways to maximize efficiency (e.g., by shaping the exhaust or something like that) but for any given fuel there are inherent limitations on how efficient you can make an engine.

At least that is my understanding at this point! Will be good to see if the local physicists and engineers give me a passing grade

Edited August 22, 2013 by Diche Bach

[Bunsen]

Or is it just an engineering issue; for a given Isp it's easier to build a 1kN engine than a 1MN engine, and conversely for a given thrust, it's easier to build a less efficient engine? AKA, add more boosters .

That's the most realistic aspect of it. There are a lot of complicating factors, and there are problems that come in at the bottom end where making engines really small also makes them inefficient, but the real-world tradeoffs between different fuels, different methods of driving turbopumps, whether you use pumps at all, what pressure your engine bell needs to operate in, etc. are a bit much for KSP's approach to engineering. So they distill it down to "efficiency is approximately proportional to inconvenience" and call it good.

[Navyguy]

I've kinda noticed and this still bugs me. The fact how fast fuel goes. When you have the biggest fuel tank it still runs out of fuel in about a minute if you use the mainsail. That has to do with the isp being alot lower than real engines isp are. Thats a comparison on isp of KSP and the real world. So don't be frustrated. Just add boosters.

[Anton P. Nym]

It's all about energy. Chemical fuel can only generate a certain amount of energy from the same mass of reactants. Cloning a quick and over-simplified summary I posted elsewhere:

We use rockets to add momentum to whatever it is we want to put into space. Momentum is conserved; so to add momentum to a spaceship we have to add an equal amount of momentum to something in the opposite direction. In a rocket, that's the exhaust. Momentum is the mathematical product of mass and velocity, so under ideal circumstances increasing the amount of momentum transferred to our spacecraft can be done by increasing either the mass or velocity of the exhaust with equal effect.

Rockets, however, aren't working under ideal conditions as they have to carry the fuel needed to make exhaust and fuel has its own mass; it's very easy to design a rocket with so much fuel that the added weight makes it unable to fly. Rocket engineers try to increase the speed of the exhaust whenever possible, by making the exhaust hotter, as that reduces the weight of fuel needed to transfer the needed momentum. (This efficiency of momentum transfer is measured by "specific impulse", abbreviated I(sp). Bigger I(sp) numbers mean greater efficiency.) That's difficult because there are limits imposed by chemistry on how hot a given mass of a particular fuel can burn.

To get around the limits of fuel chemistry we can use other means than chemical reactions (burning) to heat the fuel. Ion engines use electricity to heat xenon gas to absurdly-high temperatures, getting I(sp) figures ten times that of our best rockets, but limits on the lightweight power sources we can use in space mean it can only heat tiny amounts of fuel at a time... so the thrust is very low. In order to do better, we need much greater sources of power to get large quantities of thrust out of high-efficiency engines. We have a source of power that's much greater than burning fuel, but as always in rocket science there's a catch.

-- Steve

[Anton P. Nym]

When you have the biggest fuel tank it still runs out of fuel in about a minute if you use the mainsail. That has to do with the isp being alot lower than real engines isp are.

Not really. The sea level ISP of the Mainsail is 280s; the Rocketdyne F-1s that powered the Apollo missions had sea level ISPs of 263. It's that our fuel tanks are very much smaller than the real world ones, scaled down to match the Kerbals.

-- Steve

[SargeRho]

Well, the SSME has an Isp of 450, and it's one of the most powerful rocket engines in existence.

[Anton P. Nym]

Well, the SSME has an Isp of 450, and it's one of the most powerful rocket engines in existence.

It has an Isp of 450 in vacuum. At sea level it has an Isp of ~360.

-- Steve

[Scotius]

To get out of gravity well and atmosphere constraints fast you need power. To get far in space you need efficiency. Power we do have - aside of fancy engineering tricks like the aerospike engines, fuel crossfeeding etc. there is not much to gain in the area of chemical engines. Currently we are working towards efficiency. Electrical (ion and plasma), nuclear and fusion engines, along with more 'esoteric' methods of propulsion like warp drive and quantum thruster are worked on around the world. And all of them are characterised by Isp measured in thousands of seconds, not mere hundreds as chemical engines.

[magnemoe]

Not really. The sea level ISP of the Mainsail is 280s; the Rocketdyne F-1s that powered the Apollo missions had sea level ISPs of 263. It's that our fuel tanks are very much smaller than the real world ones, scaled down to match the Kerbals.

-- Steve

TWR is far lower than real world. Mainsails has 25, real world TWR for large engines are 75-150. Tanks also have an higher dry mass compared to fuel content.

Yes this is for balance its easy to get something small into orbit far harder with something heavy, but also the main reason asparagus works so well.

[Dangerous_Beans]

thanks, that all makes more sense now.

[edogrigqv2]

TWR is far lower than real world. Mainsails has 25, real world TWR for large engines are 75-150. Tanks also have an higher dry mass compared to fuel content.

Yes this is for balance its easy to get something small into orbit far harder with something heavy, but also the main reason asparagus works so well.

Lsp is about the efficiency of engine(burning fuel). In space high lsp is more important than high trust, as from physics high lsp => high velocity, no matter how much time you need to get that velocity.

For example: if your lsp is very high and trust is very low, you will get speed slowly, but fuel will be enough for long time and if your trust is high and lsp is low you will consume your fuel quickly but the total velocity will be lower as it burns rapidly. At surface of planets as long you haven't a chieved elliptic/circular orbit, the gravity will drag you down, especially at the launch. Here trust is MORE significant, because you need to get height quickly otherwise the gravity's affect on you will be higher.

So your first stages should have high trust and "space" engines need only high lsp, while orbiting you always will have enough time to do maneuvers

[Dkmdlb]

I've kinda noticed and this still bugs me. The fact how fast fuel goes. When you have the biggest fuel tank it still runs out of fuel in about a minute if you use the mainsail.

That's actually not too far off from reality. Take a look at this video of a kerosene tank emptying on the Saturn 1. The first stage emptied in about 2.5 minutes.

[Streetwind]

Well, the SSME has an Isp of 450, and it's one of the most powerful rocket engines in existence.

Do note that, while the Rocketdyne RS-25 (also known as SSME) is one of the most efficient engines ever flown (but not THE most efficient), it is far from the most powerful. 1.86 MN sea level thrust is decently beefy, I'll admit - that's about three Merlin 1D worth, or about half an RD-180. But the Space Transport System was a weighty fellow indeed, and needed extreme amounts of thrust to heave its orbiter off the pad. Each RS-25 mounted on the shuttle was responsible for a grand total of... wait for it... less than 6% of liftoff thrust. Despite carrying three of them, the orbiter itself was responsible for less than a fifth of the STS' total power output.

So where did the remaining 83% of thrust come from? From the very thing that deserves the real title of "most powerful rocket engine in existance": the space shuttle solid boosters. Each one produces a whopping 12.5 MN at sea level, and there were two of them on each STS launch. Nothing more powerful than these, and by a fairly comfortable margin too, has ever been flown. The Saturn V had more liftoff thrust than the STS, but it achieved it by using five large engines, not two with a bit of assistance from assorted small fry.

And limiting yourself to liquid engines: the most powerful ever flown was said Saturn V's Rocketdyne F-1 at 6.77 MN sea level thrust, and the most powerful ever developed was the NPO Energomash RD-170 at 7.55 MN sea level thrust (but to my knowledge it never flew). The currently in-development SpaceX Raptor engine is (according to latest SpaceX statements) going to fall short of beating the RD-170 at liftoff thrust, but it is expected to beat it in vacuum thrust (8.2 MN vs 7.89 MN), and with a decently higher Isp in both regimes to boot. Should the Raptor finish development in this form, it would become the most powerful liquid fuel engine ever developed, and soon after perhaps, also the most powerful one flown.

[Kryten]

And limiting yourself to liquid engines: the most powerful ever flown was said Saturn V's Rocketdyne F-1 at 6.77 MN sea level thrust, and the most powerful ever developed was the NPO Energomash RD-170 at 7.55 MN sea level thrust (but to my knowledge it never flew).

RD-170 is the engine on Zenit's first stage, it's flown dozens of times.

[Streetwind]

It did? Well, today I learned something new!

EDIT: and there's the source of my mistake, I confused it with the RD-270. Just one number off!

[magnemoe]

That's the most realistic aspect of it. There are a lot of complicating factors, and there are problems that come in at the bottom end where making engines really small also makes them inefficient, but the real-world tradeoffs between different fuels, different methods of driving turbopumps, whether you use pumps at all, what pressure your engine bell needs to operate in, etc. are a bit much for KSP's approach to engineering. So they distill it down to "efficiency is approximately proportional to inconvenience" and call it good.

Engine size versus isp is present in KSP at least before the arm parts, The very small engines have poor isp who goes up in the center and fall off with skipper and mainsail. Engines is KSP has however lower TWR than in real world.

Generally efficient increases with size however large efficient rocket engine will have some issues, they would be expensive and have more stuff who can go wrong, on the other hand tanks and fuel is cheap.

[shynung]

Or is it just an engineering issue; for a given Isp it's easier to build a 1kN engine than a 1MN engine, and conversely for a given thrust, it's easier to build a less efficient engine? AKA, add more boosters .s

That's about half. The other half has to do with propellant chosen for the rocket. I'll explain the engineering aspect.

For simplicity, let us assume a monopropellant engine, whose primary components are composed of a propellant tank, an ignition valve, a reaction/combustion chamber, and a nozzle. Let's also assume a single type of monopropellant, to ignore the chemical side for now.

The simplest type of liquid fuel rockets are pressure-fed rockets. This has no turbopump, so the pressure needed to transfer propellant to the chamber are either supplied by a separate pressurant gas, or provided by the propellant's own vapor pressure, using a heating element in the tank. This design is often used in engines designed to operate fully in vacuum, including the Apollo SPS.

While simple and efficient, this system has a limited maximum thrust, because the pressure in the chamber goes up in proportion to the thrust. At some point, the pressure generated by the pressurant gas or propellant vapor pressure will fail to exceed the chamber's, resulting in no more thrust being produced by the engine.

At this point, the solution is simple: install a pump. Turbopumps are chosen for their high power-to-weight ratio (the amount of work it does, compared to its mass), to keep the engine's overall TWR high.

But at the same time, the turbopump brings its own problems, the most notable of which is that it requires power to operate. Some designs simply route some of the gases from the chamber into the turbopump, some others install a separate chamber for the turbopump.

The next question regarding the final performance of the engine is: where does the turbopump exhaust go?

One solution used often is throwing it overboard from an exhaust pipe. This gives the maximum thrust available, as the turbopump runs efficiently on its own, delivering the maximum propellant flow rate into the chamber. However, since the propellant used to run the turbopump is thrown out, it is no longer available to produce thrust, resulting in lower specific impulse.

This design, called a gas-generator cycle, is used frequently in engines used for lower-stage work, one notable example being the Saturn V's F-1 engines.

The other solution is piping the turbopump's exhaust into the main chamber of the engine. This is to reduce the loss of efficiency associated with dumping propellant off to the side, resulting in a more efficient engine compared to the gas-generator.

However, since the turbopump exhausts into the main chamber, it has to deal with back-pressure from the reaction in the main chamber itself, needing a separate chamber with a higher pressure. In addition to lowering the turbopump's performance (leading to lower mass flow, therefore lower thrust), it also exposes the turbine blades to much harsher working conditions than a gas-generator engine's, requiring exotic and/or advanced materials. This makes the final engine more complicated and expensive, but gives more specific impulse than a gas generator, at the expense of lower TWR.

This design, the staged-combustion cycle, is used for upper-stage engines or in engines that are never staged away even if other engines are, a notable example of which is the Space Shuttle Orbiter's RS-25/SSMEs.

Note, there are other engines that work on slightly different cycles that I may have missed (one cycle simply heats the incoming propellant via regenerative cooling of the chamber and nozzle, rather than reacting it, and using the expansion of said heated propellant to drive the turbopump), but the most common chemical rocket engines in service today typically use one of these three cycles.

And that's a little bit of rocket science from the engineer's side. For the chemical side, read this book to explore the lair of the beyond-insane-yet-brilliantly-intellectual crackpots, known to the industry as 'propellant chemists'.

EDIT: ...aaand all of a sudden I just made a wall of text.

Edited November 6, 2014 by shynung

[sgt_flyer]

Be careful there's a lot of staged cycle variants RS-25 cycle being one of them

Basically RS-25 engines use two turbines + turbopumps - one oxygen rich, the other fuel rich. (Basically, they add a bit of what's missing on each side (either fuel or oxydizer) to drive the pumps (one pump per turbine), the rest of the fuel (or oxydizer) is mixed with the turbine's exhaust, and all of it is fed with the other flow, and both of the flows mixed and ignited in the main combustion chamber. Extremely complex, pretty heavy, but veeery efficient so yes, for those engines, dropping them too early would be not interesting.

Russians make use of a different kind of staged combustion, oxygen rich, where they have only 1 turbine, which drives all the turbopumps at the same time. They put all oxydizer in the turbine, and just enough fuel to drive the turbines. The rest of the fuel is mixed and ignited with the turbine's exhaust in the combustion chamber. Much simpler, but requires more advanced materials to handle a high pressure oxydizer rich environment.

A lot of russian engines use this technique, and all are used on rocket first stages. N-1's NK-33s (refurbished on aj26-58), Proton RD-253, Zenit's RD-170, Atlas V RD-180 and Angara's RD-191. Of course, all those engines are not meant to be reusable, but their simplicity and efficiency allows to limit the costs.

One of the key features of staged cycles if effectively the high chamber pressures, giving off better ISP's than the gas generator counterparts. But that requires better metallurgy techniques to handle those pressures. (Which can make the engines abit heavier)

US rocket scientists even thought that oxygen rich staged cycle was not possible to do before 1990! (When russians showed them their engines)

Edited November 6, 2014 by sgt_flyer

[Kulebron]

IIRC, NK-33 (or RD-170) has TWR of well above 100, which means it's not heavy (or it's crazy powerful).

I wonder, back in 1970 this kind of engine was the most efficient in terms of Isp. Now there are liquid fuel engines with similar Isp, how did they reach this efficiency? How using gas as a propellant improve efficiency?

[Streetwind]

I wonder, back in 1970 this kind of engine was the most efficient in terms of Isp. Now there are liquid fuel engines with similar Isp, how did they reach this efficiency? How using gas as a propellant improve efficiency?

They probably do it by doing the same things as the NK-33 did.

Isp is dependant on the exhaust velocity of the rocket engine, and that in turn is dependant largely on three factors: the molecular weight of the exhaust, the exhaust temperature, and the combustion chamber pressure. The former is based on what kind of chemical reaction happens to the propellant you use, the latter is dependant on the design of your engine, and the middle one is partly dependant on both.

If you have a given propellant, then you're getting different Isp numbers for different engine construction schemes, but similar Isp numbers for similar engine construction schemes. As a result, it matters comparatively little whether an engine was built in 1970 or 2010, despite fourty years of time in between - a gas generator cycle that throws part of its propellant overboard is not going to beat a staged combustion cycle that utilizes it all.

Case in point:

Kuznetsov NK-33 (1969), propellant: RP-1/LOX, Isp: 297s (ASL)/331s (Vac), closed cycle (staged combustion)

SpaceX Merlin 1D (2012), propellant: RP-1/LOX, Isp: 282s (ASL)/311s (Vac), open cycle (gas generator)

Kind of a sobering comparison for one of humanity's most recently completed and also fan favorite new rocket engines, no?

It's commonly accepted that today's rocket engine technology was maxed out sometime in the 1970's. There's been little to no improvements in the known combustion cycles since then. There are engines like the Merlin 1D, which scored a new world record in TWR; but you have to keep in perspective what this means. Basically, new materials and improved manufacturing allowed the shaving off of a couple kg's of dry mass from the engine compared to older designs. That's it. In the actual rocket engine itself, little has changed compared to half a century ago. The Merlin turbopumps are even derived from those on the Apollo Lunar Module's descent stage engine. Once an ideal design has been found, it's hard to come up with something ideal-er than that.

What we can still do is try new propellants, like liquid methane, which tended to get passed over in the past. Or we can try a combustion cycle that hasn't been done before, like the full-flow staged combustion variant that was briefly examined with the RD-270 but not followed up on. As a matter of fact, the Raptor will do both at once, which is why it might actually set itself apart from existing engines by a solid margin, for the first time in decades. If, of course, it actually works out.

Edited November 6, 2014 by Streetwind

[shynung]

Russians make use of a different kind of staged combustion, oxygen rich, where they have only 1 turbine, which drives all the turbopumps at the same time. They put all oxydizer in the turbine, and just enough fuel to drive the turbines. The rest of the fuel is mixed and ignited with the turbine's exhaust in the combustion chamber. Much simpler, but requires more advanced materials to handle a high pressure oxydizer rich environment.

That's because Russian engines typically use kerosene/LOX mix as propellant. In terms of molecular weight, oxygen is much lighter than kerosene (having somewhere between 6-16 carbons per molecule). In a rocket engine, for a given chamber temperature and pressure, the exhaust velocity is higher the lighter the exhaust gas molecules are. For the Russians, this means cramming as much oxygen into the chamber as possible. As a result, some of the oxygen simply becomes hot without actually reacting with the kerosene, then gets quickly dumped out of the nozzle. Oxygen, however, is notoriously corrosive at high temperatures, so the engines using this mix are understandably of non-reusable designs.

Rockets using liquid hydrogen as fuel (Space Shuttle, Ariane 5), however, face a backwards situation: their fuel is much lighter per molecule than their oxidizer (LOX). That's why their engines run fuel-rich, in addition to the extra fuel lowering their chamber temperatures (hydrogen has good heat absorption properties), giving them much longer engine service life (both vehicles used their main engines almost all the way to orbit).

Also, said Russian engines are able to use a single turbine for both fuel and oxidizer pumps, because the density of kerosene and LOX isn't very far apart.

Edited November 6, 2014 by shynung

[Kryten]

There's nothing preventing oxygen-rich SC engines from having relatively long service lives. RD-170 was designed to last at least ten flights, and NK-33s have gone through multiple series of test firings.

[shynung]

True, there isn't. But it does place quite a strain on the engines' material, so it depends on the engine's particular design and material.

[architeuthis]

Hi

Something that has been confusing me; Isp independent of thrust (unless i missed something), so why do the more powerful engines in KSP have lower Isps? (or is it just a game balance thing )

Shouldn't you be able to build more powerful versions of a rocket engine, while keeping the Isp the same, by building a larger engine that uses more fuel but has the same exhaust velocity?

Or is it just an engineering issue; for a given Isp it's easier to build a 1kN engine than a 1MN engine, and conversely for a given thrust, it's easier to build a less efficient engine? AKA, add more boosters .

I just have the sneaking suspicion that I have missed something here.

Thanks

There is no law of the universe that won't let us have rocket engines with both high thrust and high specific impulse. But there does tend to be a tradeoff. The reason for that tradeoff has to do with energy and power. So isp for a thermal rocket engine is a function of engine pressure ratios, combustion chamber temperature as well as the ratio of specific heats and the molecular mass of the propellant. Here is the equation itself:

Remember that isp=Ve/g0. So you'll notice a few things about this equation. Namely that Ve, and hence isp, goes up if the temperature gets hotter, if the molecular weight gets lower, and if the pressure at the nozzle exit is close to zero. This is why a NERVA type nuclear rocket engine gives such good isp compared to a chemical rocket, M=2 for molecular hydrogen vs. say something like 18 for the hydrogen-oxygen (for a 1:1 stoichiometric ratio) used on the very best chemical rockets. Assuming everything else stays the same the nuclear rockets get about 3 times the performance.

Now lets look at thrust

F is thrust, q is the mass flow rate, Ae is the area of the nozzle exit, pe, and pa are the exit pressure and ambient pressure respectively. Notice a few things? Thrust is a function of Ve!

Now let say you multiply F*Ve, what are the units your result would have? Force*distance/time is the definition of power. This is the problem. If we treat power as a constant, then it is obvious that if we increase thrust, Ve has to go down. This makes sense from a intuitive stand point. If you are trying to heat up cold water flowing out of a pipe, if you reduce the flow rate for the same set amount of heat the water will be warmer, if you speed up the flow rate the water coming out will be colder.

We can always build rockets with more energetic processes. For example the 200 kW VASIMR thruster currently in testing produces only 5 N of thrust (which is a heck of a lot compared to most electric thrusters, but it is still only like 1.1 lbf). If we could pour 200 MW of juice into a scaled up version of the same VASIMR we could supposedly fly to Mars in only 39 days.

Unfortunately, there is a practical problem with just continually adding 'moar power'. Waste heat will build up until it melts your engine. The pesky second law of thermodyanmics says there are no 100% efficient thermodynamic processes. As an example lets say we've got a 10,000 ton wetmass seedship accelerating towards alpha centauri at 1 g. It's rolling with a beam core anti-matter propulsion system that has a isp of .6*c/g0=18.4 million seconds:) The engine thus produces 8.8 TW of thrust power (this is about half of the world's 2008 energy production). This isn't a super realistic example for several reasons, but let's say the engine is 99.99% efficient. It still produces 882 MW of waste heat, which you then have to get rid of somehow. This is an engineering problem of the first order.