I have made the transition from a visual to a CCD observer.
When I was a visual observer I submitted a lot of fainter than observations. I am currently using a V filter only, no-transforms and on some occasions have imaged fairly faint variables with low SNR ratio's and error factors of almost a tenth of a magnitude.
Just a couple of nights ago I imaged QZ VIR and using VPHOT came up with an estimate of 16.564 error .094 and SNR of 12!! The check star was 14.732 error .036 and a SNR of 46.
When I look at recent observations for the star they are all fainter than visual observations.
My observation is not the best from a CCD perspective, but it provides MORE INFORMATION than is available looking at the recent visual observations.
If there were a bunch of CCD observations in recent days that had lower error factors and greater signal to noise ratios I would NOT submit my observation. Those would be better observations and my observation would only add noise to the system.
But there isn't.
So in cases like this. Is it worthwhile to submit a CCD observation that by CCD standards is a pretty bad image? However, when you compare it a visual observation it seems that it is worthwhile to submit the observations by VISUAL OBSERVATION standards. The star in question is a UGSU star.
Thoughts?? The image is still sitting in my analysis log.
Vladimir
Hi Vladimir,
As long as you have what you feel is a reasonable uncertainty estimate, then I would recommend submitting the measure. A researcher can then make her/his own decision as to whether to include the measure in an analysis, and as you mention, such high-uncertainty measures are often better than visual "fainter than" submissions. Even "fainter than" CCD submissions are ok IMHO, as they often indicate that something like a CV is in quiescent state and is not doing a "stunted" outburst. "Fainter than" for CCD measures is harder to make and a visual "fainter than", but can be done. I posted something on this several years ago. You can simply do it identically to a visual "fainter than" by noting the faintest comp star visible in your image, or you can do it statistically.
Arne
I have been reporting very faint, low-SNR photometry for several years now mostly for objects that are under-observed or are of high interest. As I have gained experience I have found that it is important to be sure you have the variable identified correctly. As often happens, a faint very close companion can confuse indentification. The recent nova V1391 Cas light curve has a few examples of possible mis-identifications. The old-style printed dot charts don't always show all the faint stars so it is best to check the DSS image chart. I always looks at the G-scale DSS chart for faint companions to the variable to be sure of ID. V1391 Cas underwent a dust production event and faded quickly to invisibilty even in the Univ. of Virginia 24-inch Robotic Rapid Response Telescope that I have access to. There are three faint stars plus the nova within a few arc seconds. One is very close and I suspect is what has generated a few positive measurements during the dust event. Be sure you are reporting the photometry errors. There are many measures I have seen in the database that don't report error estimates. I normally use portable a SkyWatcher Esprit 100 + Atik490EX + Astrodon BVRI filters. I put the first two items in a comment on the webobs data record plus anything else that might be useful for decision making by users of the data such as "thin clouds", "smoke", "uncertain ID", etc. To sum up: Check you have the variable identification correct and put short comments as needed in the data record.