A laymen's guide Gravity, Gravity Waves, LIGO, and how it all works

StarCharter
StarCharter
Joined: 19 Feb 05
Posts: 59
Credit: 641,079
RAC: 0
Topic 187989

I've seen more than one post asking about how this project works. There's a decent source of information for lay folks posted at The American Museum of Natural History that discusses Newton's ideas, Einstein's Theories, Graity and what it's all about. There are some decent videos, a virtual tour of LIGO, and a nifty simulator that lets you fire the laser and search for your very own gravity wave. Click HERE if you're interested.

There are two secrets to life: 1) Don't tell everything you know...

Dennis
Dennis
Joined: 19 Feb 05
Posts: 51
Credit: 4,459
RAC: 0

A laymen's guide Gravity, Gravity Waves, LIGO, and how it all wo

> I've seen more than one post asking about how this project works. There's a
> decent source of information for lay folks posted at The American Museum of
> Natural History that discusses Newton's ideas, Einstein's Theories, Graity and
> what it's all about. There are some decent videos, a virtual tour of LIGO, and
> a nifty simulator that lets you fire the laser and search for your very own
> gravity wave. Click href="https://einsteinathome.org/%3Ca%20href%3D"http://sciencebulletins.amnh.org/astro/f/gravity.20041101/">http://sciencebulletins.amnh.org/astro/f/gravity.20041101/"> HERE [/url]
> if you're interested.
>
>

Yes, I can easily find things like what a pulsar is, what are sources for gravity, how their machines work, what General Relativity predicts, how interferometers work. That is just popular press items. Although I must admitt I totally enjoyed listen to: http://web.mit.edu/sahughes/www/sounds.html
But the real question still remains - What is not what the pulsars or machines are doing but what is being done in our WU's?
It is not the Why are we looking, it is the How are we calculating?

For example, when I turn in may WU that looks at 511 Hz signals, is it verified by say.... someone elses WU because they have the same period, or phase, or what?
Are they looking for single source pulsars or do our WU's look only for binary systems?
I can easily look up what nature is doing, what I want to know is what are the WU's doing as they sit there crunching on my machine.

StarCharter
StarCharter
Joined: 19 Feb 05
Posts: 59
Credit: 641,079
RAC: 0

Dennis, I'm often not good

Dennis,
I'm often not good with words and many times get misunderstood. I'm saying this so that you won't think that this post is in any way an attack on you.

I purposefuly gave the post a heading of LAYMAN's guide to GRAVITY, GRAVITY WAVES, LIGO so that those who had questions on other items could ignore it. I did not include einstein@home in the heading because I have no definite, concrete knowledge of what's happening during the WU processing. If you already know the information in my previous post, it's likely that the post wasn't for you. I was posting it because I continue to see questions from those who have little knowledge of what's going on in the basic science behind what the project is addressing.

In actual fact, to the best of my knowledge, we aren't looking for pulsars, binary or otherwise, at all. We're looking for interference differences in data taken from (I think) Hanford. The data starts at a given RA and DEC and then steps to the next data by varying the RA slightly. Each data point (I'm assuming) consists of the data from the intereferometer. It's either 0, as in there was no difference, or some other number, probably the amount of distortion caused by the intereference of the two returning beams. Each WU (I believe) contains only data at a given frequency. Different WUs cover the same "sky spread" at different frequencies and, very likely, at different times. I'm assuming that the graphic representation of the progress of processing is accurate. If you watch it, it starts at a given RA and DEC and steps through, changing the RA while keeping the DEC constant, and then changes the DEC and resets the RA. If we were looking only at pulsars, it would seem that the graphic display would quickly skip thru the 150-250 known pulsars and not scan incrementally. Based on the data (ie, presence or absence of an intereference pattern), we wouldn't see a new pulsar unless we saw the gravity wave first.

I say all of this as an assumption because I have yet to see, anywhere, a decent writup for the steps involved in the processing of WUs. Not in the unfinished FAQS, not on the homepage for the project, and not on the LIGO website. Either Bruce Allen et al are up to their eyepieces in trying to keep things rolling or they don't, for some reason, want us to know. Either way is ok with me.

What little I know of the inner workings of the program come from reading the bazillion or so back posts. In particular, Ben Owen, a physics professor at Penn State, explained much of this back before. For example, see this thread. I have no knowledge of whether Ben Owen is attached to the project or just one hell of a nice guy (I believe those two cases are not mutually exclusive...)

As to someone else verifying a WU that you crunched, the answer is yes, At least 3 other computers have (or will have) processed the same data. I'm quite certain that, in the case of a positive result, many more computers than 4 will look at it. I'm assuming that, in the case of a positive result, most of the astronomical world will look with every form of detection possible at what caused the perturbation.

I hope I answered your questions without adding to the confusion. Many of the very old posts in the forum contain more information. I'm assuming (again) that things were a bit quieter and saner before the project opened for full-bore participation.

All of this is just my opinion. I may be wrong.

Edited Afterwards: It appears that Ben Owen is a member of the LIGO Collobration AND one hell of a nice guy...

There are two secrets to life: 1) Don't tell everything you know...

StarCharter
StarCharter
Joined: 19 Feb 05
Posts: 59
Credit: 641,079
RAC: 0

Followup: The very front

Followup:

The very front page of the project website says we're looking for pulsars. So disregard everything I said.
Except for the part about Ben Own being a nice guy.

There are two secrets to life: 1) Don't tell everything you know...

Ben Owen
Ben Owen
Joined: 21 Dec 04
Posts: 117
Credit: 33,972,457
RAC: 7,309

StarCharter, Thanks for

StarCharter,

Thanks for the "nice guy" remarks. I'll take the Fifth Amendment on whether or not they are true.

Yes, it does help to read the front page. However, we know that it would also be nice if we had some more explanation handy. There are plans to do that; for example you will see some questions in the FAQ that is being constructed. But the main priority at the moment is making sure the code actually does what it's supposed to. (It basically does, although if you read the help desk you'll see there are still plenty of bugs to clean up.) In the meantime, Ian Jones and I can answer a few questions here and there.

That museum link was nice in general. Here's a little more specific Einstein@Home info that addresses some of the questions you and Dennis asked.

Einstein@Home isn't really looking for an interference pattern between different points, but that leads to an explanation of interferometry which I think I'll postpone for another day.

The LIGO Science Collaboration is implementing several pulsar searches, but the one Einstein@Home is running is the "all-sky pulsar search" as you will see at the top of every WU page. This is not a search for the known pulsars whose locations you see on the screensaver. This is a systematic search of the sky, one location at a time, for any periodic gravitational wave coming from that location. It has to be done for each location because the frequency shifts (Doppler shifts) due to the Earth's motion are different for different sky locations. There are additional Doppler shifts for pulsars in binaries, but the current application is only looking for isolated pulsars.

Most of your CPU cycles are going into Fourier transforms. If you don't know, a Fourier transform is a way of looking at a time series as a sum of sine waves at different frequencies. Pulsar signals should be nearly sinusoidal after the Doppler shifts are taken out, so Fourier transforms pick them up pretty easily. Fourier transforms are numerically pretty efficient, but there are an awful lot of them to do in an all-sky search. That is why Einstein@Home is doing this particular search, and not for example the searches for known pulsars which can be done pretty quickly on a single computer.

Later on Einstein@Home might do some other searches, but the consensus was that this was the best fit (at least for now) because (1) it is the most expensive in CPU cycles, and thus the best use of the enormous power you all are donating, and (2) people would probably be more excited about going after something brand spanking new than a pulsar that's been seen in radio, x-rays, etc for years. I think also that (3) we know where the radio pulsars etc are, but a previously unknown one (whose radio beam isn't pointed towards Earth) might happen to be much closer and thus a much stronger source of gravitational waves. We set our long-term goals by the sources we already know, but we cross our fingers and hope for a pleasant surprise.

Hope this helps,
Ben

Dennis
Dennis
Joined: 19 Feb 05
Posts: 51
Credit: 4,459
RAC: 0

> StarCharter, > > Thanks

Message 5232 in response to message 5231

> StarCharter,
>
> Thanks for the "nice guy" remarks. I'll take the Fifth Amendment on whether or
> not they are true.
>
> Yes, it does help to read the front page. However, we know that it would also
> be nice if we had some more explanation handy. There are plans to do that; for
> example you will see some questions in the FAQ that is being constructed. But
> the main priority at the moment is making sure the code actually does what
> it's supposed to. (It basically does, although if you read the help desk
> you'll see there are still plenty of bugs to clean up.) In the meantime, Ian
> Jones and I can answer a few questions here and there.
>
> That museum link was nice in general. Here's a little more specific
> Einstein@Home info that addresses some of the questions you and Dennis asked.
>
> Einstein@Home isn't really looking for an interference pattern between
> different points, but that leads to an explanation of interferometry which I
> think I'll postpone for another day.
>
> The LIGO Science Collaboration is implementing several pulsar searches, but
> the one Einstein@Home is running is the "all-sky pulsar search" as you will
> see at the top of every WU page. This is not a search for the known pulsars
> whose locations you see on the screensaver. This is a systematic search of the
> sky, one location at a time, for any periodic gravitational wave coming from
> that location. It has to be done for each location because the frequency
> shifts (Doppler shifts) due to the Earth's motion are different for different
> sky locations. There are additional Doppler shifts for pulsars in binaries,
> but the current application is only looking for isolated pulsars.
>
> Most of your CPU cycles are going into Fourier transforms. If you don't know,
> a Fourier transform is a way of looking at a time series as a sum of sine
> waves at different frequencies. Pulsar signals should be nearly sinusoidal
> after the Doppler shifts are taken out, so Fourier transforms pick them up
> pretty easily. Fourier transforms are numerically pretty efficient, but there
> are an awful lot of them to do in an all-sky search. That is why Einstein@Home
> is doing this particular search, and not for example the searches for known
> pulsars which can be done pretty quickly on a single computer.
>
> Later on Einstein@Home might do some other searches, but the consensus was
> that this was the best fit (at least for now) because (1) it is the most
> expensive in CPU cycles, and thus the best use of the enormous power you all
> are donating, and (2) people would probably be more excited about going after
> something brand spanking new than a pulsar that's been seen in radio, x-rays,
> etc for years. I think also that (3) we know where the radio pulsars etc are,
> but a previously unknown one (whose radio beam isn't pointed towards Earth)
> might happen to be much closer and thus a much stronger source of
> gravitational waves. We set our long-term goals by the sources we already
> know, but we cross our fingers and hope for a pleasant surprise.
>
> Hope this helps,
> Ben
>

Thanks, that gives the kind of information I have been hoping to hear about "how it all works".
we are looking over the whole sky for isolated pulsars, our WU are doing transforms.

Thanks,

Dennis

marj999
marj999
Joined: 8 Nov 04
Posts: 45
Credit: 17,474
RAC: 0

Thanks Ben, thats brilliant.

Thanks Ben, thats brilliant. I think thats the first clear and concise explanation I've seen of precisely what we are actually doing :-)

Marj

StarCharter
StarCharter
Joined: 19 Feb 05
Posts: 59
Credit: 641,079
RAC: 0

> Einstein@Home isn't really

Message 5234 in response to message 5231

> Einstein@Home isn't really looking for an interference pattern between
> different points, but that leads to an explanation of interferometry which I
> think I'll postpone for another day.
>
> The LIGO Science Collaboration is implementing several pulsar searches, but
> the one Einstein@Home is running is the "all-sky pulsar search" as you will
> see at the top of every WU page. This is not a search for the known pulsars
> whose locations you see on the screensaver. This is a systematic search of the
> sky, one location at a time, for any periodic gravitational wave coming from
> that location. It has to be done for each location because the frequency
> shifts (Doppler shifts) due to the Earth's motion are different for different
> sky locations. There are additional Doppler shifts for pulsars in binaries,
> but the current application is only looking for isolated pulsars.
>
> Most of your CPU cycles are going into Fourier transforms. If you don't know,
> a Fourier transform is a way of looking at a time series as a sum of sine
> waves at different frequencies. Pulsar signals should be nearly sinusoidal
> after the Doppler shifts are taken out, so Fourier transforms pick them up
> pretty easily. Fourier transforms are numerically pretty efficient, but there
> are an awful lot of them to do in an all-sky search. That is why Einstein@Home
> is doing this particular search, and not for example the searches for known
> pulsars which can be done pretty quickly on a single computer.
>
> Later on Einstein@Home might do some other searches, but the consensus was
> that this was the best fit (at least for now) because (1) it is the most
> expensive in CPU cycles, and thus the best use of the enormous power you all
> are donating, and (2) people would probably be more excited about going after
> something brand spanking new than a pulsar that's been seen in radio, x-rays,
> etc for years. I think also that (3) we know where the radio pulsars etc are,
> but a previously unknown one (whose radio beam isn't pointed towards Earth)
> might happen to be much closer and thus a much stronger source of
> gravitational waves. We set our long-term goals by the sources we already
> know, but we cross our fingers and hope for a pleasant surprise.
>
> Hope this helps,
> Ben
>

Ben,
Thanks so much for responding. This actually helps a lot. And, yes, repetitive transforms make a heck of a lot of sense for a BOINC project. I think a lot of my confusion resulted from thinking about the nature of the detector. I couldn't see how it could detect a pulse of any kind, rather that it would detect a theoretical effect from an object that would most likely be a pulsar. I'm not sure how this necessitates that the object be a pulsar. There are other, perhaps not as likely, potential sources of gravity waves which might trigger a detection event. Your statement "This is a systematic search of the sky, one location at a time, for any periodic gravitational wave coming from that location" seems to me to be a better assesment of the goals of the project than to say that we're looking for pulsars. I understand that the organizers were looking for some verbage that would encourage folks to participate but it seems a bit "pop-sci" to me.

I'm not an astrophysicist nor have I played one on TV. I'm merely a diletante with too much time on my hands and too much curosity to watch TV.

If I had one complaint about the project, it would be that the organizers don't seem to realize that a project like this will attract a lot of "science nuts" who want to know more. I could care less about credits or scores. I have no problem being a small cog. But, I, and others apparently, have questions. If there's an "ask a physicist" page on the web, I haven't been able to find it. If you or Ian could drop in time to time (hey, if juggling semester time constraints was easy, orangutans would do it...) to answer a question. On the other hand, I certainly appreciate the demands on the team members.

One of my problems is that my fastest source of information is the web and it is far from reliable. For example, this page about pulsars on Wikipedia says that the discovery of PSR B1913+16, confirmed general relativity and proved the existence of gravitational waves. I'm not sure that's true. I do wish that you or Ian could convince a student to go over to Wikipedia and straighten that out.

To me, "nice guy" is someone who takes the time to impart knowledge to others, knowing full well that it may not be the best use of their time. Many of us have not had the good fortune to persue an academic career. Those who share their wealth are indeed nice guys.

There are two secrets to life: 1) Don't tell everything you know...

StarCharter
StarCharter
Joined: 19 Feb 05
Posts: 59
Credit: 641,079
RAC: 0

> sky, one location at a

Message 5235 in response to message 5231

> sky, one location at a time, for any periodic gravitational wave coming from

AARGH!!!
Periodic! yep, no doubt. We're looking for pulsars.

There are two secrets to life: 1) Don't tell everything you know...

deacon blues
deacon blues
Joined: 9 Feb 05
Posts: 11
Credit: 34,215
RAC: 0

Hi all - I'd like someone to

Hi all - I'd like someone to tell me if I have this right: quantum mechanics predicts that over vast periods of time all massive objects eventually loose some of their mass in the form of gravitational radiation that propagates through spacetime as waves. These waves are not electromagnetic forces but rather are ripples in the fabric of spacetime itself. Gravitational waves are not the same as an object's gravitational field. Rapidly rotating neutron stars (pulsars) are presumed to shed gravitational radiation in the form of periodic pulses. LIGO and GEO are not looking for "gravitons" but rather are set to detect these periodic pulses as minute distortions in spacetime.

So what we are looking for, as our computers do Fourier transforms to correct for the Doppler effect, is a tiny hiccup in the more or less uniform expansion rate of the universe. Is that somewhere close to the truth?

debugas
debugas
Joined: 11 Nov 04
Posts: 170
Credit: 77,331
RAC: 0

> LIGO and GEO are not

Message 5237 in response to message 5236

> LIGO and GEO are not looking for "gravitons" but rather are
> set to detect these periodic pulses as minute distortions in spacetime.

That's exactly how i understood it myself, but i am not a physisist nor associated with E@H team in anyway, so i can't confirm that.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.