Based largely on the BreurRam pump developed by Gert Breur, my version features an alternate component layout.
The simplicity of the original design is due to incorporating the air reservoir, which absorbs the hydraulic shock, into the outlet tubing, with only 2 check valves being used, one of which needs to be modified.

The hydraulic ram uses pressure pulses from the closing of the waste valve to pump a small amount of water out of the discharge/outlet check valve.

My question is, can the pulses be better managed for increased efficiency?

The following documents my ram pump build and subsequent testing.

Step 1: Theory and Method

The traditional ram pump has the inlet opposite the outlet valve with the vertical waste valve connected via a T piece in between. A typical setup is the Clemson University version, in the numbered diagram the inlet is at the right at #1, waste valve at #4 and one way outlet check valve at #5 with #15 depicting the air reservoir. During the working cycle it can be seen that the pressure pulse will be reflected by the closing valve into the bottom of the T-piece and dispersed equally back upstream to the inlet as well as downstream to the outlet valve.

I suspected that directing the pressure pulses in a more focused manner toward the outlet check valve might result in better pumping efficiency.

My method of achieving this was to place the outlet opposite the waste valve so as to receive the pressure pulses more directly. One solution was to place the waste valve at the end opposite the outlet and have the inlet between the two, entering either from above or the side, I settled on a side inlet for device stability. The pulses from the waste valve would be reflected via a modified 90 degree elbow fitting to the outlet.

I built two 25mm (3/4") devices, a traditional config and my custom one, using interchangeable check valves reduced build variations which could affect results as well as keeping costs down.

Bench tests were conducted in the garden with a 5 litre container on a 50cm length of 25mm dia pipe. The outlet was hooked over a carport beam 190cm high, simply because I have no river, stream or any other similar type of environment suitable for field tests.

Step 2: Parts

Initally I used two 3/4" check valves, but after some testing I found that a 1" valve worked far better for the waste gate than the smaller one. I used a 1" connector from the valve to a 1" to 3/4" adapter to the original 3/4" connector.

Common parts

1" barrel connector
1" to 3/4" adapter
3/4" barrel connector
2 SS hose clamps
a 70 cm length of 20mm vinyl hose
straight barb fitting 20mm to 10mm
elbow 3/4" threaded to 20mm barb
3 metres 10mm vinyl hose
25mm pvc ball valve on the inlet for convenience

My custom pump body config uses:

3/4" 90 deg elbow
3/4" T piece solvent
2 x 25mm to 20mm solvent connector

The traditional config uses:

3/4" T piece, solvent on the ends and 3/4" threaded on the top
1 x 25mm to 20mm solvent connector

Step 3: The Modified Check Valve

Threading the valve stem will allow varying the length of the valve stroke to fine tune it for field use.

The 1" spring check valve was opened, spring removed and valve stem threaded M6 to accommodate a M6 coupling nut for fine tune weights. I tapped the thread too far down the stem and it did occasionally bind, the top 5mm of the stem would be sufficient.

The 3/4" spring loaded brass check valve which I modified by taking out the spring and threading the brass spindle to allow for an adjustable valve to seat distance. The valve body needs to be dismantled to modify the innards, a bit of heat to the body assists in breaking the sealant on the threads, wedge the rubber seal off the seat to protect it from the hot brass. I then found the valve to be too light and discovered there weren’t any M5 coupling nuts so I tried the spring down force mod, which was very sensitive and I couldn't get consistent results.

Step 4: PVC 90 Degree Elbow Mod

This is just a standard 90 bend that has an internal 45 degree slope to transmit the pulses from the waste gate valve to the outlet check valve.

The 45 slope was created with super glue and fine glass bead sandblasting media which doesn't reflect laser beams. i.e. camera safe, but don’t take a chance with your eyes, rather wear some sort of protection if you use this method. I figured a 45 degree surface would show a centred beam in the exit part of the elbow.

Step 5: Testing and Results

I held a measuring jug under the outlet to collect the discharged water for comparison checking. I experimented with a "snifter" port to add additional water via a check valve, but it was unsuccessful so I blocked it off during testing.

During the custom config test, 5 litres was finished in 30 seconds, while the traditional config used 25 seconds for the same amount. The waste valve was fine-tuned with the M6 coupling nut to give 4mm of travel on the custom config for approx 200 pulses per minute and that setting was untouched when used on the traditional config.

The custom config pumped:

350ml
380ml
360ml
360ml
370ml
360ml
360ml

Discarding the highest and lowest readings gives an average of 362ml pumped water out of 5 litres.

The traditional config pumped:

260ml
300ml
330ml
300ml
300ml
310ml
310ml

Discarding the highest and lowest readings gives an average of 304ml pumped water out of 5 litres.

Step 6: In Conclusion

Although it seems to be a tiny difference of almost 60ml in 30seconds, that difference translates to approx 170 litres in a 24hr period, which is largely due to the more focused and directed pressure pulses.

The video shows the smoothed output due to the air resevoir in the outlet tube which cushions the hydraulic shock.