No reply in over 4 and a half years, so I thought I would reply with my thoughts!
I do not have a JOBO 4220 separator, but some thought reveals the following deductions.
In essence, the JOBO 4220 separator is a box which has a pneumatically powered and controlled valve inside it. To achieve this pneumatic manipulation, three standard aquarium PVC tubes (4mm inside diameter) are attached to connections marked "1", "2" and "4" on the back of the ATL-1000 or ATL-1500. The chemical outlet (marked with yellow sticker) is then attached to the box of the separator.
I think that the way it works is that as the chemicals are pneumatically pumped into the tank, the pressure is also reduced in one of these lines and this draws the valve into one of it's two positions, therefore directing flow one way or the other.
So development chemicals are sent through the outlet as normal, and then the valve in the 4220 sends them into one collection container. When the fixing chemicals are pumped into the tank, the valve in the 4220 is re-positioned and when it comes time to dump them they are now sent to the other collection container. Simple really.
I think it is important to note that the chemicals share a common path even with the 4220 installed, up to the valve in the separator. So there is still the chance for contamination.
A question for the C-41 experts: would a double tank rinse between C-41 dev chems and C-41 blix be a good idea (assuming at 37.8 degrees)? This should prevent contamination. Would this be good for the end result? No colour shifts etc?
So it seems that one could install a manual switch valve (like a Y-joint) on the JOBO and achieve a manually controlled separator. I might give it a go!
Let me know what you all think...