Probably good to reset the purpose of the "surge" containers and vacuum method.
A few months months ago, the circulating flow in the tank was achieved by separating the tank into four channels of flow with one big common volume in the back. There were four arrays of powerheads connecting the channels to the back volume. The intent was to activate the powerheads to create flow in some channels and accept the reverse flow through the inactive channels. That creates a circulating loop of flow.
Fast forward to the implementation phase and I realize that the cost of the 48 powerheads wasn't trivial. Neither was the power consumption to achieve the 4in/s targeted linear flow. The straw that broke the camel's back was the wiring and maintenance. The incursion of life would require regular maintenance of the powerheads and creating access mechanisms became a nightmare with 48 pumps with 48 wires - each connected to a power driver synchronized by an arduino. All possible - but wow the complexity.
I was undeterred until I started looking at the surge section (separate from the circulating flow) and the desire to create a separate grazing ground for turf to grow in sunlight. Here's how the thought evolved:
1. I need to create a scrubber that is only accessible to fish when I choose to. I also need to be able to get them out when I want to. I have no more tank space, but I saw the inverted tanks and though that an inverted tank made the ideal controllable grazing ground. So I went vertical up the wall where the sun was shining but nothing was using it.
2. To automatically fill and flush the grazing scrubber (to allow fish or expel them), I needed an automated vacuum and seals to hold or purge air pressure. So, the grazing grounds became a defacto surge.
3. But my problem with surges is the burden then create on the overflow with siphons and vortices, etc... this is what my wife calls the "toilet flushing" events. So I came up with the idea of separating the vertical tank into two tanks 4' wide each. The idea is that if one is full and the other side is empty, I can connect the top air pockets and create a circulating flow (not a surge) that sucks water up one and pushes water down the other. That was it... flow without flushing!!! The idea of connecting the air pockets worked (video on YouTube) but was slow compared to the surge.
4. I needed to figure out what the formulae were that controlled the rate of flow as a function of design variables... math math math... got it and yes! I can make it work.
5. So now I had a circulating flow scrubber surge (could circulate or surge)!! And the best part is that it requires no pumps in the water. No wires in the water. No maintenance of equipment in the water. The fish can swim up to graze and I can slowly lower the level so they stop. I can run it as a waterfall or updraft scrubber too when it's air filled or water filled.
6. Turn my attention back to my circulating flow powerheads.. they're now unnecessarily complicated. Their only advantage was that they were an array that allowed me to create a slug of flow down the channels that wasn't a circular pump output. So, I replaced them with an array of PVC pipe and now I get the same effect from two vacuums over four tanks.
The tanks cost $350 each. The vacuums $150 each. The valves cost $15 each x 10 = $150 per channel. I actuate the air valves instead of the water so I need 4 actuators per tank at $50 each ~ $200.. $850 per channel with only four vacuums and 16 valves to control. Compare that to 48 DC powerheads ($1000 per channel before the controls).
So easier, cheaper.. but the flow was still not as much as I'd like.
7. I realized that as I vacuum the air out of one container, I can inject it into another!! The vacuum's blower output was able to pressurize the second tank. That means that the pipes weren't just pumping water out (like a surge), they were drawing water in. Unlike a "pump fill", a vacuum fill sucks water to create a counter flow back up. Likewise, the tank rebalancing also sucks water back up. This exceeded what the one way powerheads and pumps could do... true reversible flow!
And that's where we have the current design.
