Beta Perseus

RJN · Post by **RJN** » Mon Sep 20, 2004 3:09 am

Earlier I had found that photometic values for MK for 040918 and 040917 were generally without significant systematic errors. So I went looking for a variable star to see between those two days. Beta Perseus is a known variable that was seen both days at MK.

After taking the ratio of Beta Per data on 040917 to Beta Per on 040918 I believe that I have found a light curve that indicates Beta Per coming out of an eclipse. Here is the plot:

This was done rather quickly and not checked, as yet. Using data for these nights, it is relatively simple to check stars with automaticaly created photometry tables for variability.

- RJN

Vic Muzzin · Post by **Vic Muzzin** » Wed Sep 22, 2004 2:29 am

Beta Per was at full eclipse at 3265.72657

http://www.as.wsp.krakow.pl/minicalc/PERBETA.HTM

Vic Muzzin · Post by **Vic Muzzin** » Thu Sep 23, 2004 3:49 pm

Just out of curiosity I wanted to see what our "old" method of taking ratios of two stars would show for this same night (9/17/2004).

It is quite clear that Dr. Nemiroffs latest approach works much better. This chart may include some bad images (clouds, etc), but even if those were taken out the graph would never be as clear as the graph posted by Dr. Nemiroff.

RJN · Post by **RJN** » Thu Sep 23, 2004 6:38 pm

Hi Vic,
Was this for MK data on that night?
- RJN

Vic Muzzin · Post by **Vic Muzzin** » Thu Sep 23, 2004 9:28 pm

Was this for MK data on that night?

Yes

Vic Muzzin · Post by **Vic Muzzin** » Thu Oct 14, 2004 2:56 pm

Vic Muzzin · Post by **Vic Muzzin** » Thu Oct 14, 2004 3:14 pm

Vic Muzzin · Post by **Vic Muzzin** » Thu Oct 14, 2004 3:30 pm

Vic Muzzin · Post by **Vic Muzzin** » Sat Oct 16, 2004 12:26 am

In search of a way to study Beta Perseus, I have tried a few things.

First, I decided to use comparison stars to attempt to correct for image deviations due to clouds, low altitudes, and any other weird occurences in the data. I collected photometry for Bet Per from 10/09/04 to 10/14/04. I also collected all the available files for the Alp, Gam, Eps, and Zet stars of Perseus. For each of the comparison stars I took an average C5-B using all available data. I then took C5-B for each image and divided it by the average for the star. After doing this for each of the comparison stars I took an average of the results and divided C5-B for Bet Per by this number.

This is the graph

Bet Per is known to eclipse at
JD = 2452500.150 + E x 2.867328

The eclipse for E = 275
Occured at 2453288.6652
E = 276 at 2453291.5325
These points are marked on the above graph in red.

The eclipse is supposed to last for "several hours" (anyone have a number for me?)
For epoch 276 the HL CONCAM did not see Bet Per until more than six hours after the minimum.
For 275 the CONCAM first picked up Bet Per a little over three hours after the minimum, I suspect the show was mostly over except for maybe the very end.

So, as seen in the graph, my correction attempts failed to bring low altitude data to a standard level.

Next, I decided to use the data to calculate an expected value for Bet Per (C5-B) for each integer of altitude, and use this for comparison on future nights.

When compared to these values, October 15th looks like this

For reference, the starting altitude was 21 degrees, the max was 70 and it ended at 41.

This graph did not come out as smooth as I had hoped, however this method may be useful in spotting an eclipse. I believe after a couple more cycles Bet Per will eclipse while seen by the HL CONCAM.

In the meantime I am looking to see if any of the other CONCAMS saw the most recent eclipses.

Vic Muzzin · Post by **Vic Muzzin** » Sun Oct 17, 2004 4:14 am

Same method as last graph. Data from HL.
There should have been an eclipse at 2453294.39986, which was inbetween this graph and the last.

Vic Muzzin · Post by **Vic Muzzin** » Sun Oct 17, 2004 5:09 pm

Here is October 17th, this graph was made the same way as the last ones, however I have recalculated my averages to include all data HL has collected on Bet Per (~2000 images).

The next eclipse should occur at 2453297.26719
I am hoping the HL CONCAM will see the very begining of it tonight.

Vic Muzzin · Post by **Vic Muzzin** » Sun Oct 17, 2004 5:57 pm

October 10th revisited.

On October 10th the HL CONCAM picked up bet Per about 3 hours after the primary minimum.
I decided to go back and look at that day.

I believe this is the type of behavior I can expect to see within 3 hours of an eclipse. (i.e. a magnitude change of around -.3 to -.1)

I expect the HL CONCAM to see Bet Per until about 2.5 hours before the eclipse tonight. I am hoping to see a curve showing the progression towards minimum.

Vic Muzzin · Post by **Vic Muzzin** » Sun Oct 17, 2004 6:38 pm

Did everyone go on holiday this weekend and not tell me? Ok, you will have to go back a few posts to see what I am up to.

Going back even farther I found that the HL CONCAM started collecting data for Bet Per a few minutes after the minimum at 2453285.797874 (October 7th)

Here is the graph

Vic Muzzin · Post by **Vic Muzzin** » Sun Oct 17, 2004 6:58 pm

On October 4th, there are only 13 data points. The first point is just less than 2 hours after the minimum.

These are all from 20 sec exposures. I have made a seperate baseline table for 20 second exposures. 180 sec and 20 sec exposures can both be used in these graphs will neglible differences. The graph for Oct 7 was about half of each. The change between exposure lengths can easily be seen in the previous graphs by noting the gap in the chart where the CONCAM was switching exposure times.

Vic Muzzin · Post by **Vic Muzzin** » Sun Oct 17, 2004 7:17 pm

Minimum was at 2453277.19589

Once again, all 20 sec exposures.

tilvi · Post by **tilvi** » Mon Oct 18, 2004 2:08 am

Vic,
Thats a good base. I think from these data sets, we have now enough confidence , atleat for HL and MK stations, to do systematic photometry study of stars.

Vic Muzzin · Post by **Vic Muzzin** » Mon Oct 18, 2004 1:06 pm

It looks like it is a sporadically cloudy night in HL tonight. However it is currently clear, and Bet Per is being tracked perfectly. It should eclipse at 2453297.26719, about five and a half hours from now, unfortunately the camera is going to shut off about three hours before minimum. Past results indicate there should be some drop off before the camera goes off, I expect about -.3 magnitude change. Right now Bet Per is still going strong.

jbruce · Post by **jbruce** » Mon Oct 18, 2004 7:49 pm

Hello, just joining this thread....

can I get a quick summary of the specific data redution steps...that is, how delta mag is computed?

dark subtration? flat fields? extinction coefficients? differential photometry, I assume? comparison names, aperture sizes, etc. mostly I'm wondering how extinction is being addressed in the process of computing delta m's.

Thanks for bringing a new member up to speed.

Bruce Rafert

Vic Muzzin · Post by **Vic Muzzin** » Mon Oct 18, 2004 9:51 pm

Hello Dr. Rafert,
Of course that is always the question with CONCAM data, I have read some papers on atmospheric extinction and it is actually the current topic of study in my Atmospheric Physics class. It has, however, been my expierience that none of the equations hold for correcting CONCAM images. The only law CONCAM data seems to follow is the law of averages, or a tendency to repeat what it has done in the past.

To use this tendency, I have taken all the data collected by the HL CONCAM and constructed a baseline chart of counts versus altitude for Bet Per, made up of an average of all counts ever taken at each degree of altitude (actually two charts, one for 180 sec exposures and one for 20). To examine a night I have been dividing each data point by the average count for the corrresponding altitude. It seems to work, but I can't help feeling it is not exactly scientific. If anyone has any suggestions they would be greatly appreciated.

Here is a graph of the comparison data I am using

It is approximately linear, but a linear function does not work as well the individual data points. If there is another form of equation to be taken from this data, the math is beyond me.

Vic Muzzin · Post by **Vic Muzzin** » Tue Oct 19, 2004 3:15 am

Knowing that Bet Per eclipsed on Oct 7th and was constant on Oct 6th, I thought it might be safe to simply take a ratio of Bet Per (Oct 7th)/(6th) from Mauna Kea. The images have been compared on the basis of matching altitudes.

The large gap is due to the fact that the CONCAM switched to 20 second exposures at different times on the two nights and it took awhile for the images to get back in line with each other.

RJN · Post by **RJN** » Tue Oct 19, 2004 6:07 pm

Hi Bruce!

Well, as you can guess, it's all "same problem(s), different day." But we've been gaining some useful experience. Here is a thread that specifically discussed a photometry technique: http://nightskylive.net/asterisk/viewtopic.php?t=78
A lot is explained there, and even example Excel files are downloadable.

Why doesn't atmospheric correction work? That has been looked at (again) recently by graduate student Tilvi among others. When doing my own study, I found that the ratio (C16-B)/(C5-B) changes with altitude. If atmospheric effects were all that were acting, this ratio should NOT change with altitude. Therefore C5 and/or C16 themselves have an intrinsic altitude dependence that has nothing to do with the atmosphere. Therefore they must be calibrated emperically. See for example: http://nightskylive.net/asterisk/viewtopic.php?t=129

The old fix of comparing a star to itself at the same sidereal time but on a different night still works great. Much of our differential photometry uses this, and shows constant stars to be constant within 0.003 magnitudes (!) or better, when averaged over the entire night. I think the above linked threads show some cases of this, too.

I hope this is a help!

- RJN

Starship Asterisk*

Beta Perseus

Beta Perseus

Photometric Techniques