I have noticed, if the signal is somewhat unstable at 100ns using both channels then pressing F3 (TRIG), will make triggering become stable, but if you leave F3 for another function, the signal becomes unstable again. Also just selecting the blue channel in "Vol/Time" will make the signal stable
I have FW 1.6.4
This is a very strange version. If you catch the signal in normal mode in 50 us scale (and maybe others), hold it, and use cursors to measure the signal duration, it increases twice at your eyes and so do the measurements. For example, 100 us long signal shows and measures as 200 us signal!


Also, there are many other bugs, for example, if you catch the signal, it may look at the trigger correctly, but when you try to move it, the trigger changes its place unpredictably. At 2us long signals and 2us time scale, if you press CURSOR button to adjust cursors, the signal just disappears (actually goes far left beyond the screen). MOVE button has the same pitiful experience. This firmware is a garbage, I just wanted to measure the duration and the amplitude of my signal and I cannot do it... oh no, finally I could, but I wasted ~10 minutes for the task that can be completed in 30 seconds with a reliable oscilloscope

I installed ZOYI-703S-FW-V127.ZTK for comparison. There are many trigger issues as well, but you can't believe how fast it is in comparison with 158! I wish Zotek could make a new firmware without all that garbage in 158 that slows it down so horribly, just basic functionality, but reliable.
By the way, the shape of the signal is less neat in 127, which is probably closer to reality.

158 seems to spend some CPU time for various filtering and it slows it down noticeably. I think most users would rather prefer the true shape of the signal than filtered, but I think it deals with marketing more, they think if the [really ugly] signal will look neat, it will increase sales, but I don't agree with it..