I shot the Squid this season from Bortle 5 skies using the ASI1600 at unity gain (gain=139, offset=50) with a Takahashi FSQ106 with 0.73x reducer (effective f=3.6) and 5nm Astrodon filters. I only shot OIII data on moonless hours and the total exposure time on the squid was 97x8min~13 hours. I usually target 1200 to 1500 DN for subs on the ASI1600 at unity gain. You can find my image on astrobin here. I shot a two panel mosaic to cover the whole flying bat, but you can see the exposure times for each panel in the description. I tone mapped my image but I did not have to stretch the OIII excessively. You can see the inverted OIII signal of the squid in revision G. I think there is a bit more SNR in the squid in my image but certainly not 10x.
The ratio of f-ratios between your system and mine is 1.33, so that amounts to ~1.8x imaging time to achieve the same pixel SNR (same camera in both setups). I don't think gain changes SNR calculation, that's only affected by the amount of object vs background light captured (provided read noise is negligible). It just affects how quickly you fill up the wells, which sets the sub duration. At twice the gain, you could shoot 4 minute subs instead of my 8 with similar results but more star clipping.
2 minute subs seem a bit on the short side and may affect how you build up SNR as a function of total integration time. However, if your subexposure ADU counts are in the range you reported, that's not likely the case. You seem to be well in the shot noise dominated regime. which then points to either significant light pollution, or wider band filters (what filters did you use? signal will be affected by transmission of the filters and noise by rejection band of the filters), or shooting with the moon in the sky.
finally, given the wide range of ADU counts you reported, I agree with Nico that you may want to make sure that a few poor subs are not wrecking the whole integration. Blink is a good way to go, Subframe Selector process in PI is useful as well (you can look at star count are exclude images with star count < 2 sigma from the mean or so).
Edited by lucam, 23 October 2019 - 10:24 AM.