A pump tries to provide so much volume per unit of time.
The pump tries to maintain that rate of delivery.
Let's assume that the pump outlet is connected to a manifold, and the manifold has several nozzles in various sizes - from very small to very large. The nozzles are pointed upward. Let's say that the diameter of the pipe between the pump and manifold is larger than the pump outlet. Let's say the same thing about the diameter of the manifold. Let's also say that the diameter of one of the nozzles is as large as the diameter of the manifold - that nozzle would be larger than the pump outlet.
Let's further say that all of the nozzles are of the same height and that their upper edges describe a straight, horizontal line.
+--+ +----+ +------+ +--------+
| | | | | | | |
--+ +---+ +---+ +---+ +---- etc...
MANIFOLD
-------------------------------------------
.
Open all of the nozzles and turn on the pump.
Water will be coming out of all of the nozzles - some more so than others.
The total amount of water through the nozzles in x-seconds will be equal to the total amount of water from the pump in those x-seconds.
The water is distributed among those nozzles in a manner that is inversely proportional to the resistance to flow that each nozzle presents to the source.
The largest nozzle offers the lowest restriction, thus the lowest resistance. That nozzle experiences the most volume of water.
All of the other nozzles offer more and more restriction as their orifices get smaller. As the nozzles get smaller and smaller, they experience smaller and smaller volumes of water.
Since the nozzle ends are at the same level, and since water tends to seek its own level, water will be present at the tops of each nozzle. I shouldn't expect to see the water doing anything more than simply "over flowing" at the nozzle tips. That is, I wouldn't expect to see any "height" to the escaping water.
So... because the large nozzle (and supply pipe and manifold) are of a diameter larger than the source (the pump), the "net restriction" is "0-what-cha-ma-callits". This huge amount of water is being provided at what appears to be "0-psi!" The velocity of the water at the nozzles appears to be somewhere near zero. There has to be at least some velocity - the water is moving, after all. But the nozzle-velocity (muzzle-velocity) is very, very close to zero. Near-Zero velocity translates into Near-Zero height.
Now, close the largest nozzle.
The pump continues to provide the same volume. An equal amount of water is still leaving the nozzles. However, with the large nozzle closed, the volume is distributed across fewer and smaller nozzles. In this case, the net restriction is greater than zero. A pressure develops in the manifold. In order to have the same volume of water escape, the velocity at each nozzle must increase.
You now see the water developing height. The height of the water escaping from each nozzle should be consistent. They won't be the same but, assuming that the pump is reasonably consistent, they should remain consistent relative to each other.
Continue to close off the largest remaining valves, one at a time. The water height at the remaining nozzles becomes progressively higher and higher.
With only the smallest nozzle remaining open, the pressure in the manifold is at its greatest (unless you close all nozzles). Look at the height of the water column. It will be such-n-such.
Now, pick any of the closed nozzles (except the largest) and open it. What happens to the height of the water at the smallest nozzle? It changes - it gets lower. Close the nozzle you just opened. The water column resumes its previous height.
Repeat the test using various nozzles and various nozzle combinations - all the time watching the height of the column at the smallest nozzle. Repeat several of the combinations to see if the results are not consistent - they will be consistent!
So... with the smallest nozzle being used as the primary nozzle and the other nozzles being used as "trimmers", you can produce, and reproduce, any proportional fraction of height at the smallest nozzle. Instead of aiming the "trimmer" nozzles upward, aim them downward or back into the supply tank.
You will have to play a "sizing game" with the various "trimmers" to ensure that you end up with the expected heights. I suggest using "trimmers" that are sized in a binary fashion... 0^2, 1^2, 2^2, 3^2... etc. Then correlate the various heights to a binary value. Then apply the binary number to the bank of trimmers.
Sending "1" turns on trimmer "A"
Sending "2" turns on trimmer "B"
Sending "3" turns on trimmer "A & B"
Sending "4" turns on trimmer "C"
Sending "5" turns on trimmer "A & C"
...etc.
