Recently we looked at Fast Radio Bursts (FRBs) and the ongoing effort to identify their source (see Fast Radio Bursts: SETI Implications?) Publication of that piece brought a call from my friend James Benford, a plasma physicist who is CEO of Microwave Sciences. Jim noticed that the article also talked about a different kind of signal dubbed ‘perytons,’ analyzed in a 2011 paper by Burke-Spolaor and colleagues. Detected at the Parkes radio telescope, as were all but one of the FRBs, perytons remain a mystery. As described in the essay below, Jim’s recent trip to Australia gave him the opportunity to discuss the peryton question with key players in the radio astronomy community there. He has a theory about what causes these odd signals that is a bit closer to home than some of our speculations on the separate Fast Radio Burst question, and as he explains, we’ll soon know one way or another if he’s right.

by James Benford


A few weeks ago I visited Swinburne University in Melbourne Australia. I was invited there to give a public address about the controversy surrounding METI (Messaging to Extraterrestrial Intelligence). I also visited the radio astronomy group and discussed how to search for an explanation for the Perytons, dispersed swept–frequency signals. My host was Ian Morrison. I also spoke for several hours with Emily Petroff, Willem van Straten and Matthew Bailes, the head of the group.

Ian had sent me the Burke–Spolaor paper before I arrived, so I knew what the basic observations were. In our discussions, I learned that there have been other observations of Perytons and that they have the same general features: They occur from one portion of the sky (which is a clue), happen around midday, and peak in the southern hemisphere’s winter around July (another clue). The shape of the frequency versus time curve is not quite the same as true dispersion measure (DM) signals. There are kinks and dropouts in the frequency-time graph. And the shape is a bit off of the standard DM scaling, 1/f2. And they always occur at the same frequency, about 1.4 GHz.

Although they had concluded in the 2011 Burke–Spolaor paper that signals came from the horizon and were not local to the Parkes radio telescope site, they are now beginning to think that it might be emission from some local electronics. Since they knew of my knowledge of microwave sources, they asked me whether or not whether microwave ovens could be an explanation.

I had already concluded that it was a likely explanation because microwave ovens are highly nonlinear devices and can produce several frequencies. They are designed to stay on a single frequency, 2.45 GHz, but can oscillate at a variety of frequencies. If the magnetron voltage changes, other oscillation modes, with lower frequencies, will occur.

Although microwave ovens when they leave the factory have Faraday shields around them and do not radiate into the environment, over time these precautions can fail due to wear on the equipment. The primary means of preventing radiation from an oven into kitchens is redundant safety interlocks, which remove power from the magnetron if the door is opened. Microwaves generated in microwave ovens cease once the electrical power is turned off.

Parkes Visitor Centre

Image: The Parkes Visitor Centre with the 65-m dish in the background. Credit: Jim Benford.

After describing magnetrons in some detail I offered a hypothesis that leakage of microwaves was occurring from the magnetron, which generates them, and then leak from the enclosing metallic cabinet. They can easily develop separations between the metal case that the microwave magnetron is in and also in the outer case, of which the door is a part. The door is the weakest part of the shield.

Microwave ovens operate at a single frequency and are not dispersed, as their Peryton signals are. But magnetrons sometimes fail to produce a single frequency, due to mechanical disturbances or changing electrical characteristics.

Conditions change in the magnetron when it is turning on or turning off. The voltage on the cathode rises at the turn on and falls at the turn off. That changes the resonance condition and thus excites different oscillation modes, with lower frequencies. This ‘mode hopping’ may explain the observed Perytons fall in frequency.

People simply opening the door, interrupting operation, could well cause this odd radiation. Yanking the door open shuts down the voltage on the cathode of the magnetron, but the electron cloud in the resonator takes a short time to collapse because the cathode is still hot, and still can emit electrons as the voltage falls. Therefore the magnetron will continue to resonate until both the voltage goes to zero and the cathode cools down.

oven magnetron

A microwave oven doesn’t cease to radiate instantly when the electricity drops off. The timescale for the cessation is not well documented. It’s a contest between the L/R timescale of the electrical circuit and the cooling of the cathode. The Peryton signals last a few tenths of a second, which could be consistent with the fall time of the voltage and the time for the electron cloud to collapse. The frequency shifts as the voltage falls because a resonant condition, which depends on the ratio of applied voltage to insulating magnetic field (V/B), is changing. (The magnetic field doesn’t change because it’s produced by a permanent magnet.) V/B is proportional to the circulation speed of electrons in the cavity of the magnetron, which relates to the resonant frequency of the device. (For more on magnetron operation, see High Power Microwaves, Second Edition, Benford, Swegle & Schamiloglu, Taylor & Francis, 2007).

Image: Microwave magnetron from a microwave oven. Credit: Jim Benford.

I was thinking that the radio telescope at Parkes looks at a small part of the sky. But it also has side lobes through which the telescope is less sensitive to signals at other angles. The most important of these is the back lobe exactly opposite to the direction in which the telescope is pointing. This occurs because any source directly behind the dish radiating signals will diffract around the edge of the dish. This diffracted signal will arrive coherently at the receiver at the focal point of the dish.

So I inquired as to what was directly behind the dish when it was pointed at the Peryton location. They said it was the Visitor Center.

I realized at once that the Visitor Center was the microwave oven location that would have the most use and fitted all the clues. That use would occur primarily in the midday. And that in the southern hemisphere winter, more people would visit Parkes in the outback.

So I made a prediction: That they would find that the Perytons were coming from the microwave oven in the Visitor Center. I suggested the Swinburne researchers could check on that by several tests:

1) The simplest thing would be to simply remove the old oven and replace it with a new one. The Perytons would cease. But that would require taking a lot of data over time to see if they had really disappeared since Perytons are infrequent phenomena.

2) They could replace all ovens with a non-microwave cooker. That’s also a slow approach.

3) A more aggressive approach would be to rewire the oven, to defeat the safety interlocks and turn the oven on, allowing it to radiate directly into the Visitor Center. Then the signal should be quite evident and they would see a lot of Perytons. Turning the oven off and on would prove it to be the source. (Of course one would evacuate people and whoever turns the oven on would need to be behind a conducting radiation shield.)

I hear the Swinburne team is going to conduct such experiments. I hope they get a clear result. If my hypothesis is proved true, it may call into question whether the famous Lorimer Burst of 2001 was in fact a Peryton. If so, it was not extragalactic, as its large DM was taken to mean.

Perhaps we shall soon know the origin of these mysterious signals.

The Burke-Spolaor paper on perytons is “Radio Bursts with Extragalactic Spectral Characteristics Show Terrestrial Origins,” Astrophysical Journal Vol. 727, No. 1 (2011), 18 (abstract). The paper on the ‘Lorimer Burst’ is Lorimer et al., “A Bright Millisecond Radio Burst of Extragalactic Origin,” Science Vol. 318 no. 5851 (2 November 2007), pp. 777-780 (abstract).