Skip to main content

Data scatter.....questions.

21 posts / 0 new
Last post
Data scatter.....questions.

Thanks guys, for all the help in my previous discussion topic.

I downloaded the AAVSO photometry manual and have been working my way through it. I have a hard time with linear thinking mode....I much prefer running in circles!! I have made my FIRST circle through the manual and actualy have some measurements.

I ended up with the AP900 mount, with a 8inch f4 newt and the ST-7XE at my roll-off roof observatory.

I have taken a "bad" set of calibration flats...first run through the circle. I don't like my flats.

I usually observe T CRB visually and normally report 10.1 as a magnitude. My CCD image, through a V filter, showed a magnitude of 9.7.  That got me thinking about data scatter. I don't know the magnitude of T CRB that night visually since I did not observe it visually.

However, I did take a image of SS CYG and caught it in eruption. Using the 86 comparison star I showed a magnitude of 8.813 for SS CYG. For the 98 comparison star I showed a magnitude of 8.734.

Checking in the database, those are in the ballpark.

In one of the observations I clicked on.....the comparison stars were shown as "ensemble".

So are there any rules for chosing more than one comparison star??? And how do you average the different values for reporting the magnitude??

Given the issue of flats....are there rules of thumb for deciding what is a good light flat.


Eric Dose
Eric Dose's picture
Nice equipment for first CCD

Nice equipment for first CCD work.

Are your flats twilight flats, or panel flats? Why do you say that yours are bad? Can't advise much without knowing this.

T CrB is a fairly red star. So are you transforming your CCD data? If your really must observe without transforming your data, you'd probably do better by starting CCD observations with less red stars.

V ~ 9.7 is within T CrB's normal range; you might be better off comparing your CCD data with others' CCD data (LCG tool, or even with the WebObs page), preferably in multiple filters, rather than trying to compare with visual results, which have their own problems with red stars. Starting with low-amplitude and/or long-period stars will minimize the time factor, and starting with well-observed stars will build confidence in your technique and results.


Yes, checking the data base 9.7 with the V filter fits fairly well with others observations. I was surprised at the difference between the V filter observations and the visual observations. But I only did the month of August for T CRB for comparison with my observation.

Being a visual observer I never thought of T CRB as red. I guess I will start paying more attention to "colors". I am not transforming the data and quite frankly, do not understand transforming at this point on my learning curve.

The flats are panel flats. I really want to stay with using panel flats rather than t-shirt flats. Looking at the flat field, my assumption has been that from the middle of the frame off to all sides the gradiant should be balanced.

My flats have a bright spot, just a bit off-center with one edge slightly dimmer and the other edge much dimmer. Those a relative terms....the pixel values are pretty close.

My first guess is that is an artifact of positioning the flat panel in front of my scope. But I really don't know at this point if it is significant.

On my next run around the "circle". I will switch to t-shirt flats and that should give me enough info on whether it is the light panel position or my optical path. How's that for a plan??

That AP mount is wonderful for finding variables. Really almost as quick taking pictures as it was to do visual estimates. The nice thing, I still have the original data if needed,not just my human memory banks.

Given the expense of filter sets I am thinking of just staying with V filter work.

I am using AIP for this first circle through the process. My astrophotography friend is recommending that I manage my dark and flats through SkyStacker.  He says it is simpler, quicker and easier than AIP. Thoughts??

spp's picture
evaluating flats


I think choosing and perfecting a flat method is one of the biggest challenges we face when we start learning to do CCD photometry.  A big part of this challenge is evaluating the flats.  

For most of us, and especially for users of SCT's and reflectors, I think the most important function that flats provide is to correct vignetting.  To can check how well your flat does this using the AIP Profile Tool (in the Measure drop down menu). 

You should start with a starfield image that has a good SNR in the sky background.  In the example, the image was dark and bias corrected, but this may not be necessary if the SNR of the sky background to good enough.  

In the unflatted image the Profile Tool shows the vignetting pattern, the sky background hump in the center.  In the flat corrected image the sky background is flat, as it should be. 

The Profile Tool may also reveal other gradiants in your images to help you track down problems in your raw images and flats.



spp's picture
using ensembles

"In one of the observations I clicked on.....the comparison stars were shown as "ensemble".  So are there any rules for chosing more than one comparison star??? And how do you average the different values for reporting the magnitude??"

The AAVSO comp sequence, in whole or in part, can be used as an ensemble.  The usual guidelines for picking comps, single or ensembles, are discussed in the CCD Photometry Guide.  But often, you just have you use what is available. 

If you want to try your hand at ensembles without having to learn new software tools, there is an good candidate variable only 22 arcm west and slightly north of T CrB.  Use the VSP to plot a finder chart for ASAS J155754+2600.1

Also, use VSP to get the VSX information for this star. 

If you place the variable in the NW part of your image you should be able to get three AAVSO comp stars in your field:  99, 112, and 94.   You would then use each of these stars to make three independent magnitude measurements.   The average of those measurements would be reported as a single measurement.  The s.d. would be the uncertainty.  

You would want to get your flat field technique down pretty well before reporting these measurement, but this might be an interesting practice exercise. 

Most observers use ensembles with software that does all this automatically, e.g. VPhot.  I think AIP may also handle ensembles.



spp said....

spp said....

"I think choosing and perfecting a flat method is one of the biggest challenges we face when we start learning to do CCD photometry.  A big part of this challenge is evaluating the flats. "

When I tried going through the manual last time this is where is came to a halt!! Fortunately, this time my astro-imaging friend has managed to gently get me through the flat process. Sometimes, it took five explanations before I finally understood!!!

He suggested using monitor flats in my observatory, instead of purchasing a light box. Given my "personality" and mode of working sky flats are not going to work for me.

Thanks for the info on the profile tool. I will check some of images.  That does seem like a handy tool.

Ok, I did check a couple of my raw images without a flat applied to them. Completely straight across the field of view. Given the small size of the ST-7 chip and the large field of view with the 8inch F4 newt is doesn't look like vignetting is going to be an issue.

Is the ADU numbers shown in the profile tool....useful for image analysis?? And how would you use them??

I will try and image the suggested field for T CRB. Is that a possible variable or the three comparison starts are to be used for estimating T CRB??

I am not reporting ANYTHING until I have a work flow that I can trust and replicate.

I am also hoping to being able to take the next choice class next time it is presented.



I just saw I can upload my

I just saw I can upload my they are...title contains the basic info.



spp's picture

I'll try to respond to a few of your comments and questions from your previous post.

Taking CHOICE courses:  VPhot is offered only once per year in September.  I'd recommend that you sign up for this course immediately so that you don't have to wait another year.  You'll want to take CCD1 and CCD2 in the winter-fall of 2020.  Knowing VPhot will be a advantange for those, especially for CCD2.  You could learn VPhot on your own, but you'll be learning it by drips and drabs.  I think the taking the course is better,  and certainly quicker.

SkyStacker:  I'm not familiar with this program, so I can't give you any useful advice specifically about that program.  Since you are already using AIP, I'd recommend you stick with that for image calibration and analysis.  Lots of AAVSO observers use, or have used, AIP, so you can get software specific help and instructions from the forums when you run into problems.  Another popular program is MaximDL ($$$).

Posting images:  JPEG's are good.  FITS are better.   Forum members can load your FITS in their own software to help you with problems.  Also, the FITS header contains lots of useful information that can be used to run down problems.  

The odd thing about these forums is that they don't permit FITS attachments.  You can zip compress the file(s) to post them in the forums.

Flats:  Your flat image shows vignetting in the left upper and lower left corners.  My concern is that it doesn't show the same pattern on the right side.  This makes me think that the illumination in the flat method you are using may not be uniform.  Use the AIP profile tool on your FITS image corner to corner, left to right, top to bottom to help you see the gradients.  In the typical vignetted flat, there is a roughly symmetrical circular bright center with all the corners darker.

Filters:  A single V filter is fine for starters, but since you are obviously serious about learning to do CCD photometry you will want other filters sooner or later.  I'd suggest you keep your eye on the Cloudy Nights and Astromart classified ad's for a filter wheel to fit your camera.  SBIG no longer makes the wheels for the ST-7 series cameras.  Once you have a filter wheel, you can add filters at your leisure.



Can I upload MY images to

Can I upload MY images to VPHOT for analysis??  I assume so.  

Is there a dowloadable version of VPHOT.....lots of times I am outside of any web access.

MZK's picture
Upload images to VPhot


If you are an AAVSO member, you can upload your fits images to VPhot. Download and look at the VPhot Guide. Make sure the images have the required fits headers or enter them in Upload Wizard..

No, VPhot does not have a downloadable version.


PS: If you have any further VPhot questions, you should put them on the VPhot forum.

One last question....if I

One last question....if I upload my images to VPHOT is there a simple way to keep them from entering the database??

Right now...I am trying to learn the system and debug my workflow. So anything I upload to VPHOT should NOT be a observation.

spp's picture


Please post this question on the VPhot forum so other new VPhot users will get the benefit of the answers.  It is a pretty active forum.  You can learn a lot there.


MZK's picture
Not to worry, try it!


Answer.... IF you upload images to VPhot, they will NEITHER automatically generate photometric results NOR submit them to the AAVSO International Database (AID).

There are several steps that you must complete to generate a photometric magnitude and you must subsequently submit the result to the AID.

Upon uploading, the images will appear in your VPhot account. This assumes you have an observer code (SVD?) and that you have logged into VPhot.

HTH, Ken


thanks.....guys. I will get back on the other issues. 

Flats...It appears the flat image I posted is in my imaging trail, not the light box. Last night, I noticed a major problem with "trailing" which I think is really not having the chip square to the light path. I am going to work on this...

So one step at a time. Back to flats...the "squareness" issue IMHO is related to the distribution of the light curve in the flat.

So how does this flat look??  I might switch to this....other possibility is a solar filter flat.


File upload: 
spp's picture
much better flat

This al_flatframe is a big improvement over the previous flat. 

There seems to be an additional shaddow on the right side of the image.  Take a good look at your filter to see if you can see something that could be causing that shaddow over the right 25 or 30% of the image. Another way you could test this is just remove the filter, then shoot the flat again with everything else unchanged.  I'm guessing the filter is the culprit.

I suppose it t could also be a smudge on the cover slip over the CCD or even something in the way the flat was illuminated. 

Was this flat made using a flat box or an EL panel?   

If you can't find the cause in the filter or camera for that exra right side shaddowing , try rotating the flat box or EL panel 180 degrees with respect to the camera (as compared with the orientation you used to make the posted image) to see if the shaddow switches sides.

Here's something else to check.  With the camera well focused, look down the sky end of the scope to see if the focuser drawtube is in the light path.


Went out to the observatory

Went out to the observatory and shot a bunch of flats using and all of them showed the "unbalanced" light pattern.

My friend, suspects it is the collimation of the 8f4 newt.  He did manage to "collimate" the scope so the flat-field was even, but at that point the scope was totally out of collimation.

We were not quite prepared for a extended session and didn't have the complete set of collimating tools.

Plan is to collimate the scope. Then try the various methods of doing flats and see if they are consistent. I guess because of the small field of the ST-7 the stars remained good in the field of view.

There was nothing intruding into the light path. Moving the camera or lightbox did not shift the image pattern.

We will give it a couple of days rest and try again. 

HQA's picture
flat field

The basic flat looks about right for an 8" f/4, but with fairly strong vignetting.  You are making good progress!  You want the center of that vignetting to be in the center of your camera field of view - then you know the optical axis is properly centered.  The "smudge" on the right hand side of the flat looks to be broken up into a number of "dust donuts", which is good.  There is a calculator on-line:

which can use the donut size and information about your telescope and camera to determine the distance from the sensor to the "dust".  My guess is that it is on the entrance window to your camera; filter donuts at f/4 are quite large.

However, first get the right tools and re-collimate your telescope to give nice star images!


Another flat field....

Al and I had a very "busy" and fruitless day on Sunday. I think the issue is trying to get a good flat with the lightbox. That doesn't seem to be working. We tried working on the collimation and ended up collimating for the lightbox, not the image. I don't think that is a good idea. So we did manage to get a "decent" flat.

So change is strategy. Giving up on trying to get a decent flat with the light box for now. I will go ahead with t-shirt flats.

But first, will go back and re-collimate for best image and worry about the flats after that.

Right now this is the flat after "de-collimation" on Sunday. This is a t-shirt flat that I took tonight.

If folks look back on the CCD image of SS CYG uploaded a couple of comments back, that collimation looked real good.

My "expert" is down in Oregon at a tennis tournament so when he returns we will recollimate and then shoot some new t-shirt flats. Hopefully those will pass mustard.

spp's picture
latest flat

This flat looks pretty good to me except it seems that the collimation is way off.  I'd suggest you recollimate and repeat with the same technique, maybe also with a longer exposure.  The maximum pixel ADU in this flat is 255.  Ideally, it should be about 10X that.  The guidance for exposing the flat is about 1/2  way to saturation.


spp's picture
latest flat

As part of your flat experimentation consider measuring the saturation level of your camera.  There is a dicussion of this in the CCD Photometry Guide.  If you do this first you'll have a better idea of how long to expose your flats.


That adu value is because it

That adu value is because it is a jpeg. 

My camera saturates around 60,000 so I am looking for around 35,000 adu's for my "final" flats. 

Looks like fall has arrived in eastern Washington and we are looking at cloudy weather for the next couple of weeks!!  Also pheasant season opens on Monday!!  

Hopefully, I am going to recollimate the 8inch f4.  Then get a "good" t-shirt flat. 

Thanks guys....I appreciate the support and comments.

Log in to post comments
AAVSO 49 Bay State Rd. Cambridge, MA 02138 617-354-0484