ELECTRON — New Zealand, 2018

These guys are making a success of the tiny rocket business, as they cut the cost per launch to $6 million despite having no reusability yet. One market is “cubesats” — tiny orbital devices consisting of one or more cubical sections with a standardized size (10 cm) and limited mass allowance per cube (1.3 kg). This rocket could launch a hundred or more of them at a time. They originally aimed to schedule such launches very frequently, like once or even twice a week, so that small budget satellite customers wouldn’t have to wait for a berth on a big rocket, which can take months or even years, and might make you settle for a less than ideal orbit. The company is called Rocket Lab, and one of the seed investors is actually named Mark Rocket, though he is not a founder. They moved the company to Los Angeles, but the main launchpad is still in New Zealand, on a cliff at the end of a peninsula, making it the prettiest launchpad in the business, as well as (they hope) being capable of really fast turnaround for frequent launches.

For 2020 they built a new launchpad at Mid-Atlantic Regional Spaceport on Wallops Island, Virginia. This will be more suitable for equatorial orbits. They didn’t launch from it until early 2023. They also expanded the New Zealand site by giving it two launch pads within the same complex. By mid-2024 they’d done fifty launches, making them the fastest company to ever reach that milestone, and still no strong competitors aside from Chinese military-derived solid rockets.

design

This little carbon fiber missile is a two stage kerosene burner based on a very compact and inexpensive engine they call the Rutherford, after the famous New Zealand-born physicist. It’s small enough for one person to easily pick up, and can be made very quickly with 3D printing. It was the first rocket engine to use electric motors to pump the fuel — hence the rocket’s name. (I do not know if the name was also chosen to complement Russia’s big Proton.) This electric motor makes the engine much more efficient in its use of fuel, with the tradeoff that the rocket has to lift a big pack of lithium batteries. Each engine has two soda-can-sized brushless DC motors of 37 kilowatts each, or 50 horsepower.

Like a quarter-scale Falcon, the Electron’s booster uses nine Rutherfords, and the second stage uses a single Rutherford with a vacuum bell. Their first flight attempt came up just short of reaching orbit, and their second delivered a payload. They’ve also got a tiny third stage, or “kick stage” as rocketry nerds call it, available for tasks such as raising or circularizing orbits. It doesn’t extend the rocket stack, it just fits inside the fairing. It has a little engine called the Curie which consumes monopropellant. It doesn’t use hydrazine — they’ve got some “green” alternative. I don’t know what it’s made of, but it might be based on hydroxylammonium nitrate (NH₃OHNO₃) liquified in a solvent. The miniscule thrust allows orbital insertions to be extremely precise.

They went on to evolve this kick stage into a general-purpose satellite bus onto which customer devices could be mounted, so you can just build the instruments you want and not the complete satellite. They call this the Photon. It can be had in various sizes, with monopropellant or bipropellant. A basic Photon takes up around 100 kg of your payload allowance. They have a stretch version of it which is intended to send a tiny payload to the moon, and they had a customer who wanted to use two electrons for a “Moon Express” mission. (That launch was supposed to happen in 2019, but was delayed for years and I don’t think they’re pursuing it anymore.) This version has an upgraded “Curie 2” or “HyperCurie” engine which has electric pumps where the original was pressure-fed, and is still not much bigger than your hand, aside from the flare of the bell. It has higher specific impulse than most kick-stage engines... but they aren’t saying what its fuel is. The Photon has solar cells on the back around the Curie nozzle in a style kind of like a Starliner service module, and on the sides too.

They never got their launch cadence anywhere near a weekly pace, but they are still easily the most successful of the new batch of smallsat launch companies so far. With just four launches in 2018 and six in 2019, they hoped to get to a fortnightly pace soon, if they could streamline and automate enough manufacturing steps. But 2020 saw only seven and 2021 dropped back to six, before they managed to climb to nine in 2022. Yet despite the shortfall relative to their stated ambitions, even the moderate pace they’ve got so far is enough to position them solidly as the dominant leading brand among new small-launch companies.

reuse

And now, after CEO Peter Beck saying they would never pursue reusability, Beck has officially eaten his hat (shredded with a blender) and started a reuse project. The approach is to put a para-wing in the interstage, and as a drogue, a small balloon which will add drag at high altitudes during reentry. (Beck says they might also look at air brakes on the bottom of the stage.) They will not burn any fuel to slow down the booster — they’ve got none to spare. They will have some tiny cold-gas thrusters to keep it straight, but these are only needed at subsonic speeds. Before that, surviving the reentry heat is quite a challenge, but they already had a lot of heat shielding on the bottom just to protect the parts from the rocket flame. After some early tests, they also found benefit in adding shiny foil to the sides of the booster, to reduce the heat intake of the underlying black finish.

The last step of the reentry was originally going to be to pluck the para-wing out of the air with a helicopter, which is a fairly well-understood technique that the Air Force was doing fifty years ago. Beck said helicopters are way cheaper to operate than ships. In 2022, on the first attempt to catch a booster from a real launch, the chopper did indeed catch it, but then the pilot felt it doing something nasty to his stability, and cut it loose again. The chute reopened and it splashed down softly, with little or no damage. They kept trying... and eventually gave up and decided to just let the boosters splash down. It only needed a bit more waterproofing.

Though chutes and balloons are lightweight, with a rocket so small that weight becomes significant, so even the lightest possible approach does take a bite out of their payload capacity. Fortunately, weight on the first stage imposes only a fraction of the cost that second stage weight does, because it applies to only part of the flight, so from what Beck says, the capacity penalty of the chutes may be well under thirty kilograms.

They originally said that the whole purpose of reuse was not to launch cheaper, but to launch more often — without reuse, their manufacturing process was keeping them stuck at a cadence of at most one launch a month. But this is not an issue anymore, as a pace of one launch every five or six weeks was the most they’re finding buyers for, and they had manufacturing capacity going underused. At this point it is about saving money, not time. But in 2024 some customers started buying in bulk, and their cadence jumped sharply. But at the same time, they quietly dropped most attempts at booster recovery, perhaps because these payloads were too heavy.

In early versions the total payload capacity was only 150 kilograms, but soon they raised it to 225, and in 2020 they got it up to 300 by improving some components and lightening the batteries.

Reliability has been pretty decent. The first test flight in 2017 failed, but it had no payload, and the problem was actually on the ground. The next failure was in 2020 on the thirteenth launch, from bad wiring in the upper stage. The next year saw the twentieth launch fail with a different upper stage problem, and in 2023 the fortieth had yet another upper stage electrical issue.That still leaves then with more successful launches than most of their commercial competitors combined.

It’s not uncommon for rockets to have the majority of their problems up top, even though booster stages are more complex and under more stress. I’d guess this might be because boosters are a lot easier to test under realistic conditions, and a lot easier to examine the pieces of after a failure.

Neutron

Another thing Beck said they wouldn’t do is develop a larger rocket. But in 2021 they announced the forthcoming Neutron, which would have an eight ton capacity with reuse of the booster... and be human rated. The idea is to use an ultralight upper stage, enclosed in a five meter fairing which is part of the booster and gets reused. (Five meters is a very common fairing diameter for large rockets.) Normally, a significant part of the dry weight of a second stage is a sturdy outer wall to support everything stacked above it, but on the Neutron that wall goes back down with the first stage, so the second is basically just bare tanks, hanging suspended from the support platform that the payload rides on. (An existing rocket with a bare second stage somewhat like this is the Astra.) The main stage was originally going to use fixed fins as legs — an idea SpaceX pursued for a time but then dropped. They chose this because it was less trouble than having legs with hinges, but then they changed their minds and added fold-out legs, making the fins slimmer. It will have upper steering fins for use during landing, so it can glide a bit.

Despite the human rating, Beck says the Neutron will be mainly intended for launching swarms. Their customers are still people with small satellites, but some of them want to launch them in batches. The Neutron’s first stage will do vertical landings like a Falcon. At first he said it would always go back to the launch site — no sea landings because ships are expensive — but later they admitted that this was a mistake and they would use a landing barge. Its short thick body will be carbon fiber — even the outer heat-protection layer is just another type of graphite composite, and they say it’s good enough to not need a re-entry burn. The engines for it are being newly developed, and will burn methane. They originally went for a conventional gas generator cycle, as the electric pumps in the Rutherford don’t scale up to that size nearly as well as turbines do. Beck wanted this engine to be “boring”. But then they switched the approach to a fuel-ruch staged combustion cycle like the Blue Origin BE-4. They also changed the booster’s engine count from seven to nine at this time (in the earliest drafts they had intended to use just four), and bumped the payload capacity to 13 metric tons. The engine, now dubbed “Archimedes”, breathed its first fire in mid-2024. They won’t launch the Neutron from New Zealand, but only from Wallops in Virginia. Their goal is to fly it cheaper than a Falcon 9, which is probably doable, but they don’t expect to beat the Falcon’s price per ton.

Rocket Lab started bidding the Neutron for government launches, and promising them that it would be ready within the time frame of other competitors, which meant they’d have something flying by the end of 2024. The government took a look at this assertion and didn’t buy it. An internal report accused Rocket Lab of a “campaign to misrepresent their launch readiness in an effort to gain competitive advantage”, and it leaked, as well it should. Beck has said that they are “tracking [the deadline] pretty closely”, meaning that even the maximum possible amount of startup optimism left no room for anything to slip. This plan had them reaching orbit in ten months when the engine hadn’t even reached the test stand yet.

Beck has been accused of abusive management. As the company grows and matures, he still expects people to work tons of overtime as if it were still a little startup, promising big rewards later for “key contributors”. Words like “culture of fear” and “toxic” were used. Some of the pressure may be because the company is still losing money every year.

Electron: mass 12.5 t (early ones were lighter), diam 1.2 m, thrust 224 kN, imp 3.4 km/s, electric pump (kerosene), payload 0.3 t (2.4%), cost $20M/t, record 47/1/3 through August 2024.