Did i miss how the new water enters the system or are you still working on that?
Also a slight delay between opening the NC Solenoid and the closing the NO Solenoid might be a good idea so as not to make to much back pressure if the one closes faster then the other opens or something.
I got a little lazy on the drawing.
There's a manifold before the MD 7(in the latest diagram). There are 4 lines coming into it. 1 co2, 2 ferts, and the 4th is refill water. Here's where I got lazy, the drawing doesn't include the refill solenoid.
The solenoids that are on the drawing are wired together, so that N.O. is open when water goes into the tank, and N.C. is closed. Then, for water change, they switch. They'll pump for approximately 3 minutes, and then close. There is a mini-float switch in the tank, that is will be looped with the refill solenoid and controller. A minute or so after the drain solenoids are done, the controller will turn on power to the refill solenoid, and leave power on for 5 minutes(or whatever I determine is the amount of time to refill and top off), but the solenoid valve will be shut off by the float switch when the tank is full. Then, a minute or so later, power will be shut off to the solenoid/float switch loop until the next time it's scheduled to run.
If you are familiar with systems like this, that made sense to you. If not, just hang out and I'll get a diagram with control up here in a couple days.