I'm starting to think that my idea of switching to the 183 with the rasa8 may be a fool's errand altogether. I currently use the ASI1600 and it has much larger pixels, theoretically they should collect more photons than the 183 and even the higher QE of the 183 may not be able to offset the larger pixel size of the 1600.
You will likely have more clipping issues with the ASI1600. IMO one of the biggest benefits of using the IMX183 on a RASA is the fact that it has tiny pixels, and those pixels have excellent capacity. In fact, the per-micron effective FWC is worse for the ASI1600...it is 1778e-/µm^2 for the ASI1600, and 2604e-/µm^2 for the ASI183. This would be for minimum gain...which you should probably be using with the RASA. Further, the ASI183 has smaller pixels, by a factor of ~2.5x. Accounting for the higher Q.E. (~1.4x higher with the IMX183), that means you should, technically, be able to expose with the IMX183 ~2.7x longer than with the ASI1600 before you start to clip. Or, alternatively, expose for the same amount of time...and clip less!
IMO, clipping is the biggest issue using the ASI1600 with any super fast, large aperture scopes. There are far too many threads on these forums with people complaining about star clipping when using the ASI1600 on Hyperstar or RASAs. In general, I think this is a particular issue with these super fast scopes in general...star intensity grows much faster when you enlarge the aperture and reduce the focal length than extended object signal...so you clip faster and faster, while your background signal doesn't grow fast enough, and you end up with dynamic range problems.
The best cameras to pair with a RASA or Hyperstar are either cameras with huge pixels and massive FWCs, or cameras with tiny pixels and large FWCs (which, in fact...at minimum gain, the native FWC of the IMX183, when scaled out to the pixel area of a KAF-16803...is a MONSTROUS 210,000e-). So basically, FWC is of supreme importance with RASA/Hyperstar. The IMX183 gives you tons of FWC on a unit-area basis. The ASI1600 does not.