Visual walkthrough — Expander cycle — hydrogen-cooled nozzle drives turbine

This is the child of 3.3.24 Expander cycle — hydrogen-cooled nozzle drives turbine (Hinglish). If a word here feels unexplained, that is a bug — tell me and I fix it.

---

Step 1 — What is "mass flow" and why does everything start there?

WHAT. Before any physics, we need one number: how fast fuel is moving through the engine. We call it the mass flow rate, written $\dot{m}$ (read "m-dot"). The little dot on top is a physicist's shorthand for "how much per second". So $\dot{m}$ means kilograms of stuff passing a point every second, measured in $\text{kg/s}$.

WHY this quantity and not, say, total mass? A rocket does not care how much fuel exists in the tank at one instant; it cares about the stream — a river of fuel flowing continuously. Everything downstream (heat picked up, work done) is a rate (per second), so we must start with a rate of mass.

PICTURE. Look at the pipe below. Every second, a slug of hydrogen of mass $\dot{m}$ crosses the dashed cyan line. That is the river we will follow all the way through the engine.

---

Step 2 — The cooling jacket steals heat: $Q=\dot{m}{c}_p\Delta T$

WHAT. The cold hydrogen enters channels wrapped around the nozzle at temperature $T_{in}$ and leaves hotter at $T_{out}$. The heat it soaks up per second is $Q=\underset{\begin{array}{c}\text{kg per}\\ \text{second}\end{array}}{\underbrace{\dot{m}}}\cdot \underset{\begin{array}{c}\text{joules to heat}\\ \text{1 kg by 1 K}\end{array}}{\underbrace{c_p}}\cdot \underset{\begin{array}{c}\text{kelvin the fuel}\\ \text{warmed up by}\end{array}}{\underbrace{(T_{out}-T_{in})}}$

WHY this formula? The symbol $c_p$ is the specific heat capacity: it is defined as the energy needed to warm one kilogram by one kelvin (at constant pressure — that is what the little $p$ means). So joules = (kg) × (joules per kg per K) × (K). Multiply by "per second" from $\dot{m}$ and $Q$ comes out in joules per second, which is watts ($\text{W}$) — a power. We use $c_p$ and not some other property because our question is literally "how much heat to warm this gas?", and $c_p$ is the quantity built to answer exactly that.

PICTURE. Follow the temperature bar along the channel: cold blue at the inlet, hot amber at the outlet. The area of heat flowing in (red arrows from the flame) is the $Q$ we banked.

---

Step 3 — Heat becomes enthalpy, the turbine's food

WHAT. All that heat raised the gas's enthalpy — think of enthalpy $h$ as the "usable energy content" of the flowing gas. For an ideal gas there is a beautifully simple link: $h=c_{p}T$ Each symbol: $h$ is energy per kilogram ($\text{J/kg}$); $c_p$ we just met; $T$ is absolute temperature in kelvin.

WHY do we switch from "heat" to "enthalpy"? A turbine is not fed heat directly — it is fed a hot, high-pressure gas stream. The property of that stream the turbine can convert into shaft work is enthalpy. So we translate the heat we stole (Step 2) into the currency the turbine actually spends: enthalpy. The equation $h=c_{p}T$ is the exchange rate, and it is exact for an ideal gas.

PICTURE. A thermometer rising = enthalpy tank filling. Hotter gas = fuller tank = more the turbine can extract.

---

Step 4 — The turbine spends enthalpy by dropping pressure

WHAT. The hot gas flows into the turbine at pressure $p_1$ and leaves at a lower pressure $p_2$. As it drops in pressure it also cools, and the enthalpy it loses becomes shaft work. The ideal (loss-free) work rate is $W_{turb}^{ideal}=\dot{m}{c}_p{T}_1\left[1-{\left(\frac{p_2}{p_1}\right)}^{\frac{\gamma -1}{\gamma }}\right]$

Term by term:

• $\dot{m}{c}_p{T}_1$ — the enthalpy rate arriving at the turbine inlet (Steps 1–3, with $T_1=T_{out}$).
• $\frac{p_2}{p_1}$ — the pressure ratio; less than 1 because the gas expands to a lower pressure.
• $\gamma$ (gamma) — the gas's heat-capacity ratio; for hydrogen-like diatomic gas $\gamma \approx 1.4$. It measures how "springy" the gas is.
• The exponent $\frac{\gamma -1}{\gamma }$ and the bracket come from the isentropic rule below.

WHY this exponent — where does $(p_2/p_1{)}^{(\gamma -1)/\gamma }$ come from? An ideal, loss-free expansion is isentropic (constant entropy — a fancy word for "no wasted disorder"). Isentropic flow ties temperature to pressure by $\frac{T_2}{T_1}={\left(\frac{p_2}{p_1}\right)}^{\frac{\gamma -1}{\gamma }}$ (see Isentropic flow relations). The work is the enthalpy drop $\dot{m}{c}_p(T_1-T_2)$. Substitute $T_2=T_1(p_2/p_1{)}^{(\gamma -1)/\gamma }$ and factor out $T_1$ → exactly the bracket. We use the isentropic relation because it is the best possible case; it tells us the ceiling on turbine output.

PICTURE. Gas balloon pushing a paddle wheel as it expands from high (amber) to low (cyan) pressure; the shrinking temperature bar is the enthalpy being handed to the shaft.

---

Step 5 — Real turbines waste some: multiply by ${\eta }_t$

WHAT. No real machine is loss-free. We tack on a turbine efficiency ${\eta }_t$ (eta-t), a number between 0 and 1: $W_{turb}=\dot{m}{c}_p{T}_1\left[1-{\left(\frac{p_2}{p_1}\right)}^{\frac{\gamma -1}{\gamma }}\right]{\eta }_t$

WHY multiply? ${\eta }_t$ is defined as (real work) ÷ (ideal work). So real work = ideal work × ${\eta }_t$. If ${\eta }_t=0.75$, the turbine captures 75% of the theoretical maximum; the other 25% escapes as swirl, friction and leftover heat.

PICTURE. Two bars side by side: the tall ideal bar and the shorter real bar, the gap labelled "losses".

---

Step 6 — What the pumps demand: $W_{pump}=\frac{\dot{m}\Delta p}{\rho {\eta }_p}$

WHAT. The shaft that the turbine spins drives the pumps. A pump raises a fluid's pressure by $\Delta p$. The power it needs is $W_{pump}=\frac{\dot{m}\Delta p}{\rho {\eta }_p}$

Term by term:

• $\Delta p$ — the pressure rise the pump must supply ($\text{Pa}$).
• $\rho$ (rho) — the fluid density, kilograms per cubic metre. Dense fluids need less volume moved.
• $\frac{\dot{m}}{\rho }$ — volume flow (${\text{m}}^3/\text{s}$): mass per second ÷ mass per cubic metre = cubic metres per second.
• ${\eta }_p$ — pump efficiency; we divide because a real pump needs more input than the ideal.

WHY does pushing on volume cost $\Delta p$ per unit volume? Work is force × distance; pressure is force per area. Push a piston of area $A$ up a pressure step $\Delta p$ over distance $d$: work $=(\Delta p\cdot A)\cdot d=\Delta p\cdot (Ad)=\Delta p\times \text{volume}$. So each cubic metre lifted through $\Delta p$ costs $\Delta p$ joules. Multiply by volume-per-second and you get watts. We divide by ${\eta }_p$ because friction means you always pay extra.

PICTURE. A piston shoving fluid up a pressure "hill"; the swept volume shaded, height of the hill = $\Delta p$.

---

Step 7 — The self-sustaining condition (the whole point)

WHAT. Assemble Step 5 (supply) with two copies of Step 6 (demand), one per pump. The engine keeps running only if the turbine makes at least what both pumps eat: $\boxed{\underset{\text{supply — from nozzle heat}}{\underbrace{{\dot{m}}_{H_2}{c}_p{T}_1[1-(p_2/p_1{)}^{\frac{\gamma -1}{\gamma }}]{\eta }_t}}\ge \underset{\text{demand — fuel pump + ox pump}}{\underbrace{\frac{{\dot{m}}_f\Delta p_f}{{\rho }_f{\eta }_{p,fu}}+\frac{{\dot{m}}_{ox}\Delta p_{ox}}{{\rho }_{ox}{\eta }_{p,ox}}}}}$

Term by term on the demand side:

• ${\dot{m}}_f,\Delta p_f,{\rho }_f,{\eta }_{p,fu}$ — the fuel (hydrogen) pump, exactly the Step 6 formula with fuel subscripts. Note ${\dot{m}}_f={\dot{m}}_{H_2}$ — the same hydrogen stream we followed through the jacket and turbine.
• ${\dot{m}}_{ox},\Delta p_{ox},{\rho }_{ox},{\eta }_{p,ox}$ — the oxygen pump, the same formula with oxidizer subscripts and its own efficiency.

WHY an inequality and not an equals sign? If supply exactly equals demand the engine can only just tick over; any hiccup stalls it. Real engines want $W_{turb}$ larger than demand so there is margin for startup and throttling. If the left side falls below the right, the shaft slows, pumps weaken, flow drops — the engine spins itself down.

PICTURE. A balance beam: turbine power on one pan, summed pump power on the other. The cycle "closes" when the turbine pan sits level or lower.

---

Step 8 — The degenerate cases the beam must survive

Every scenario must be covered, so let us push the formula to its limits.

Case A — no heat picked up ($T_{out}=T_{in}$). Then $Q=0$ and the gas never gets hot, so the turbine inlet temperature $T_1$ stays low. The bracket $[1-(p_2/p_1{)}^{(\gamma -1)/\gamma }]$ is unchanged (it depends only on pressures and $\gamma$), but it multiplies $T_1$ — and $T_1\to$ small drives $W_{turb}\to$ small → cycle cannot start. This is why a cold engine needs a start cartridge or tank-head bleed to get the first trickle of hot gas.

Case B — zero pressure drop ($p_2=p_1$). Then $p_2/p_1=1$, the bracket $1-1^{...}=0$, so $W_{turb}=0$. A turbine that does not drop pressure does no work — obvious, and the formula agrees. Expanders deliberately keep this drop modest (Example 2 used only 12→8 MPa) because the gas must still be high-pressure enough to enter the chamber afterward.

Case C — scaling up (the square–cube trap). Multiply every length by $k$. Heat is picked up over the nozzle surface area, which grows as $k^2$, so heat supply $\propto k^2$. But the propellant mass flow the pumps must move scales with the chamber's throughput — with volume/throat, i.e. $\propto k^3$. Since pump power $W_{pump}\propto \dot{m}\Delta p$ and $\dot{m}\propto k^3$ (at fixed $\Delta p$), the demand grows as $k^3$. Compare:

$\frac{\text{demand}}{\text{supply}}\propto \frac{k^3}{k^2}=k.$

So as the engine grows, demand outruns supply and the right pan of the beam eventually crashes down. This is the Square-cube law and it caps closed expanders at a few hundred kN of thrust — a geometric limit, not laziness. (This is why big engines use gas generator or staged combustion instead.)

---

The one-picture summary

Here is the entire derivation compressed: the hydrogen river (Step 1) enters the jacket and drinks heat (Step 2 → enthalpy, Step 3), spends it in the turbine as it drops pressure (Steps 4–5), the shaft carries that power to two pumps (Step 6), and the beam decides whether the loop closes (Step 7), forever shadowed by the square–cube law (Step 8).

Recall Feynman retelling — say it back in plain words

Cold hydrogen flows in a stream ($\dot{m}$). It runs through pipes around the roasting nozzle and drinks heat — how much depends on its enormous appetite $c_p$ and how many degrees it warms ($Q=\dot{m}{c}_p\Delta T$). That heat becomes "usable content" (enthalpy, $h=c_{p}T$). The hot gas then squeezes through a turbine: as its pressure falls from $p_1$ to $p_2$ it cools, and the lost enthalpy turns a shaft ($W_{turb}$). Real turbines keep only a fraction ${\eta }_t$ of the ideal. That shaft drives two pumps — one for hydrogen fuel (${\dot{m}}_f$, efficiency ${\eta }_{p,fu}$), one for oxygen (${\dot{m}}_{ox}$, efficiency ${\eta }_{p,ox}$); each pump's cost is "how much pressure I add times how much volume I move, divided by how leaky I am" ($\dot{m}\Delta p/\rho {\eta }_p$). The engine runs forever only if the turbine's output is at least the two pumps' combined bill. It won't start with no heat (because $T_1\to 0$), does nothing with no pressure drop, and — because heat comes from area ($k^2$) while pump work comes from mass flow ($k^3$) — it can never be made huge.

Recall Quick self-test

Why can't we replace $c_p$ with a different property in Step 2? ::: Because $c_p$ is defined as heat per kg per kelvin — it is precisely the number that answers "how much heat to warm this gas?" Where does the exponent $(\gamma -1)/\gamma$ come from? ::: From the isentropic $T$–$p$ relation $T_2/T_1=(p_2/p_1{)}^{(\gamma -1)/\gamma }$, the best-case loss-free expansion. Why divide by ${\eta }_p$ for pumps but multiply by ${\eta }_t$ for turbines? ::: Pumps need more input than ideal (÷), turbines deliver less output than ideal (×). What do the subscripts $f$ and $ox$ mean, and why does each pump get its own ${\eta }_p$? ::: $f$ = fuel (hydrogen), $ox$ = oxidizer (oxygen); each is a different machine moving a different fluid, so each has its own efficiency ${\eta }_{p,fu}$, ${\eta }_{p,ox}$. Why do expanders stay small? ::: Heat supply grows as area ($k^2$), pump demand as mass flow ($k^3$); demand/supply $\propto k$ eventually breaks the balance.

---

See also: Specific impulse · Regenerative cooling · Isentropic flow relations · Turbopump design · Square-cube law · Liquid hydrogen properties