Dennis Means, our diligent Image Inspector for BSM_TX, has been noting that images seemed weird over the past few months. We finally tracked down the problem as being bad -5C darks that had been taken in May. These darks had two different count levels, so that the top 1/2 of an image was at one mean number and the bottom 1/2 was at a slightly different count level. The result was a top half/bottom half brightness level as seen in the image thumbnails at
http://images.aavsonet.aavso.org/bsm_tx/
We've obtained new darks, thanks to Dave Cowall's late night efforts. These do not exhibit the top/bottom behavior. While this is not a large error, as each star photometric measurement uses local sky which will either be at one level or the slightly different level, it could affect faint star photometry. Therefore, we've made the decision to reprocess late-May to June, and then September/October, datasets to correct the problem. The nights that will be corrected are 20230522-20230623 and 20230904-20231008.
This means that you will be getting a bunch of "you've got data!" messages for images that you thought you had already analyzed. What I'd do is analyze a night or two to see if there is any real difference in your photometry. if so, update your webobs submissions.
We apologize for the error, and really apologize for not finding the problem quickly.
Arne