If you use a Windows PC, I recommend that you play the videos with a program such as Irfanview which allows you to change the playback speed (the movies normally run at 25 frames per second).

I recorded these images over the course of one week in May 2000. You may want to read my description of how I did this. It took a considerable amount of time to set up each launch, record it, and then download the images to computer. From the outset I did not know how many launches I would have time to record, so it was hard to decide what input parameters (bottle type, amount of water, pressure, etc.) to focus on. I was largely guided by my own curiosity and the constraints of time, safety, and my supply of bottles. Certainly there are interesting combinations of launch parameters I was not able to test, and questions I was not able to investigate. The whole filming process was a learn-as-you-go kind of thing. As ‘Doc’ Edgerton said, “That's the nature of research—you don't know what in h*** you're doing.”

Each launch is designated a letter, A through O, in the chronological order they were recorded. Here I group them by loosly defined categories. In many of the movies you will see a scale on the edge of the frame. This is a strip of masking tape, marked off in 50-cm increments. In some movies you may notice it “flapping in the breeze” after the rocket passes.

2L with 25cm launch tube

Video link & file size

295 KB

354 KB

222 KB

487 KB

Flight A E D O
Rocket volume 2100 mL 2100 mL 2100 mL 2100 mL
Rocket mass 159 g 160 g ? 109 g
Water volume 480 mL 514 mL ? 705 mL
Pressure 88 psi 100 psi ? 100 psi
video speed 2000 frames/s 3000 fps 2000 fps 3000 fps
Length scale 146 pixels/m 77 pixels/m ? 434 pixels/m
Comments below 1 2 3, 4 5, 6
  1. water volume estimated 10mL—there was some leaking during pressurization. The bottles were "two-liter" bottles but have an actual total volume of 2.1 L after pressurization (the bottle stretches).
  2. estimate of final velocity from last four frames is 73 m/s
  3. flight D was an attempt to answer the question: “what happens when the launch tube slides through the liquid? does it leave a hole behind?” The recording is inconclusive.
  4. Note that the top 2cm of my launch tube is colored with black ink.
  5. in flight O, the light has been placed behind and below the bottle to increase visibility of disturbances in the water in the bottle. No milk was added to the water this time to maximize visibility.
  6. estimated velocity as bottle clears the end of launch tube, based on 8 frames, is 12.5 m/s

2L sans launch tube

Video link & file size

120 KB

136 KB

Flight B C
Rocket volume 2100 mL 2100 mL
Rocket mass 161 g 166 g
Water volume 416 mL 476 mL
Pressure 88 psi 88 psi
video speed 2000 fps 2000 fps
Length scale 146 pixels/m 146 pixels/m
Comments below 1 1, 2
  1. notice the unexpected symmetry of the "cloud burst" near the end of water thrust. Read my interpretation of the cloud burst in the analysis section below.
  2. estimated final velocity from last four frames is 59 m/s

Smaller bottles

Video link & file size

166 KB

71 KB

Flight F M
Rocket volume 1550 mL 600 mL
Rocket mass 128 g 43 g
Water volume 382 mL 116 mL
Pressure 100 psi 75 psi
video speed 2000 fps 2000 fps
Length scale 77 pixels/m ?
Comments below 1, 2, 3 4, 5, 6, 7
  1. seltzer water bottle—contains efficient straight taper—same bottle as flight H (below).
  2. estimated final velocity from last five frames is 78 m/s
  3. launch tube used.
  4. green sprite bottle—somewhat efficient bottle taper
  5. estimated final velocity from last 22 frames is 57 m/s
  6. launch tube not used. Notice the water spraying up from the launcher after the rocket is clear due to residual pressure in the air line.
  7. water volume estimated 3 mL due to leaking during pressurization.

Horizontal launch

Video link & file size

333 KB

Flight H
Rocket volume 1550 mL
Rocket mass 103 g
Water volume 259 mL
Pressure 90 psi
video speed 2000 fps
Length scale 124 pixels/m
Comments below 1, 2
  1. recording this video required some improvisation. The foam backstop was relocated to a far wall. The lights were moved to new locations. There were various pieces of equipment 0.5 m away from parts of the flight path as the rocket crossed the room. I had to very carefully (!) aim the rocket so as not to hit this equipment. I decided not to press my luck more than once.
  2. See my later attempt to perform a horizontal launch that does not have the flight-path-perturbing water wave.

Gas pulse

I merged 5 separate flights into one movie, in order to have a virtual race between the rockets. Four of the five flights had a capful of water added. I intended for the helium-filled rockets to be 100% helium. Only when it was too late did I realize that I had not flushed out the air that was present in the bottle and line before pressurizing the bottle with helium.
Video link & file size

191 KB

Flight I J L K N
Rocket volume 2100 mL
Rocket mass 95 g
Water volume 7 mL 8 mL 7mL 7 mL -0-
Pressure 88 psi 51 psi 51 psi 88 psi 88 psi
Gas He + air He + air air air air
Effective MW (g/mol) 7.6 9.6 29 29 29
Gamma = Cp/Cv 1.61 1.58 1.40 1.40 1.40
video speed 2000 fps
Length scale 77 pixels/m
Comments below 1, 2, 3 1, 2 1, 2 1, 2 1, 2, 4
  1. I was surprised at what a close finish there was between the He+air and air flights. Some work needs to be done to explain this.
  2. It is also interesting to compare the moment of internal fog formation in each of the flights.
  3. From the small size of the water plume in the video, the volume of water in this rocket appears to be less than the 7 mL added. There may have been leaking of water during pressurization (that's the best explanation I can come up with).
  4. estimate of final velocity from last four frames is 67 m/s

Analysis of water-rocket thrust

The thrust phase of a vertical water-rocket flight is composed of the following general parts:

  1. Launch tube impulse - If a launch tube is employed, then it acts as a piston with force equal to bottle pressure times cross-sectional area of tube. In flight O I got a pretty good picture of what happens as the launch tube slides through the liquid. In the wake of the tube a hole starts to form in the liquid, but the hole does not reach to the throat (it possibly could with lesser amounts of water). The high acceleration of the rocket strongly favors a flat interface between the liquid and gas. Also, when the rocket gets to the end of the tube, a small “explosion” is observed. This burst comes from the compressed air stored in the launch tube. Water thrust then initiates.
  2. Water impulse - Water is pushed out of the rocket due to the pressure difference between the inside and outside of the rocket. As the amount of water remaining in the rocket decreases, the rocket mass decreases and hence the acceleration of the rocket increases. If the velocity of the rocket relative to the earth exceeds the velocity of the exhaust relative to the rocket, then the exhaust will have a positive (upward) velocity relative to the earth. This is observed in many rocket flights and explains how the water column chases after the rocket. In some of my movies you can see the rocket coming down after rebounding off the ceiling while at the same time the water column is still on its way up!
          A phenomenon that appears near the end of the water phase but while there is still water in the bottle is a set of plumes that radiate outward and upward from the water column. These plumes are amazingly symmetric. I believe they are caused by instability in the flow associated with a phenomenon called “vena contracta.“ Vortices are shed from the wall of nozzle and these grow in size as they progress down the water column outside the bottle. This hypothesis is supported by the fact that the plume is larger when the bottle has a stronger taper at the throat. Compare the flights on this page to the flight of a strongly tapering water-cooler bottle.
  3. Air (or gas) impulse - The exit of the pressurized gas at high velocity is sometimes thought to be an instantaneous pulse. However, for 2L rockets this phase actually lasts for about the same amount of time as the water-impulse phase. Because the nozzle is converging only, the exit velocity of the gas is “choked” or limited to the speed of sound. The gas exits the rocket at a pressure exceeding atmospheric, thus it continues to expand after exiting. The exiting air dramatically impacts the previously expelled water, creating a large cloud burst. As the bottle depressurizes the internal temperature drops (due to adiabatic expansion) and eventually the water vapor contained in the gas condenses. This results in a fog trail observable behind the rocket in some of my videos and photos.