Why SETI is a waste of time – take 2

Remember SETI – the Search for Extra-Terrestrial Intelligence?

Here is reason #2 that it is a waste of time (based on physics):

Radio Frequencies and Power and all that

There are 2 basic law of nature – which cannot be changed. These are

1. When you transmit radio frequencies, the received power level falls with distance squared.

2. All technology used to receive radio signals is imperfect, and this manifests (at the very least) as noise.

Taking the first point.

If you transmit a signal with a given power, then when you receive that signal from a distance of 1 metre you will receive a different power. The difference is due to the types of antennas you use, and the distance. Using special antennas helps – but you can’t get out more than you put in (if that were true then perpetual motion machines would also work, really well!).

Lets assume you transmit X, and receive Y, where Y less than X.

Mathematically, Y = c X / (d ^ 2)

Where c = a constant (less than 1), and d = distance (in metres).

Now, increase the distance to 2 metres. You will receive a power less than Y, and it should be about 1/4 as much. Increase to 4 metres, you get 1/16 as much received power and so on.

So… Assume that the little green men are transmitting the sort of powers that we transmit for radio or TV transmissions – maybe a few thousand watts, or a few tens of thousands.

Then take into account the vast distances, it does not take much to work out that the power we COULD receive (if everything worked perfectly) is very very small.

Mathematically, lets assume an alien transmits 10,000 Watts, and that there are no losses, and they are 100 light years away (which is pretty close really).

The most wildly optimistic receive power is 10000 / [( 100 light years, in metres) ^2] Watts (because this does not take into account the frequency dependant effects).

(1 light year = distance for light to travel in a year = 3 x 10^8 [metres/sec] x 365 [hours / day] x 24 [hours / day] x 60 [hours / minute] x 60 [minutes / sec] = 6.46 x 10^15 metres)

So, the best possible receive power is roughly 10000 / (9 x 10^35) = 10^-32 watts.

In engineering speak, that’s -290 dBm.

This is a number so mind bogglingly small that it needs a damn good receiver to pick it up.

Taking the second point

When we build a receiver for a radio signal, it has noise. We have all heard noise – it’s that hissing “static” sound in a radio that’s not tuned on. This is an extreme example.

If we ignore all other sources of noise and concentrate only on Thermal noise (which is often a big contributor anyhow), the amount of thermal noise is given by:

N = kTB (in Watts)

N = noise power,
k = Boltzmann’s constant, 1.38 x 10^-23
T = temperature in Kelvins
B = Bandwidth in Hertz

The important point about noise is that it is additive – so it is added to the signal we receive, and to detect a signal we need lots of it, so the amount of noise added needs to be small compared to the signal.

If we fix the bandwidth at 1 Hz (which is impossibly small) we have a noise power at room temperature (approx 290 Kelvin) of 4 x 10^-21 Watts (or -174 dBm in engineering speak).

It does not take much to see that if the noise power in a ridiculously small bandwidth is MUCH LARGER than the best possible receive signal power, then we are not going to detect a signal.

We can make things better by reducing the temperature of the receiver. Let’s cool the receiver with something jolly flash, and get temperature down to (say) 20 Kelvins – which is damn cold.

This gives a noise power of 2.76 x 10^-22 watts, or -185 dBm. Still does not cut it. And that is in a 1 Hz bandwidth – still impossibly small.

The only way that we can see the transmissions of the little green men is if they have a DAMN BIG antenna with a DAMN BIG transmitter, POINTED RIGHT AT US.

And even then we would have to be looking at the same frequency, and at the same time.

Those who wish to lecture me about antenna gains and correct my maths with a more exact RF link budget – go right ahead – it will get closer but the point still stands – both sides need antenna gains of > 50 dB to come even remotely close. We can be pretty sure that the little green men will not have a focussed antenna pointing at us, so there won’t be 50 dB of antenna gain on their side. We might have it on ours – maybe even more. Still does not come close.

So… waste of time. It’s trying to break the laws of physics.



Isn’t noise chaotic? And wouldnt signals be somewhat less random? doesnt that offer the seti team some chance of detecting something if it exist?


Comment by Duncan | August 30th, 2005 9:51 am | Permalink


just on the physics side Duncan is correct. I am not a SETI proponent as I still think the odds are stacked mindbogglingly against us but the physics can be overcome. Proof are the Voyager probes that are still able to transmit data back without hugh antenna gains on the spacecraft and they are something like a lightyear or so away from us by now. The trick is to use spread-spectrum techniques, very high gain antennae here on earth and as you say crygenically cooled low-noise amplifiers to get a high G/T. The signal processing then takes care of the rest.


Comment by Ulrich | August 30th, 2005 10:13 am | Permalink

Ulrich is correct – you CAN establish a communication link by using very clever techniques.

What you cannot do (without prior knowledge) is detect a signal with a power that is much lower than the noise power. Spread-spectrum uses correlation to do a few neat tricks but you MUST know what was transmitted.

When the alieans transmit an unknown signal of an unknown modulation in an unknown bandwidth, you CANNOT use a correlating technique.

And because the noise power is so much higher than the signal power (by more than about 1000000 times) the chances of detection are mind-bogglingly small, and the chance of demodulation (extraction of information) is even more mind bogglingly small.

Comment by Wally | August 30th, 2005 8:25 pm | Permalink

You guys are Freaks, I have no idea what you’re talking about, physics is a mystery to me :)

If noise is random and signals that are played over and over again like a “we’re here, we’re here, we’re here” message are not random cant looking for “noise” that isnt random potentially work? even if the intelligence that it contains couldnt be decoded?


Comment by Duncan | August 31st, 2005 9:03 am | Permalink


Comment by Adam | August 31st, 2005 11:40 am | Permalink

Well, it turns out that the more noise-like a signal is the more efficient your comm’s link becomes. So sending “we’re here…” repetitively is actually a very inefficient transmission (i.e. short distance in our context). Totally scrambling the message to make it virtually the same as noise (but still recoverable if you know the scrambling sequence) lets you achieve much longer distances. You can recover such signals that are way below the noise floor. In fact any civilisation will develop more and more noise-like modulation techniques (i.e. CDMA, 802.11 in all its incarnations, Bluetooth, ZigBee etc) as their technology progresses simply because of the physics. That’s why I think SETI is pointless because it becomes increasingly harder to detect these signals and finding the scrambling sequence. There are techniques to do that (essentially exploiting the ever tinier amount of non-randomness of the scrambled “we’re here” message) but they take an incredible amount of compute power (hence the idea of SETI @ home).

Comment by Ulrich | August 31st, 2005 4:34 pm | Permalink

SETIs goal has never been to find a bonafide signal, if you look at what the seti@home project is trying to achieve for instance is to look for the incidental effects of transmitted signals in order to purely detect (with a reasonable amount of certainty) alien communications. The SETI project hasn’t, any probably won’t worry about even the concept of “decyphering” any signal, until it knows theres one there. Although I believe its unlikely (as in trillion to 1), I would never discount it, purely based on physics/statistics alone.

Comment by Paul Cleland | September 2nd, 2005 8:02 am | Permalink

I fail to see how you can detect a signal created by an intelligent being without doing some level of basic decoding on it.

And to decode a signal you need signal to noise being high.

For example: to detect a carrier you need it to stand out above the noise. (There is NO information content in just a carrier). But suppose a wide-band FM signal is transmitted? In that case detecting a carrier will be very difficult with a receiver that has a narrower bandwidth than the transmitter….

The “incidental effects of transmitted signals” sounds like PR bullshit to me.

Comment by Wally | September 2nd, 2005 1:33 pm | Permalink

Leave a Comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Live Comment Preview

Comment by Somebody

Powered by WordPress 2.8    Rendered in 29 queries and 0.223 seconds.    CleanBreeze Theme