3 Ideas
Sketch Model Review
Mockup Model Review
Assembly Review
Tech Review
Final Presentation

Archimedes's Steam Cannon


Think that a half-cup of water is harmless?
Think again.

Watch this video, and don’t even consider trying to do the same thing at home.

A one-pound projectile traveling at the speed of sound with 10 times the kinetic energy of an AK-47 bullet is not a toy. This experiment was performed under carefully designed and controlled conditions.

Curious? Read on.

(video also available as:
.mov 3.3 MB, .mov 11.7 MB, full resolution dv)


Don't see the video?
Get Flash player first.

Click on each picture to see a large version.
Archimedes's steam cannon, from
Leonardo da Vinci, Volume II
Istituto Geografico de Agostini
Novara 1956.

In episode 54 the MythBusters took on the myth of Archimedes’s steam cannon, and after a gentle lob of a few feet, the myth was busted. A large boiler-driven cannon was used for the show’s finale.

Being frustrated with the outcome, having a visit to MIT in their schedule, and remembering last year’s Archimedes’s death ray collaboration, the MythBusters asked us to take a look. They wondered if there was a simple, plausible solution compatible with da Vinci’s vague drawing and explanation of Archimedes's steam cannon.

The drawing has been interpreted as a direct injection concept, where water rapidly sprays into a hot breech and the resulting flash of steam launches the projectile.

But, some back-of-the-envelope calculations suggest otherwise. A 540 °C  (1000 °F) copper breech can provide a heat flux of roughly 100,000 kiloWatts per square meter to water that is in direct contact with its surface. And if water can be sprayed onto the breech in a uniform 1mm thick layer, 2,000 kiloJoules of energy per square meter of surface will produce 400 PSI steam—a pressure that would generate a reasonable projectile velocity.

direct injection concept sketch
direct injection test

The pressurized steam must be generated quickly, before the projectile exits the cannon, perhaps in 0.005 seconds. But at 100,000 kiloWatts per square meter, the breech can only provide around 500 kiloJoules per square meter in such a short time, not nearly enough energy to produce high-pressure steam.

Our conclusion? Direct injection is not likely to produce much more than a burp (.mov, 0.23 Mb).

In 1981, Sakkas (see figures 23-25) recreated Archimedes’s steam cannon and incorporated a wood board into the concept to hold the projectile in place while some pressure developed, although da Vinci’s drawing does not show a similar feature. Even so, this produced a fairly low velocity projectile (perhaps 70 mph) that traveled only 50 meters.

The Archi
MITes's steam cannon

Is there another way to interpret da Vinci's drawing of Archimedes's steam cannon?

We think yes. 

In fact, the solution is so simple, effective, and dangerous that we have decided not to disclose the details.


mit cannon design
cannon shell

The design specifications and calculated performance estimates for our small-scale cannon were:

  • barrel length:  2 feet
  • barrel bore: 1.5 inches
  • volume of water: ½ cup
  • projectile mass: 0.5 kg (1 pound, 1 oz)
  • estimated firing pressure: 3,500-4,000 PSI
  • estimated muzzle velocity: 280 m/s (630 mph)
  • estimated range: 1,200 meters,  shell drag coefficient of 1.1

But, how did the cannon really work?

The test

November 5, 2006

Test day began at 6:05 AM with the MythBusters, approximately 30 students and a two-hour bus trip from MIT to the test site, complete with breakfast and a showing of episode 54.

bus trip
cannon fire pit

For safety reasons, the cannon was placed below ground in a cinder block and steel lined fire pit.

cannon test berm
remote camera setup We observed the firings, which were aimed directly into a sandy berm, on remote monitors behind a hill 150 feet away from the cannon.
image from remote camera
safety zone for testing

After explaining the design, two different cannons were tested.

Two rounds were fired from the first cannon—a design simpler than da Vinci’s but using the same concept.

Then, a second cannon that mirrored da Vinci’s drawing was tested.

Both designs performed comparably.

mytbusters hear about cannon design
high speed cannon animation of shooting

The time needed to reload and shoot was roughly 2 minutes.

High-speed camera footage showed that the muzzle velocity was over 300 m/s (670 mph)!

The shell’s kinetic energy was calculated to be approximately 23,000 Joules — 1.3 to 1.8 times the energy of a 0.50 BMG caliber bullet fired from a M2 machine gun.
shell imact animation

On impact, the thick nose of the shell was driven into its body and it was filled with soil, packed to the consistency of concrete.

shell in berm expended shell sectioned shell
MIT da Vinci cannon

Our conclusion?

We can’t say if Archimedes did it, but this small-scale test shows that a simple, powerful steam cannon could probably have been made using technology of his era.

And in tribute to da Vinci (since they are his drawings after all),
Mito Pericoloso!

Watch it again!

Interested in other 2.009 estimation exercises? See the demise of the Pony Express.