13 replies to this topic

[terrain_inconnu]

I tried to do photometry with the Seestar S50 on the variable star HD 50880 (Gaia DR3 2919158394848316800, period: 115.6431 min, range: 9.04 - 9.27 G according to AAVSO) by using the default settings (10s exposure, 80 gain, dithering etc) and processed the data in ASTAP. Turns out I could not detect any variability (see chart below). This is from Bortle 9 so I was not expecting a nice smooth light curve. Any recommendations on how to retry this given the light polluted sky I have to operate under?

 

 

[SeymoreStars]

I tried to do photometry with the Seestar S50 on the variable star HD 50880 (Gaia DR3 2919158394848316800, period: 115.6431 min, range: 9.04 - 9.27 G according to AAVSO) by using the default settings (10s exposure, 80 gain, dithering etc) and processed the data in ASTAP. Turns out I could not detect any variability (see chart below). This is from Bortle 9 so I was not expecting a nice smooth light curve. Any recommendations on how to retry this given the light polluted sky I have to operate under?

 

  HD 50880.png

Are you able to post an image from your SeeStar so we can visualize what you're seeing?

[Airship]

I do all of my photometry, including my Seestar S50/30 photometry, from my Bortle 8 backyard. My first thought is whether you are using the source images or the enhanced Seestar images. I use 10 second subs, save all of my source images, stack them in Deepsky Stacker, split the color channels in Nebulosity, and use the green channel for my photometry. I haven’t done a rapid variable yet, but my plan is to average them in sets of 3-5 minutes in each set.

[terrain_inconnu]

Are you able to post an image from your SeeStar so we can visualize what you're seeing?

Here is a single frame from the sequence (arrow indicates the star I am trying to do photometry on):

 

[terrain_inconnu]

I do all of my photometry, including my Seestar S50/30 photometry, from my Bortle 8 backyard. My first thought is whether you are using the source images or the enhanced Seestar images. I use 10 second subs, save all of my source images, stack them in Deepsky Stacker, split the color channels in Nebulosity, and use the green channel for my photometry. I haven’t done a rapid variable yet, but my plan is to average them in sets of 3-5 minutes in each set.

My process is similar (or at least I think so, maybe I am wrong as I am new to this): I copied the 10 sec individual source frames (FITS) from the Seestar to my pc and then run them through ASTAP, i.e. extract the green channel, solve, align and then do the photometry.

[Fabricius]

This star has been observed by the TESS satellite.

 

 

Obviously, it's a multiperiodic low-amplitude variable. The amplitude is only 0.0113 m, so it's probably too ambitious to detect the variations if you use a Seestar from Bortle 9.

 

BTW, I could not find a period of 115.6431 minutes in the TESS data. The dominant period is about 0.5399 days.

I think it's a Gamma Doradi star, not a Delta Scuti or SX Phoenicis variable.

Edited by Fabricius, 11 February 2025 - 05:53 PM.

[Airship]

Ahah! It sounds like you are extracting photometry data from the individual frames. That works, but you’ll pick up the frame to frame noise in the background. That’s why I like to stack my frames in groups of at least 16 frames. This reduces the random noise by about a factor of 4. You might achieve a similar level of noise reduction by binning your date in groups of 16 data points (+/-8 points) or to use a traveling average that is 16 points wide. A more sophisticated filtering approach would be to use a Savitsky-Golay filter. An S-G filter is fairly easy to implement in Excel and a smaller window on the order of 11 points wide might work well. Anywho, if you have your source images you have a lot to tinker with. I keep mine until I’m happy with the data and then delete them (keeping the stacked images).

 

My Seestars have proven to be wonderful tools for photometric work and once you get your process worked out you should get good results. I’m hoping to do some work on fast variables once we get past winter and I start to get my observing campaigns stood back up (I kinda got distracted by the pending T CrB eruption ).

[terrain_inconnu]

This star has been observed by the TESS satellite.

 

HD50880 TESS lightcurve.png

 

Obviously, it's a multiperiodic low-amplitude variable. The amplitude is only 0.0113 m, so it's probably too ambitious to detect the variations if you use a Seestar from Bortle 9.

 

BTW, I could not find a period of 115.6431 minutes in the TESS data. The dominant period is about 0.5399 days.

I think it's a Gamma Doradi star, not a Delta Scuti or SX Phoenicis variable.

The discrepancy between the data from AAVSO and TESS is quite large (e.g. amplitude is given by AAVSO as 0.23, i.e. mag varies between 9.04-9.27). I just checked TESS and on the AstroView map it is indeed the same star. Is there any other explanation for the difference aside that either AAVSO or TESS is wrong?

[terrain_inconnu]

Ahah! It sounds like you are extracting photometry data from the individual frames. That works, but you’ll pick up the frame to frame noise in the background. That’s why I like to stack my frames in groups of at least 16 frames. This reduces the random noise by about a factor of 4. You might achieve a similar level of noise reduction by binning your date in groups of 16 data points (+/-8 points) or to use a traveling average that is 16 points wide. A more sophisticated filtering approach would be to use a Savitsky-Golay filter. An S-G filter is fairly easy to implement in Excel and a smaller window on the order of 11 points wide might work well. Anywho, if you have your source images you have a lot to tinker with. I keep mine until I’m happy with the data and then delete them (keeping the stacked images).

 

My Seestars have proven to be wonderful tools for photometric work and once you get your process worked out you should get good results. I’m hoping to do some work on fast variables once we get past winter and I start to get my observing campaigns stood back up (I kinda got distracted by the pending T CrB eruption ).

I tried to group frames together but found no automated way to do this (I tried via AstroImageJ but could not make it work). Do you group the frames in DSS?

 

Re S-G filter: just skimmed through the basics and seems straightforward to implement. Thanks for the tip.

Edited by terrain_inconnu, 11 February 2025 - 06:31 PM.

[Airship]

Ugh. No, I'm olde school. I do much of this manually. I thought that I had a semi-automated way of doing this when I was observing fast Cepheids many years ago, but I can't recall exactly what how I did it. I currently just stack them in groups. It's not too bad.

 

Fun fact... the original paper by Savitsky and Golay was once the most referenced paper in the open literature. We used to have photocopies of it tucked away everywhere. It's still a nice read, and I find it easier to follow than the Wikipedia page (which is still okay). Golay was one of the pioneers in modern gas chromatography and the S-G filter was developed to smooth IR spectra without losing fine detail. Savitsky was his graduate student.

 

Neat stuff.

[RedLionNJ]

The discrepancy between the data from AAVSO and TESS is quite large (e.g. amplitude is given by AAVSO as 0.23, i.e. mag varies between 9.04-9.27). I just checked TESS and on the AstroView map it is indeed the same star. Is there any other explanation for the difference aside that either AAVSO or TESS is wrong?

Is it possible the long-term variation can exhibit a brightness between 9.04 and 9.27, while the apparent 115-minute variation is only 0.01m? Something of this order would be extremely difficult to detect with the level of calibration available with the SeeStar. You'd definitely need to align & stack multiple frames to arrive at each data point and even then, the noise would likely still exceed the 0.01m variation.

[terrain_inconnu]

Is it possible the long-term variation can exhibit a brightness between 9.04 and 9.27, while the apparent 115-minute variation is only 0.01m? Something of this order would be extremely difficult to detect with the level of calibration available with the SeeStar. You'd definitely need to align & stack multiple frames to arrive at each data point and even then, the noise would likely still exceed the 0.01m variation.

I do not know enough about this, but this is the second star I was not able to measure it's variability, the first one being BL Cam. My next try should probably be with a longer period / larger amplitude star.

[rutherfordt]

 

  HD 50880.png

Although I am not familiar with the target star, so cannot comment on it specifically, I do have a couple of comments about your graph.

 

1)  Normally, when graphing magnitudes on the y-axis, the axis is inverted so that brighter is toward the top (remember, the magnitude scale is an inverted scale).

 

2)  Just glancing at the graph with my eye, it looks like there is a general trend of the graph going upward over time.  If you invert the y-axis as I mentioned above, that would be a gradual decrease over time.  If this matches what the star was supposed to be doing (dropping in brightness) during your imaging session, then perhaps you did capture something after all.

 

3)  Error bars on the points will help address whether you captured something or not-- if they are larger than the apparent drop that I mentioned, then it may not be real-- if they are smaller then perhaps you did detect something.

 

Tom
 

[StupendousMan]

A good practice when performing photometry is to measure two stars: the target of interest, and a nearby star of similar brightness (and color, if you can manage to find one).  After reducing all the data to produce a list of instrumental magnitudes for each star in each frame, compute the difference:

 

           delta(m)  =  (instrumental mag of target)  -  (instrumental mag of nearby star)

 

Variations in brightness due to clouds or extinction will tend to cancel out in this differential magnitude, providing a better idea of the target's behavior, and giving a decent indicator of the uncertainty in the measurements.