The new GW run announced in the science forum has just begun, i.e. teh workunit generator was started. The runs short name is S6BucketLVE, and it follows and extends the previous runs S6Bucket and S6LV1. It will run for a few months and will probably be the last "all-sky" GW search running on Einstein@Home for the foreseeable future. The next run that is already being prepared will be a "directed" search, digging deeper into space at rather few sky positions.
BM
BM
Copyright © 2024 Einstein@Home. All rights reserved.
New GW run Gravitational Wave S6 LineVeto search (extended) star
)
I'm curious about the significance of 0.013682292Hz; all the new WU's I've gotten so far are offset by that amount from the 0.05Hz steps used in the last run.
Hi! This is just a WU
)
Hi!
This is just a WU naming change. The WU itself are designed to be a seamless continuation of the run into a higher frequency range.
Cheers
HB
That partially explains the
)
That partially explains the first character of bogofrequency; although not the remaining 7 or why the rename couldn't be done somewhere else in the string as IIRC was done in the past.
Is it related to the 'wings'
)
Is it related to the 'wings' that are used with the frequency steps, to provide some overlay at the band margins? So that we don't miss a signal on account of our choice of frequency slicing? IIRC they were previously 0.0125 for 0.05 steps??
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Hi! When talking about
)
Hi!
When talking about signal frequencies, one has to distinguish between:
a) the frequency we (hope to) get from a given GW source at the detector(s) on the surface of Earth
b) the intrinsic frequency of the GW source itself (or more precisely, the frequency at some reference point that we can assume to have constant radial velocity wrt. the source over the time of the observation (so no frequency modulation because of a Doppler effect), in our case we pick the Solar System Barycenter).
Both are in general different, and they are both changing over time: the GW signal from a pulsar itself will slowly change its frequency over time because that's what pulsars do: most of them slowly but steadily decrease in spin frequency, some few spin up tho, but anyway they change their spin frequency over time, and enough so that it matters for our search.
But even a constant signal would be modulated at the detector by the Doppler effect, caused by the Earth's spin around its axis, by the orbit in the Solar System and by whatever radial velocity the source may have with respect to our solar system. To deal with the Doppler effects, we look for signals like they would appear at the Solar System Barycenter (SSB), the point in the Solar System that all planets and the Sun are orbiting, as our reference point.
So, when we are looking for a GW source of frequency f (at some reference time) with a given decrease of frequency over time, we know that we have to look at a whole band of frequencies in the detector. The higher our search frequency, the broader this band has to be (Doppler modulation works as a factor applied to the intrinsic frequency).
The new names of the WUs have both types of frequencies in their name for convenience: the low end of the band of frequencies we look for in the detectors (which come sliced in 0.05 Hz chunks) and the minimum frequency of a signal at the SSB that we can hope to find in the given task.
I might have simplified one or two things in this description (I'm not a physicist) but I hope you all get the general drift :-).
Cheers
HB
The task (and workunit) names
)
The task (and workunit) names always start with the name of the (lexicographically) first data file this task needs. The (frequency domain) data file name includes the lowest frequency contained in that file. This has always been the case since we started using data files segmented by frequency; I think with an S4 run around 2007/8.
Indeed the actual analysis of that tasks starts at a somewhat higher frequency. We need some "wings" around the actual "analysis frequency" to compensate modulations of the signal (doppler modulation by the detector moving, spindown of the source). How much more data we need depends on the point in parameter space that we are looking at (and, of course, on the program used for analysis).
The "analysis frequency" previously was only visible in the command-line to the program; in the task name and thus in the result file name you could only see the data file frequency. This caused some significant confusion and trouble when uninitiated scientists looked at the results. To help them in post-processing, we now changed our workunit generator to also expose the (lowest) "analysis frequency" in the task name.
Honestly I don't know the difference between these two frequencies from scratch, I just wouldn't trust it to be constant. The workunit generator tries to create workunits of roughly identical run times. The frequency spacing grows finer with frequency, the workunit generator might, for instance, find that it can't cover 0.05Hz with one workunit anymore and split the range up in two or more. In this case you will even end up with tasks that refer to the same data file (frequency), but have a different analysis frequencies in the name.
BM
BM