On the night of February 3-4, 2021, a student and I obtained a 4h 20m hi cadence (3 20-second images/minute) V-band light curve of U Gem using a 10" f7.9 RC and QHY 174 GPS CMOS camera. I am still fairly new to photometry but very enthusiastic and interested in learning how to obtain the best data I can with my equipment. I have been practicing on U Gem as it seems like it should be going into outburst soon. In reading about it, I learned about the "orbital hump" and the short eclipse that occurs at the top of this hump. It sems the eclipse should have been detectable with my setup, but for some reason it doesn't show up. I'm curious if my lack of detection was due to these eclipses being variable in their depth or irregular in their prescence, or more likely due to being beyond the ability of my technique or equipment to detect. I have posted an image showing our light curve and an image of one of our 20-sec subs. Any thoughts or comments about the detectability of U Gem's eclipses or my techniques would be greatly appreciated.
I've never done a time series on U Gem (and now I wonder why I haven't - eclipsing systems are a favorite of mine), but I did a little digging and I'd be surprised if the eclipses have disappeared. The period appears to be slowly changing, as expected for a system with mass transfer, but it would be very odd for them to disappear suddenly. Looking at the AAVSO Light Curve Generator, I found several nice recent eclipse examples like this one: https://www.aavso.org/LCGv2/index.htm?DateFormat=Julian&RequestedBands=&view=api.delim&ident=u+gem&fromjd=2457800&delimiter=@@@.
Like most cataclysmic variable systems, there's a lot going on in the light curve all the time so it occasionally may be hard to spot an eclipse. All that "noise" from the disk, reflection effects and variable mass transfer flickers create a lot of bumps and dips.
Your setup should be good for picking up the eclipse; a short integration time like you're getting is important since the eclipse is quite short. FYI: your image and light curve didn't make into the forum