This website uses cookies to ensure you get the best experience. Learn more

Asterex 2.0 update: The math, the paywalls, and steel melting

Published on:
June 4th 2018

The MAV, or Mars Ascent Vehicle, is the "rocket" we will use to get off Mars surface. It will propel us to our mothership for transfer back home to Earth. Powered by a cluster of 9 rocket engines, one of them is scheduled for test firing this summer.

Planning for Mars we always felt comfortable leveraging our expedition experience from earth survival to space. Life support (food, water, oxygen), handling extreme environments, staying mentally motivated and physically fit during prolonged discomfort, meticulous planning, rigid time schedules and so forth are all similar.

We also felt confident applying our background as software developers; in 2016 we built a prototype space sensor network that, after 842 days, still logs 100 data points/sec.

As for other hardware/software, after 15 years of using and building tech for expeditions to the most remote and inhospitable places on this planet, we felt competent in that area as well.

Rocket engines and rocket fuel though - that was a different ball game.

We rarely open the hood of our Ford Expedition truck. Jorge, our trusted mechanic up in the Sierras takes care of oil changes and repairs. For Space naturally, we planned to follow Elon Musk and Jeff Bezos, and simply hire über smart rocket engineers.

But 'Man plans and God laughs', so here we are, building a rocket engine from scratch, which includes cooking our own rocket fuel.


Lacking the space tech-moguls' financial resources to recruit rocket talent we decided to build our own engine for the MAV. Initially, it felt like "at least we'll learn something along the way" sort of project. As we dove deeper into the design and added pieces to the puzzle, our skill-set grew, and with it our self-confidence.

The first Asterex version 3D printed in December 2017. We hacked the dimensions by re-engineering historic rocket-engines and adding some crude math. The model was aimed for cold flow tests only (spraying cold water instead of hot fuel) and lacked calculations for thermodynamics, flow mechanics, tensile strength, etc.

The next step is taking the prototype to a fully functional rocket engine. You would think this type of work requires long days at the machine-shop but most of our dev time is actually spent by the computer. The required skill set is very polymath including math, physics and chemistry.

Getting close to our first hot firing (and earning the title of "rocket engineers" by practice) it's time to jot down some key experiences.

By the way, if you're new here, the name Asterex is a false portmanteau of astro (space) and rex (t-rex), plus a tribute to cartoon persona Asterix and the superhuman strength of the magical brew.

Asterex - quick tour

Asterex is a 5 kN bi-propellant rocket engine. 9 Asterex will be used in an engine cluster for Mars landing and take off.

5 kN (kilo Newton) is a thrust equivalent of a Ferrari engine, so leaving Mars will be like sitting on top of 9 sports car engines gaining full speed within a second.

Bi-propellant refers to "gas" running the engine. A mixture of a fuel and an oxidizer, in Asterex case it's very similar to what propelled the Apollo Lunar Lander in 1969: MMH and NTO.

Space might be hard, but space research is harder

Rocket scientists like to proclaim that "space is hard." It's a bit too early for us to oppose that statement but after diving into the rocket engineering literature, we can say it's possibly the most confused technical information we have encountered.

Accustomed to the clarity, simplicity and logic of computer science (if you mess up - the code simply won't run), we expected the same from rocket scientists, reputably the smartest minds of our generation. We found a messy, unstructured world of information, hard to access, but not that complicated when reaching the core. Add to that a great deal of secrecy due to military or business reasons.

On the bright side, fantastic, in-depth information for free on subjects like chemical equilibrium. Detailed technical reports from the fifties and sixties, still highly valid. And almost unparalleled enthusiasm for rocket engineering.

Here go some of our key take-aways.

Meter or feet

The units are a major cluster-f*ck. While the rest of world use the "metric" SI system (meter, second, kilo) three countries in the world prefer the imperial system (foot, second, pound): Liberia, Myanmar and the United States.

That would be fine if it wasn't for the real issue; NASA, along with most US literature and papers, use both systems!  

We've spent hours and hours decoding this unnecessary riddle of units. For NASA it has caused serious errors, the most well-known being the JPL mix-up of newton-seconds and pound-seconds resulting in the loss of the $125 million Mars Climate Orbiter in 1999.

The Pythom project is on strictly enforced SI.

Science/Space paper paywalls

Federal and states contribution to higher education is more than $100 billion per year. This finances more than 50% of University budgets in the US. That's great! Our taxes pay not only for DC corruption and nasty things like wars but also for knowledge and research to benefit all mankind.

And yet, in times of internet with publishing cost nearly zero, almost all US scientific papers are locked behind paywalls.

There is a great Anonymous Documentary about Aaron Swartz who paid a high price fighting the lock-down and monetization of public information.

Unless you are on campus, to access a scientific paper online (that you already paid for with your taxes), will cost you $25 - $35. Researching ignition times, stress tests, combustion temperatures and what else for Asterex, we easily encounter 10 papers per day, of which about eight will prove irrelevant. In the remaining two we'll scan 20-30 pages and collect typically one or two items of interest.

We are prepared to pay $100-$150 for an 800-page book detailing rocket engine performance, but papers at $30 each? At our speed that's $300 per day, or $100 k per year, for little return and a major unit salad on top.

India rocks, China not so much

Beyond US, the scientific papers we need are mainly from Europe or India (no speak Russian). Indians are well known for advances in math and software engineering, and today many qualitative studies come out of India. So far all foreign papers have been open to us.

China, not far behind US and Russia when it comes to Space these days, has been least valuable. We didn't find one single space-related document of Chinese origin during our research.

What happened after March 7, 1970?

Ever wondered why we didn't go to Mars straight after the Moon?

This excerpt from President Nixon's speech on March 7, 1970 shows the killing of the space exploration momentum:

"...we must also recognize that many critical problems here on this planet make high priority demands on our attention and our resources. By no means should we allow our space program to stagnate. But--with the entire future and the entire universe before us-we should not try to do everything at once. Our approach to space must continue to be bold--but it must also be balanced."

Not only did the 70s end human exploratory space travel, something seems to have changed also in rocket science.

In our research, most of our sources are from pre-1970. It's mind-boggling. Despite all the tech advances since we stood on the moon, to figure out fuel mixes and engine parameters we use information that's 50-70 years old!

An important book on liquid propellant rocket engines is from 1971 (Huzel and Huang). 'Rocket Propulsion Element', one of the four books Elon Musk presumably read when starting SpaceX, was published in 1949. It has had several excellent updates since, but the main text is 70 years old.

Our rocket engine library consists of about 10 books, plus around 50 papers and reports. The vast majority are American, all with old publishing dates.

In the modern section, after 2010, we got seven useful papers: 2 American, 1 German and 4 Indian.

The invaluable help from CEARUN

One research team needs special mention. First though, a fast rundown of how rocket engines work.

The essence is combustion, or the release of extreme heat energy when combining two compounds. Think your regular camping stove. The propane hits oxygen in the air, you ignite the mix, and get a flame. That's combustion releasing heat energy.

There is no air in space so we must bring an oxidizer in addition to fuel. We can choose between different mixes, which determines the heat that will create thrust which will propel the rocket. The hotter, the faster. For comparison:

A propane stove gives off around 350 deg C of heat. A rocket engine will typically reach 3000 deg C.

Steel melts at 1500; the surface of the sun is around 5000 deg C.

The mix happens in the chamber of the rocket engine. The blazing heat is channelled through the throat and then the nozzle, where it leaves the engine full force (thrust).

Obviously, the numbers guiding the design of the process are crucial. Some of them can't be calculated, but have to be tested.

In the late 1940's Virgina Morrell and Vearl Huff at a small NACA (now NASA) lab started to systematically measure the combustion forces of different propellants. Read the history here. The results are freely available (thank you NASA) online through CEARUN and a huge help to calculate the exact dimensions of our engine.

What's Next for Asterex

While the equations are mostly pre 1970, our methods and tools are very 2018 as in our last update "...the Space Revolution".

All our code is written in the Python language (not related to our space company, Pythom) and much of our initial work has been to convert 50 years old equations to working python code. Jupyter Notebook has been our main tool for fluent code and instruction intermixed. The result is flexible code, which means we can change fuel mixes and adapt engine dimensions for new thrust levels in less than an hour.

It took 6 months to nail down the parameters. Design of Asterex variable injector (to achieve throttle) took another month. The final CAD work started last week,  manufacturing of the engine, plumbing and tanks will all begin this month and if all goes well we should be firing the engine in Mojave desert later this summer.

Next week we'll also start testing fuel mixes (yes - watch out for the next explosive update).

Tina Sjogren
Tina Sjogren
Share this update:

Design of Liquid Propellant Rocket Engines, by Huzel and Huang from 1971. Must read for aspiring rocket engineers. Foreword by Wernher Von Braun, "Success in space demands perfection... perfection begins in the design of space hardware".

The original Apollo Operations Handbook for the Lunar Module. 800 pages of detailed and still valid design. Pythom MAV will use the same fuel mix and also fuel injectors very similar to the Apollo lander.

Asterex 1.0 cold flow engine was 3D printed last December. The next version will be 3D printed in steel, pack the punch of a Ferrari and have a movable injector for thrust control.

The "Asterex Design Parameter" in Jupyter Notebook (a great tool to combine text, formulas and code in one document). 60 pages calculate temperatures and speed of gases at different locations in the engine and the plumbing. The code is fully dynamic so a change of fuel, for instance, renders a completely new set of parameters within less than an hour.