Federico Rossi: “After making the first aerospike and methalox engine in Europe I can retire”


Engineering is not an easy job. If we talk about what is behind creating a unique engine in the world, it is another level. Simply, not everyone is qualified to work at these levels and it is not even an individual job. The number of hours of work, variables to study and analyze, and the complexity of the problems that occur in a rocket engine is unique to these systems. But interviewing the entire team was impossible, so we spoke to the person behind everything, Pangea Aerospace's propulsion director, Federico Rossi.

CGI of what Arcos will be like
Computer generated image of what Arcos will look like. The copper color stands out due to the alloy in which it is made. Source: Pangea Aerospace

A few minutes after speaking with his boss, Pangea CEO Adrià Argemi, we met with him. One of those responsible for the company's success also has reasons to be happy: they have just tested the Unyx satellite engine.

However, we had to start big.

AstroAventura: We start with what I believe is the most important part of an engine, the turbopump. In what state is the development of yours?

Federico Rossi: We have a team of four people that works on it. They have a lot of experience at Blue Origin, at General Electrics in turbines, which has allowed us to PDR (preliminary design review) of the turbopump in just three months. Very powerful software that we use for design has also influenced.

We are now in the post-PDR phase. We are analyzing in CFD [Compuer Fluid Dynamics, computer fluid dynamics simulations] off design features. This serves to characterize the cavitation [phenomenon in which a liquid fluid turns into a gas in a turbine due to pressure and/or speed]. And we are going to test the first components in the second part of the year with compressed nitrogen gas.

What design have you chosen for the turbopump, electric or conventional machinery? Can you tell us about that?

It is more conventional, at first we were big fans of the electric one, when we started also because the Meso was smaller and we had less power in the engine. Now that the Meso and the motor have grown, up to 300 kN, basically, the batteries can't handle it. When we started there were many promises about the batteries that have not been fulfilled, they are still poor for this. We have two turbopumps of almost 2 mega newtons each, imagine the gigantic batteries we would need.

You have mentioned that you will carry out the first tests in the second part of the year, so for now, are you limited to CFD?

Yes, at the moment we are limited to CFDs and we are now developing the test benches, to see how we can work in the extreme trading range.

Three months is a very short time for a PDR, how has the experience provided by other team members influenced it?

A lot, this is the key, we have a member of the team who has already worked at Blue Origin, at Ursa Major as well, which is another company in the United States that makes engines. And in the turbine part, the most complex part of the turbopump, another member has twelve years of experience at General Electrics. There he did not work on turbines for rockets, but for aviation, which are even more complicated.

Another leg of the table is the use of commercial software to develop it that is very well done. Boeing has also used it and it allows you at the preliminary design level to do super-fast development of exploring many configurations at the same time.

This is space, it is very complicated to give dates, but when do you expect to have a first operational functional model?

In one year, of the turbopumps.

And being more realistic?

In a year and two months. Yes, we do things quickly, I always say it, we do what we promise. We made the Demo P1, the first aerospike in the history of methalox, between four people and in a year and a half.

The Arcos will not use liquefied natural gas, not even methalox but bio methalox. Can you tell us a little about the differences between these fuels and the percentage of loss of one versus the other?

Natural gas depends a lot on where you get it. In Russia there is a lot, but now it cannot be taken out, there is also in Norway. So the purity roughly is about 98.5%, with the rest being propane and a little bit of nitrogen. While bio methane is more precise because it is always the same process to generate it, we have a purity of 99.95%.

Was your department the one that most promoted the idea of using methalox?

I would say yes, we have pushed a lot for reuse. It is key and methalox is, of the hydrocarbons, what allows you to reuse the most. Hydrolox might be even better, but it's expensive. In addition, it makes the motor very large due to its low density. And in an aerospike it is essential that the engine not be enlarged by the propellant.

One of the problems with aerospikes is that they are large, which makes them weigh a lot. Approximately how much does your engine weigh dry?

We have a TWR target that is between 85 and 100. It will weigh about 300 or 350 kilos. But, we also share many structural parts with the vehicle. Due to its shape it can be integrated much better with the rest of the rocket and have a much more efficient thrust transmission.

You mentioned a TWR target of between 85 and 100, that, discounting the Raptor, places it as the best TWR in the world at sea level.

Yes, correct, it is a very light engine also thanks to additive manufacturing and light alloys. Another help to lower weight is the pressure in the combustion chamber.

About how much pressure are you working on right now?

Less than 100 bars.

How much exact thrust will the engine have?

300 kilonewton at sea level. We have to try it, because an aerospike has never been tested at altitude but we are convinced that we can handle them for CFD simulations. You can't rely on simulations alone, but they are very good nowadays and give you a good intuition of what is going to happen.

Why not engine combinations? It remains one of the things that surprises me most about your approach.

In fact, we have a cluster at the end. If you see the ring, this ultimately works as a cluster. It allows you to be much more flexible, in addition, each of these modules [points to the attached piece below] are the most complicated part of the engine and can be produced more in series which makes them cheaper.

Mock up ARCOS aerospike engine
Collection of models of the Arcos engine at the Pangea headquarters. Those that look like conventional nozzles are called engine modules (the first one in front). Source: Pangea Aerospace

How does it feel to be the developer of the first aerospike engine in Europe?

Very good, one of the best things is when I studied the subject and read papers on aerospikes. Because here in Europe there has been a lot of research and there are very important names. And after the Demo P1 tests, having these people come and ask me the questions was great. After the tests I thought, I can retire, everything I do from now on is unnecessary.

But you have not been the only ones to build and test an aerospike engine in Europe but also the first to build and test a methalox engine in Europe...

Correct. That for me is also a big challenge, but I have to say that it has always been natural for us. Since we did not have people with experience in kerosene engines, it was not a risky decision. Furthermore, I have studied all my life in Italy, that methalox is very strong, there is a lot of research. And I also worked with the Avio M10 engine that will use this fuel.

We return a little to the technical questions, how do you start the engine?

Ignition is done with an augmented spark igniter, which is basically a very small rocket engine that is inside the engine that ignites it, using the same propellants, oxygen and methane, but gaseous.

You can comment on the oxidant-oxidant ratio.

Yes, around 3.4 like all engines in this range.

How many combustion chambers will the Arcos have?

24, it will have that number because each of the chambers has 12.5 kilonewtons of thrust so it is easy to prove it that way.

Also because of those 24 it is symmetrical, and you can make good CFDs and everything, well that is joke. [Referring to CFD being never really an easy task].

No, but this way we can separate it into four sectors, in each one there will be a camera with linear actuators that will allow the TVC to be carried out. [thrust vector control].

PLD Space, the other major Spanish rocket company, has worked on the Miura 1, a suborbital launcher. Have you thought about or looked for some type of low-level jumper or vertical fire tests for your engines?

We have thought about it many times and in the end the problem is money. With Demo P1 we did try to look for entities interested in putting it in a demonstrator but without success. I understand that it is an asset to have a vertical stand because you are recreating exactly what happens in the rocket, but in the end, at these flow rates, gravity does not matter.

So, in your opinion, it is not very important to test the motors vertically.

No, except the turbopumps. Gravity does matter in them and you have to test them vertically. In Arcos we will do it in a test bed of the French space agency that allows it.

What is the upper limit you have with Construction Bows? What is the limitation?

Arcos is limited by the test benches in Europe. Due to its modularity and that the bigger you make the aerospike, the better, we can reach 2 Mega Newton, which is what the DLR bank, the P5, allows.

Ariane 6 upper stage installed for tests scaled
Image of the installation of the upper stage of an Ariane 6 at test stand P5. Source: ESA

We also move on to the second stage. Why simply use a conventional nozzle, even if taken from the Arcos, instead of an aerospike that would allow complete reentry and reuse of the rocket?

They are priorities, basically it has to work, soon. Making a hypersonic reentry takes more time to develop. We have two heads, one that has to develop things soon and another that looks to the future. The second head already thinks about the aerospike in the second stage because it is exceptional. Stoke Space, is a great inspiration and is doing what we want to do. The advantage of the aerospike as an upper stage is that you can maximize the expansion ratio by shortening the engine and also reduce interstage weight.

The other head has to look at what can be done in a limited time. And since we have something modular with Arcos, we have seen that it adapts very well to be upper stage modules. So we are going to reuse the same technology as with Arcos.

Will they use a turbopump?

Yes, we are going to reuse exactly what we are using with Arcos. Although we have looked at using electric pumps, because for a higher stage they are not bad.

You have mentioned that you are going to use generator gas, but what other pressure cycles have you proposed?

We have to maximize performance and cost. We have gone for a gas generator. But the company's ultimate plan is to make an engine with an expansive double cycle aerospike.

With the amount of people there are with knowledge of hydrolox, this is the fuel of Europe. Why haven't they opted for him?

It is a good question, in the end we are not closed to hydrolox and in the future we will surely investigate it. Furthermore, it is very fashionable and its waste is water, but when the company started, hydrogen was too big a challenge.

There is an issue with the ignition of the second stage, there are two types: hot or cold. Which one do you choose?

It is one of the last things we have said, it will be after the separation [cold].

We also move on to questions about reuse. What is the part that worries you the most?

Reentry has never been tried with an aerospike and I imagine there will be problems. In the end, I'm not worried about the engine failing, it won't because we cool it during this phase. But later, if there is a fracture, the engine reconditioning becomes complicated and economic return may not be achieved.

Meso reentry
Image of what the re-entry of the Meso would be like. Source: Pangea

What number of launches do you think you can achieve with these engines?

We have a goal of ten, so we are designing it to hold forty.

You can talk about the complexities of using CFD. Because computer simulations have improved a lot, but they are still not exact. What kind of problems have you encountered with these types of issues?

CFD is delicate, we have validated many of our models thanks to Demo P1 tests. So, if I have to give clearly CFD advice, it must be validated. At first, it can give you a good orientation, then you have to check.

Recovering the second stage, the Demo P1 lasted five minutes in ignition. But a second stage has to burn a lot of time, a lot of fuel. How long are you going to have it on and how long will you have to test it?

With Demo P1 it was tested for 5 minutes because the propellant ran out. With the second stage we want to test until it breaks. Simulations show that we can do it for thirty minutes easily.

So how long is the second stage going to be burning fuel in a real launch?

More than ten minutes.

And the first stage?

Two and a half minutes, more or less.

I'm surprised by so little time in the first stage.

Yes, but we are going to use the aerospike where it is best, which is in the atmospheric part.

What will be the TWR of the second stage?

Something higher than the first, between 120 and 150. Something conventional, the second stage has nothing special.

How do you perform recovery without retro propulsion?

Catching it in the air. Like Rocket Lab on the Electron, with a helicopter. After many studies we have come to the conclusion that it is the best way. [Among them some with real, scale models financed by the ESA].

Having to reuse a rocket, what part is most complicated due to reuse?

Many things (laughs). To start the ignition system. Because it could be something pyrotechnic because you throw a bomb in the combustion chamber, you shoot and that's it. And not. That has to be turned on many times.

But, the most complicated is the regenerative cooling system. Many things must be taken into account here. Especially the low cycle fatigue that will be how this engine breaks for the first time.

To operate a rocket you need TVC, but also cold nitrogen propellants or similar, are you already working on RCS?

We simply have an external supplier with engines of this type. They are an old company, and so we take the propellers from the catalog based on the specifications.

I also want to talk about the 1N engine that you have for propulsion in space. How did the idea come about? What has the development process been like?

Very interesting. The idea arose because we have seen that now satellites have to move and have to deorbit in a very short time and there was a lot of market for these systems.

There are other systems, but they use hydrazine, which is carcinogenic, or they can only fire for a few seconds before stopping to cool down. Our engine is green, has hydrogen peroxide and something else I can't comment on and can fire for a virtually unlimited time. It is something unique in the entire world.

I tried it at the time with both Rocket Lab and the Capstone propulsion suppliers and they did not want to comment on it either, because there is a lot of competition with the issue of green rocket propellants.

Yes, it is an incredible alchemy, to select the fuel, we have made a comparison between thirty different fuels with hydrogen peroxide and in the end we have arrived at this. It's not as easy as oxygen and methane, there's a lot more science here.

As an engineer, was it fun to find that alchemy between fuels and hydrogen peroxide?

Yes, because it's like cooking, you try a recipe, it turns out wrong. Try something else... Although we have done few fire tests. But yes, it is fun and elegant, the fuel determines what the rest of the system will be like later, it is very elegant.

We've talked about thrust, but the other much less known metric of both aerospike and conventional rocket engines is ISP, specific thrust. Can you tell us about your engines?

Yes, I can't give you numbers, but I can tell you that this is where we have business, in specific impulse. The aerospike has 15% more integral ISP over the trajectory depending on the engines we compare thanks to the expansion ratio.

It can be simple, modular and gas generator like ours, but the specific impulse will be brutal and there the performance of simplicity is recovered.

What is your favorite engine?

It would be the XRS-2200, which is the aerospike engine par excellence.

1200px Twin Linear Aerospike XRS 2200 Engine PLW edit
Static fire test of the XRS-2200 engine, developed by NASA at the beginning of the century for the replacement of the space shuttle. Source: NASA

What if I pointed out one from today...?

The Raptor is the king of all rocket engines. It will be very difficult to achieve it because it has very good performance and is also cheap.

If you weren't in bio methalox aerospike engines where would you like to be working?

In the Raptor probably, in a full flow stage combustion. I would say that is the current challenge.

Do you think there is a better engine they are developing out there now?

Yes, a higher performance engine would be a dual expander aerospike, which we want to develop ourselves since no one is developing it. That would be the most efficient chemical engine in the world.

If we turn to other things, nuclear propulsion promises a lot, but of course, it is also more complicated, not so much for technical reasons.

Martin Morala Andres