@Fixup: thanks for that link, tried that. The files "sys.inf" and "logotype" surprisingly did already read "dst1202b". Probably because I did that "mv /dst1062b /dst1202b" already yesterday. Probably the DSO application did that during the reboots.
Now the scope reads "DSO5202B" in its system information.
By the way, the change can also be done via USB-DSO-Tool instead of soldering a UART:
plug USB stick in scope
"cp /logotype.dis /mnt/udisk"
plug that USB stick into normal computer and change text to "hantek_DSO5202B"
put that stick back in scope
"cp /mnt/udisk/logotype.dis /"
"chown 1000:1000 /logotype.dis"
However, analog behavior is still the same, that last change probably only has a cosmetic effect.
Anyhow, I think today I got a clue to it. I can sometimes see the signal shape "jump" between two different bandwidths. I made a sceenshot (screenshot hantek20_1.png) with infinite persistence that shows what I mean. Interestingly, the scope only seems to do that at 2ns and 4ns timebases. At 8ns or more, it always shows the full bandwidth. (Yes, I am absolutely sure that it is not my test signal that is causing that strange effect...
)
When I manage to capture the waveform at the "better" bandwidth, I can confirm ~ 1.8ns transition time as expected for 200MHz BW. I conclude from this that the hack was successfull in general, but the designers do some strange things with the data. It seems to depend on the exact signal shape. The above signal was taken with my 500MHz passive probe. When I take the same signal with a better bandwidth (5mm coax cable, screenshot hantek21_1.png), the scope always seems to "choose" the higher bandwidth. As if the software seems to analyse the waveform and judges from the slopes if it should apply some smoothing to it or not, in order to make the result look better....
Has anyone observed a similar behavior with his Hantek?