Failure to launch – the satellite industry problem

In this two-part feature Cosmos space correspondent Jamie Seidel looks at the satellite industry where 50% failure rate is normal and tomorrow he explains how to get your satellite into orbit. You can read part two here.

The space race has changed. No longer is it the exclusive domain of big-budget “Formula 1” teams. Our orbital highways are now seeing the equivalent of family sedans and delivery vans. And even e-scooters.

“Space is finally becoming accessible,” says Dr Matthew Tetlow, CEO and founder of Adelaide-based Inovor Technologies. “It’s no longer just the domain of nations and militaries.”

Nor is everything that is launched a single massive piece of infrastructure costing hundreds of millions of dollars. Now, dozens of 2kg lunchbox-size devices can be boosted into orbit at once, each for a price tag in the low millions.

With this ease of access comes new possibilities.

Space is finally becoming accessible. … It’s no longer just the domain of nations and militaries.

Dr Matthew Tetlow, CEO and founder of Inovor Technologies

It’s not just about critical national infrastructure anymore. And the economies of scale are being applied to space for the first time by the likes of SpaceX and its Starlink constellations.

Others want to try something different – to test new technologies, ideas and systems – and to do that, they’re willing to take risks.

“Being at the low end of the space scale, cubesats are where universities and startups begin playing,” the Tetlow, who is a satellite manufacturer, explains. “And neither have very deep pockets. So many of these projects are run on the smell of an oily rag.”

And that’s all part of the risk-reward equation behind statistics showing up to 50 per cent of cubesats launched into low-Earth orbit break down.

“Success or failure is very context-dependent,” says Dr Andrew Barton, Research Program Manager of SmartSat CRC. “If it’s a highly experimental new system, you’re going to have to accept a higher potential failure rate.”

But dramatic reductions in risk can be made through investing in resilient and redundant technology and intensive testing. If you can afford it, and if it’s economically viable.

“Take SpaceX – they rely on having routine access to space,” Tetlow says. “So they can spread the risk over multiple satellites. Therefore, the overall business risk is low even if the technical risk on a per satellite basis is still significant.”

Quantum leap

For universities and startups, a dysfunctional satellite doesn’t mean failure.

Obviously, we’ve tested everything the satellites can do ad nauseam. But there are just one or two things that can’t be tested until they’re actually in orbit.

Dr Matthew Tetlow

“Success is a difficult thing to measure in that arena,” says Tetlow. As an academic at Adelaide University, he led about 50 students and a dozen staff in building an experimental cubesat for the European Space Agency in 2017. “While the QB50 project was only marginally successful in itself, the whole program was extremely successful because it’s basically what spawned INOVOR today.”

Cubesats are designed to fit modular systems attached to commercial rockets that have space and weight capacity when launching their primary payload. Such “piggyback” schemes spread the cost over a multitude of customers to make launch affordable.

“Universities do it primarily for an educational purpose,” adds Barton. “And the fact is, most of that learning happens on the ground before it’s launched. They can afford a higher failure rate as they’re still getting most of their educational outputs.”

INOVOR Technologies is one of a handful of Australian businesses partnering with the Federal Government’s satellite cooperative research centre (SmartSat CRC) in nano, cube and small satellite research. 

These, however, have defence and commercial goals. Not just experimental ones.

“Once you get to companies like us and Myriota, we must spend millions of dollars on our test assurance programs,” says Tetlow. “That’s the only way to repeatedly get mission success.”

The first INOVOR-designed and built spacecraft will launch this November.

The SpIRIT 6U (six cube unit) cubesat project is lead by the University of Melbourne. It’s been supported by the Australian Space Agency. It will carry two astrophysics payloads – including a gamma-ray detector from the Italian Space Agency. And it will test an Australian innovation, the Neumann Space metal-fuel thruster system.

Its second shot – SA Space Services Mission Kanyini 6U cubesat scheduled next year – is also the first satellite mission to be launched by an Australian State Government.

Proving ground

Kanyini (which roughly translates to respect the land’) represents a multitude of firsts. Not least among them offering near real-time environmental monitoring services to the South Australian government.

“I think it’s fair to say they (the SA government) want to be a part of this small satellite revolution,” says Barton. “They’re taking advantage of that drop in costs to get some very specific but useful low Earth observation and communication services that they couldn’t have dreamed about even 20 years ago.”

Many of SpIRIT’s and Kanyini’s components were designed and assembled in Australia along with the software bringing it all together which was coded here.

“I’m pretty comfortable with how it’s gone,” says Tetlow. “Obviously, we’ve tested everything the satellites can do ad nauseam. But there are just one or two things that can’t be tested until they’re actually in orbit.”

In the case of Kanyini, the verification process involved building an exact replica – neatly dissected and spaced out on a workbench as a “flat-sat.” That way, each component and connection could be easily monitored, inspected, modified or replaced during testing.

“When you’re talking about small satellites, it would be more expensive to build all the simulation models needed to create a digital twin than the real thing,” says Tetlow.

You could have a high-energy particle punch through your satellite 30 seconds after you turn it on and take out the flight computer – and there’s nothing you can do about it.

Dr Matthew Tetlow

“But we’ve put probes on every part of every circuit to work out what’s going on. We’ve picked up all sorts of niggles in software and hardware because you can really interrogate with very fine detail as you try to isolate a problem. That’s invaluable. And it gives us confidence we’ve tested everything possible.”

Orbital autopsies

Radiation. Vibration. Heat. Cold.

“All those things contribute to a satellite’s death,” says Tetlow. “We select components that can operate in the expected radiation and temperature ranges. But of course, this is all statistical. You could have a high-energy particle punch through your satellite 30 seconds after you turn it on and take out the flight computer – and there’s nothing you can do about it.”

“All you can do is hope none hit you. The odds are low. But if you do get hit, it could destroy a whole system. And this is where budgets come in. It costs more to put in redundant systems. That’s the risk-reward trade-off.”

Operational cubesats are built with life expectancies of between six months to five years.

“It comes down to the amount of testing that was done and what quality of components you invested in, that kind of thing,” Tetlow adds.

Many small satellites – generally under 350kg – use “off-the-shelf” commercial electronics to keep costs low. Others in the same weight class may spend millions more on carefully tailored and hardened custom components.

But Barton says low-Earth orbit is relatively safe. Relatively.

“There’s no way anyone would use commercial-grade electronics in much higher geostationary orbits and expect to get away with it,” he quips.

Failed communications are the most common symptom of small satellite losses. 

“Those satellites that are dead on arrival in orbit are the scariest ones,” he says. “In many instances, we never determine what went wrong. And that means we’ve lost a chance to learn.”

Tomorrow on My Cosmos: Australia’s satellite industry on steep learning curve