For the second time ever, astronomers have found a fast radio burst ( FRB) that repeats on a regular, predictable cycle. This could mean that - at least in some cases - the wild unpredictability of mysterious deep-space FRBs could actually be a problem with our detection capabilities.

FRB 121102 is already famous for being the most active FRB discovered yet, spitting out repeated bursts several times since its discovery in 2012. It was thought that there was no rhyme or reason to it - but new analysis of those bursts has revealed a pattern.

According to a careful study of new observations and previously published ones, FRB 121102 exhibits repeated burst activity for a period of about 90 days, before going quiet for about 67 days. Then this whole 157-day cycle repeats again. If this analysis is correct, the source should have entered a new activity cycle around June 2 this month.

It's a jaw-dropping discovery, and one that could help to rule out proposed causes for these mysterious signals. But, at the same time, it's a very neat demonstration of how strange and difficult-to-identify these signals really are.

"This is an exciting result as it is only the second system where we believe we see this modulation in burst activity," explained astronomer Kaustubh Rajwade of The University of Manchester.

"Detecting a periodicity provides an important constraint on the origin of the bursts and the activity cycles could argue against a precessing neutron star."

FRBs - one of the most fascinating things space throws at us - are hugely energetic flares of radiation in the radio spectrum that last just a few milliseconds at most. In that timeframe, they can discharge as much power as hundreds of millions of Suns.

Most of them flare just once, randomly, and we have never detected them again. This makes them impossible to predict, although scientists are getting a lot better at tracing these one-off bursts to their home galaxies.

A smaller number of sources show repeat activity. This repeat activity was also thought to be random - until earlier this year, when a source called FRB 180916 was found to be repeating on a cycle. For four days, it would burst once or twice an hour, before falling quiet for 12 days. In all, its cycle is 16.35 days.

That, you may have noticed, is nearly 10 times shorter than FRB 121102's cycle. But if we assume that the two sources are similar, and that the periodicity is being caused by orbital motion, that range can be compared to known objects to narrow down what could be causing them.

"If we consider now also orbital motion to be the cause of the observed periodicity in FRB 121102, the large range in the observed periods (16-160 days) can constrain the possible binary systems," the researchers wrote in their paper.

"High-mass X-ray binaries are systems with a neutron star in an orbit with a massive O/B star. HMXBs in our Galaxy and the Small Magellanic Cloud have a large range of orbital periods, ranging from few tens to hundreds of days… On the other hand, binaries where the donor star fills the Roche lobe of the system have much shorter periods (

So far, possible explanations for these powerful signals have included neutron stars, black holes, pulsars with companion stars, imploding pulsars, a hypothetical type of star called a blitzar, a connection with gamma-ray bursts (which we now know can be caused by colliding neutron stars), magnetars emitting giant flares, and aliens (it's not aliens).

The discovery of a magnetar inside the Milky Way galaxy spitting out an extremely FRB-like outburst at the end of April is strongly suggestive that magnetars are the source of some fast radio bursts. But there is a lot of variation within the category of fast radio bursts - repeaters and non repeaters are just one example - so it's possible that there is more than one source.

It's also possible that we can't detect all activity from each FRB source - that they all repeat, but with varying strength, with the weaker bursts below our detection threshold.

And now, with the discovery of a second regular repeater, we have to consider the possibility that all repeaters are regular - we just don't have enough observation data to discern the patterns. It's a powerful argument in favour of dedicating a lot more telescope time to staring at repeating fast radio bursts.

"This exciting discovery highlights how little we know about the origin of FRBs," said physicist and astronomer Duncan Lorimer of West Virginia University. "Further observations of a larger number of FRBs will be needed in order to obtain a clearer picture about these periodic sources and elucidate their origin."

The research has been published in the Monthly Notices of the Royal Astronomical Society.