There is one thing I’ve always wondered while watching Starship launch videos.
The flames erupting from the 33 engines under the Super Heavy booster have a uniquely clear and transparent orange glow.
Compared to the engines of the Falcon 9 or Saturn V I’ve seen in the past, those flames seemed more opaque and yellowish. As I looked into it, the terms “mixture ratio” and “full-flow engine” kept appearing whenever the Raptor engine was discussed.
At first, I thought that adding more oxygen would simply increase the firepower, but as I dug deeper, I realized this wasn’t just a simple fuel-mixing issue; it was a variable that dictates the entire structure of the engine.
What is the Mixture Ratio (O/F ratio) and Why Does It Shake Up Engine Design?
The mixture ratio is the mass flow rate of the oxidizer divided by the mass flow rate of the fuel.
If you set the oxygen to 1 and burn the corresponding amount of fuel, the mixture ratio becomes 1. The problem is that performance does not necessarily peak at the stoichiometric point.
In a domestic study, a small bipropellant engine using hydrogen peroxide and kerosene was tested with mixture ratios ranging from 3.8 to 11.0. The combustion efficiency was highest in the 5~6 range and tended to drop outside of that interval.
Another experiment involving a methane/liquid oxygen engine injector showed slightly different results, maintaining a high combustion efficiency of over 90% across the entire tested range.
This implies that the sensitivity of performance to the mixture ratio varies depending on the propellant combination.
Looking at this, one might think, “Then why not just run experiments and set it to the best point?” But once you move into actual engine design, things get complicated.
The Peak Specific Impulse vs. The Actual Design Point
Specific impulse (Isp), which indicates rocket engine performance, is determined by the product of characteristic velocity and the nozzle thrust coefficient. The factors that most influence characteristic velocity are combustion pressure and the mixture ratio, with the latter showing a curve that peaks at a specific value. However, some engines do not use that peak point at all.
The 75-ton class engine used in the first and second stages of the Nuri rocket is a prime example.
According to data from the Korea Aerospace Research Institute (KARI), the mixture ratio for this combustor is designed at 2.45, even though the peak specific impulse occurs at approximately 2.5~2.6. This is because the lower ratio was chosen to account for combustion chamber cooling. This liquid oxygen/kerosene engine was designed with a propellant flow rate of 243.6 kg/s, a combustion pressure of 6.0 MPa, and a target characteristic velocity of 1730 m/s—intentionally deviating from the optimal point. The reason is simple: while increasing the mixture ratio raises the combustion gas temperature and improves specific impulse, it also increases the thermal load that the chamber walls must withstand. In fact, other domestic analyses of 75-ton class combustors explain that as the mixture ratio increases, the combustion gas temperature rises, which is disadvantageous for cooling. Furthermore, because a portion of the fuel flow is used for film cooling, the nominal mixture ratio that provides the best specific impulse is actually set to a lower value.
It is a constant compromise between performance and survivability. It was fascinating to learn that this isn’t just about tweaking a single number, but a decision that determines how many times an engine can endure and be reused.
Mixture Ratio Troubles During Thrust Adjustment
Even if you nail the mixture ratio at the design point, it doesn’t end there.
If you move the thrust control valve during launch to increase or decrease engine output, the mixture ratio inside the gas generator fluctuates as well.
According to research by the Propulsion Control Group at KARI, when the thrust control valve operates and the engine thrust level changes, the propellant mixture ratio in the gas generator also changes. This change can lead to fluctuations in gas temperature, which may cause damage to turbine blades or a decrease in specific impulse. Therefore, a separate device called a “mixture ratio stabilizer” was developed to manage this instability. It’s a structure where, just to control thrust, the mixture ratio goes haywire, and another component is added to fix it. At this point, I started to get a sense of why developing a full-flow engine is so difficult.
While a gas generator engine only needs one stabilizer to catch mixture ratio fluctuations, a full-flow engine has two independent pre-burners that must be controlled simultaneously as they fluctuate with different mixture ratios whenever thrust is changed. With two turbopumps and two mixture ratio control loops, it’s clear that even a single throttling maneuver requires much more sophisticated real-time control.
Mixture Ratio Dictates the Engine Cycle Itself
Going one step further, the mixture ratio issue doesn’t end inside the combustion chamber. It determines the very method used to drive the turbopumps.
The simplest method is the gas-generator cycle. A portion of the fuel and oxidizer is diverted into a separate combustion chamber (gas generator) to burn, and the resulting gas drives the turbine to power the pump before being exhausted into the atmosphere.
The Nuri 75-ton engine and the SpaceX Merlin engine use this method. While the structure is simple and development is fast, you lose specific impulse because the propellant used to drive the turbine doesn’t enter the main combustion chamber.
To reduce this loss, the staged combustion cycle was created. The exhaust gas from the turbine is not discarded but pushed back into the main combustion chamber. The problem is that with only one pre-burner, you can only stably create either a “fuel-rich” state or an “oxidizer-rich” state.
Interestingly, during the Cold War, the US and the USSR pushed these in different directions.
The US favored the fuel-rich side, like the Space Shuttle Main Engine (SSME).
The USSR had already deployed oxidizer-rich staged combustion engines for ICBMs, such as the RD-263 and RD-0233, since the 1970s.
The Soviets had secured the material technology to prevent turbine blades from burning in the high-temperature, high-pressure, oxidizer-rich environment earlier, a path the US had long avoided.
Full-Flow Engine — The Idea of Burning Everything
So, what happens if you use both fuel-rich and oxidizer-rich flows simultaneously? This is the starting point of the full-flow staged combustion cycle (FFSC). By using two pre-burners, one burns most of the oxidizer and a portion of the fuel to drive the oxidizer turbopump with oxidizer-rich gas, while the other burns most of the fuel and a portion of the oxidizer to drive the fuel turbopump with fuel-rich gas. No propellant is wasted; the entire flow performs work in the turbines before entering the main combustion chamber.
Why is this good? Because you can lower the gas temperature required to achieve the same turbine output.
NASA’s full-flow engine demonstration program data explains that because the entire propellant flow is burned in the pre-burners, there is more mass flow to drive the turbines. This excess power allows for lower turbine temperatures, which increases the turbine lifespan—a key factor in reusable engine durability.
Additionally, by ensuring that oxidizer-rich gas only touches the oxidizer turbine and fuel-rich gas only touches the fuel turbine, you eliminate the need for complex propellant sealing devices between the pump and turbine. Structurally, this reduces the risk of different phases of fluids mixing.
The Concept is 60 Years Old, So Why Did It Only Fly Now?
The idea of a full-flow engine is not actually new.
The RD-270, developed by Soviet rocket designer Valentin Glushko’s team starting in 1962, was already a full-flow staged combustion engine. It used nitrogen tetroxide and UDMH as propellants, reaching a vacuum thrust of 6,713 kN, a vacuum specific impulse of 322 seconds, and a combustion pressure of 26.1 MPa, intended for the Soviet UR-700 heavy-lift launch vehicle. However, after the US reached the Moon first in 1969, the Soviet rocket program was canceled, and development of the RD-270 stopped along with it. It wasn’t due to a lack of technology, but because the rocket that would have used it disappeared.
In the US, the NASA and Air Force Research Laboratory conducted the Integrated Powerhead Demonstrator (IPD) program in the 1990s.
Using liquid hydrogen and liquid oxygen, this demonstrator featured dual oxidizer-rich and hydrogen-rich pre-burners, aiming to keep the engine cool during flight and increase system efficiency. It was designed to produce about 250,000 pounds of thrust and be reusable up to 200 times. In July 2006, it completed 21 out of 26 planned tests, reaching 100% power output. However, subsequent funding for full engine development was not allocated by policymakers, and the design was never completed. The technology was verified and tests were successful, but it stopped due to a lack of budget.
Ultimately, the title of the first full-flow engine to actually fly went to the Raptor engine by SpaceX.
Through the Starship test flights, it holds the record as the world’s first full-flow staged combustion cycle engine to successfully fly. It uses cryogenic liquid methane and liquid oxygen, producing a sea-level thrust of 2.26 MN (Isp 327s) and a vacuum thrust of 2.53 MN (Isp 380s) for the Raptor V2. Since the concept first reached practical design in 1962, it took over 60 years to actually fly.
The core of this delay seems to be material science and precision control.
Preventing metal from burning in an ultra-high-temperature, oxidizer-rich environment, enduring long periods without soot in a fuel-rich environment, and ensuring that two independent pre-burner/turbopump loops operate with millisecond-accurate flow control. You could say it took half a century to combine what the Soviets solved on the oxidizer-rich side and the Americans solved on the fuel-rich side.
Below is a summary of the full-flow staged combustion, gas-generator, and oxidizer-rich staged combustion cycles based on actual engine specs.
Comparison of Representative Rocket Engine Cycle Specs
| Engine | Cycle | Mixture Ratio | Vacuum Isp | Development Status |
|---|---|---|---|---|
| Nuri 75-ton Engine | Gas Generator | 2.45 | - | In Operation |
| RD-263 (USSR) | Oxidizer-rich Staged Combustion | 2.67 | 318s | Deployed (Retired) |
| IPD (USA) | Full-flow Staged Combustion | - | - | Development halted after demo success |
| RD-270 (USSR) | Full-flow Staged Combustion | - | 322s | Canceled |
| Raptor V2 (SpaceX) | Full-flow Staged Combustion | - | 380s | In Flight |
Is Full-Flow Always the Answer?
After summarizing all of this, the full-flow engine seems like the ultimate solution, but that doesn’t necessarily seem to be the case.
Because it requires two sets of pre-burners and turbopumps that must be precisely synchronized, the development difficulty and parts count increase significantly. The fact that there has been only one successful case from the RD-270 to the Raptor illustrates this complexity.
Looking at the Nuri development process, it’s easy to understand why not every launch vehicle aims for full-flow.
KARI produced 17 units of the 75-ton engine and conducted 152 combustion tests, totaling 15,091 seconds. For a medium-lift launch vehicle with a fixed budget and schedule, building reliability through the proven gas-generator cycle was a rational choice. On the other hand, SpaceX had the goal of clustering dozens of engines like Starship and reusing them hundreds of times, giving them the incentive to take on 60 years’ worth of risk to fundamentally eliminate turbine life and sealing issues.
I’m not sure if all large reusable engines will converge toward the full-flow design or if cycles will continue to diverge based on their purpose.
However, knowing that the single number called “mixture ratio” is tied to everything from combustion chamber cooling to turbine materials and the choice of the entire engine cycle, I think I’ll look at the color of the flames differently the next time I watch a launch.