There are a number posts on the ASIAIR Facebook page showing what appears to be images from the ASI120 camera tearing up during guiding (and causing star loss and guide failure).
I experienced the same thing when checking out the ASI174MM-mini earlier this evening as a main camera in Preview mode (image and log already sent to ZWO).
It appears that it may not just be a ASI120 camera problem, and also not just a problem with guiding.
Curiously, I run an All-sky Camera using a ASI178MC (uncooled) continuously 24/7 and have not come across image tearing. I also have not seen it with the ASI2600MC (pro, cooled).
One interesting thing is this... in the Facebook posts and also with what I encountered with the ASI174, the image tearing appears to start at the bottom half of the image plate. I happened to have the crosshair turned on, and the tearing began right below the horizontal red line in the Preview window.
I don't know if this is coincidental. If it is not, and if you are losing stars during guiding, try choosing a guide star that is in the top half of the guide window. Just tap on a star instead of letting ASIAIR choose a star, ASIAIR will move the green square to the star you have chosen (there usually is a small delay), then start guiding.
Normally, star loss is caused by degraded signal to noise ratio, often due to loss of contrast, and can also be caused by the intensity of the star dropping below some threshold. (A passing thin cloud can be the root cause for both). Choosing a guide star on the top half of the guide window will not help this case. Plus, until we have more examples, I am not at all certain that the tearing only happens to the bottom half of the guide image. But it might be worth trying if you are out of options.
I normally use an ASI290MM mini as my guide camera and occasionally, the ASI462MC; I have not yet encountered image tearing from those guide cameras. But I also don't have as many clear nights for the past 60 days to have much time with ASIAIR v1.5 and newer updates to say for sure that they don't suffer from image tearing.
Chen