Select a Catagory:
Breaking News
This area does not yet contain any content.

Science Snacks

The Exploratorium's Teacher Institute calls small science activiites "Snacks". These activities are tried out with students and teachers and have all the information you need to make them work. The instructions are 100% free, and we don't sell parts or kits. If you need help finding something, well, that's part of my job.

These days I work for the Exploratorium, so much of my work now ends up here. The work doesn't have my name on it, since we work collaboratively, but some I helped bring to the page. 

Some of my favorites include

Modulated LED



Siren Disk


Balance Boards


Over Your Feet

for a Science of Dance workshop

Things don’t fall when their center of mass is over their base

Recommended Grade Level: 6 to 12

NGSS Science & Engineering Practices

       Asking questions and defining problems

       Developing and using models

       Planning and carrying out investigations 

Time: 30 minutes in class, 2 hours to make the first time

Materials Needed

(per group)


  •  3 12 x 24 x ¾ inch plywood boards
  •  1 x 12 x ¾ inch plywood strip (You may be inclined to use some other wood, but plywood is especially resistant to splitting.)
  •  ½ x 12 x ¾ inch plywood strip
  •  8 1 ½ inch drywall or wood screws
  •  12 inches 2 or 2 ½ inch PVC pipe
  •  24 inches 2 x 2 inch (actually sized 1 ½ x 1 ½ ) cut into two 12 inch pieces


Background Information

When dancers balance they need to keep their center of mass over their feet. When they are on they are on their toes or hands, this becomes especially hard since the area is much smaller. This is a way for non-dancers to get a sense of how difficult that is and is similar to a way that some dancers practice.

To Do and Notice

1.   Attach the 1 inch wide strip of plywood to the center of one of the plywood boards using drywall or wood screws. Drilling 1/8 inch pilot holes in the strip before you put in the drywall or wood screws will greatly decrease the risk of splitting the strip. Using wood glue along with the screws will make a stronger, longer lasting joint. Twist the screws enough to set them below the surface to avoid marking you floors. Repeat with the ½ inch strip.


2.   Try to balance on one board then the other. How are the two boards different?


If you need to save space, you can build a single balance board with one strip on each side.

3.   What are you doing to keep your balance? If you use the expression “shifting your weight”, what do you actually mean by this in a physics or biomechanics sense?

4.   Time to build the last balance board. Find the center of the third board. Place the PVC pipe in the center. Space the two 2x2’s so that they hold the pipe in place. Mark the edges of the 2x2’s on the board. Flip the board over and drill two 1/8-inch pilot holes for each of the 2x2’s. Screw in the drywall or wood screws from the topside into the 2x2’s. It won’t hold the PVC pipe from falling out, but it will keep it from rolling.

5.   Have a person try to balance on the PVC pipe board. Most will find it much more difficult. Why is it harder?


What’s Going On?

·    To maintain your balance you need to keep your center of mass over your base. The wider base is easier since it gives you more chance for error than the narrower base. Once you get your center of mass at rest over the base, you will stay balanced as long as your center of mass stays there. This is called “passive stability”.

·    The pipe is much harder to keep your balance on because its “base” is very, very small, so there is virtually no way to keep your center mass over the base. You will be on one side of the base or the other, and you will have to move actively your center of mass back and forth. This is called “active stability”.

·     Dancers have to maintain their balance over their feet. So since she extended her right leg and to a smaller extent her right arm, she needs to lean to her left to make sure that her center of mass stays over her feet. Notice that her left arm is straight up. Why doesn’t she want to extend it away from her body (to her right) like her left arm?



Going Further

Surprisingly, you can make the pipe board harder to balance on by making the pipe smaller. If you make the board with ¾ inch PVC pipe instead of 2 ½ inch, it will be the same height as the other balance boards. And it will be nearly impossible. There are several reasons why it is more difficult but the main one is distance and time. With a smaller diameter pipe, the board end of the board will be closer to the floor, meaning that it will touch sooner. This gives you less time to correct for being off in your balance.


Ballet picture by Harclade, Used with Permission.






Cyan, Magenta, and Yellow Markers That Mix Together to Make Green, Blue, Red, and Black. Really.


  • Inkjet Printer Ink Bulk Refill (like this pack of BCH inks from Amazon) Most of the packs come with 100 mL of cyan, magenta, and yellow ink as well as 300 mL of black ink.
  • Crayola Marker Refill Kit (like this kit from Amazon) The kit comes with 12 blank markers.
  • Small (approximately 10 mL) clear plastic cups.

 Set Up

  1. Discard the ink that comes with the Crayola kit. It works okay, but the bulk printer ink is much purer in color.
  2. Each pen consists of a barrel, a nib, a foam cylinder, endcap, and regular pen cap.
  3. Place the nib into the barrel and then insert the foam cylinder. Use a pen to push the foam onto the nib.

  4. Open the ink bottles. You won’t need the black ink. You can set them aside.
  5. Using a long needle and syringe that came with the ink, extract 2-3 mL of ink from the cyan bottle.
  6. Press the needle just below the surface of the foam and push down the plunger slowly.
  7. You’ll see the ink move down the tube via capillary action.
  8. When the ink reaches the nib (usually takes a few minutes), test to be sure the pen works. If so, press the endcap onto the barrel. You’ll hear an audible click. If the pen doesn’t work, you should add another milliliter of ink.
  9. Repeat with magenta and yellow.
  10. Wash out the syringes and needles with plenty of water.

Do and Notice

  • Make three overlapping circles of cyan, magenta, and yellow to make seven regions. Notice that where cyan and yellow overlap, you get green; where cyan and magenta overlap, you get blue; and where magenta and yellow overlap, you get red. Where all three overlap, you get black (or at least a really dark color).

  • What color would you get if you mixed yellow and blue? You don’t have a blue marker? No big deal. Make one. In a small cup add 2 mL of cyan ink and 2 mL of magenta ink. Mix. Add that to a blank marker.
  • Place 10 mL of water in a clear cup. Add a drop of yellow ink. Add a drop of magenta ink. Mix. Note the color. Add another drop of magenta ink. What color is that? Notice that you can mix different inks together to get many different colors.
  • Notice that adding ink always reduces the amount of light getting through, adding water allows more light to get through.

What Is Going On

Michelangelo was once asked (perhaps, appocyphorally) how he carved David. He said that he started with a block of marble and removed everything that wasn’t David.

Let’s say that you want a green splotch on a piece of paper. You are probably doing this in white light on white paper. (You might consider how they’d be different if you were in green light.)

White light contains all hues that are available to your eye. So, you already have green light hitting the paper. However, you also have red, yellow, chartreuse and many, many more hues along with the green light. You need to subtract all the other colors.

Is there an ink that subtracts the red to yellow region of the spectrum? Yes, that color is called cyan. Pure cyan would only let blue to green through. So if you put cyan on the paper, only blue and green light could be reflected.

However, you don’t want blue and green light, you want just blue. Is there a color that removes green blue light? Yes, that color is yellow. Yellow lets green to red through but absorbs other colors.

So, if you put a layer of cyan down and a layer of yellow down, only the green light from the all the hues in the original white light can get through.

Only red can get through magenta and yellow, and only blue can get through cyan and magenta. No color can get through all three.

More Advanced

What if you want orange? Orange is in between red and yellow. We want all the red light from the white light, but only part of the green light, and none of the blue light. Pure yellow ink will block all the blue light so we’ll use it. If we use dilute (or only cover half the surface) with magenta ink, it will block only some of the green light. So, if we use pure yellow and dilute magenta we should be able to make orange ink. Other inks can be generated by combining different amounts of yellow, magenta, and cyan inks.




Heating by Freezing: Freezing Lauric Acid

As liquids cool (lose kinetic energy), their temperature goes down, as you might expect, but when they reach their freezing point and turn from liquids to solids, they continue to lose kinetic energy to their surrounding but their temperature starts to level off and reaches a plateau until the material becomes a solid.

How do I know that the system is still losing heat if the temperature isn't going down? The surrounding area is heating up! If you measure the surrounding area, you will see that it's temperature keeps going up even though the temperature of the liquid turning into a solid is staying the same.

Water does this in a freezer and a thermocouple probe that you snake out the door and stirrer can show it. Boston in the winter works as well. Still, students often thing that this a special property of water and not generally true of liquids. So, it would be good to have another liquid to show.

Lauric Acid is a common additive to soap and candles and is easily and cheaply found online. (eg

Lauric acid melts/freezes at 43 C or 110 F; so, it makes a good additional example to water to show freezing curves.


(per lab group)

  • 16 grams lauric acid
  • Test tube (large)
  • 2 thermometers (digital or analog)
  • Styrofoam cup
  • Ice water to fill the cup
  • Holder for the cup (i.e. ring stand, larger beaker, cup filled with sand)
  • Graph paper and paper for notes
  • Stopwatch
  • Test tube tongs/holder or hot gloves

 (per room)

  • hot plate
  • 250 mL beakers (three or four test tubes will fit in each beaker)
  • thermometer

Set Up

  1. Fill a large beaker with ice and water and allow to come to 0 to 5 C.
  2. Put 16 grams of lauric acid into each tube.  I usually make a couple of extras.
  3. Place test tubes into a beaker. Each will hold about four test tubes.
  4. Fill the beakers with water so that they are a water bath until a couple of centimeters below the rim.
  5. Place the beakers on a hotplate. Put a thermometer into the water. Allow the water to reach 80 C.


  1. Use a ring stand, a larger beaker, a cup partially filled with sand, or some other method to keep the Styrofoam cup from tipping over.
  2. Fill the Styrofoam cup with 0 to 5 C water. Do not put any ice into the water.
  3. Place a thermometer into the water. Record the temperature.
  4. Place a thermometer into the lauric acid test tube. Record the temperature.
  5. Make a data table with 30-second intervals to record the temperature of the lauric acid as it cools and the water as it warms.
  6. Place the test tube into the water. Work carefully and quickly.
  7. Record the temperature of both the water and the lauric acid each 30 seconds. A good plan is to divide up the work. Continue until the water reaches 30 degrees or 20 minutes has passed.
  8. Plot the data. (Sample data at the end)

Things to Notice

As in most phase change situations, there is a plateauing region on the graph. This is where the lauric acid is freezing – going from a liquid to a solid. While the temperature of the lauric acid is barely changing, notice that the water is still heating. Compare this heating curve to the curve after the lauric acid is a solid. Notice how the heating curve at that point is nearly flat, indicating that little thermal energy is entering the water.



Heat Transfer: Radiation, Conduction, Convection Sequenced Activities

I put together a series of activities to teach heat transfer. Here they are in Microsoft Word format.

Many of them use liquid crystal thermometer plastic. They are plastic that change color as they change temperature. You've probably seen thermometers that are made this way. You can get them from a few places, but I am big fan of the 12 inch x 12 inch squares from Edmund Optical. They work well and one sheet is enough for a whole class.

What's great about liquid crystal thermometers is that they require very small temperature changes to have a noticeable change.

Activity H1.01 Five plates – aluminum, plastic, wood, brass, and steel -- of similar dimensions are arranged as shown above. An identical ice cube is placed on each plate with a rubber o-ring keeping the ice cubes from sliding off.

Activity 2.01   You have a white index card with a liquid crystal thermometer on the top. Discover what it does when you hold it between a thumb and a finger, let it go and hold it in the air, hold it again and then put the bottom against a table, press your thumb against it while it is on the table.

Activity 2.02    A piece of the liquid crystal thermometer is glued to a thin piece of aluminum. The LC thermometer starts off black. If the end of the thermometer is placed into warm water what will happen to the color of the thermometer?

Activity 2.03   You are given a piece of steel or a piece of plastic with a liquid crystal thermometer glued on. Will it conduct at the same speed as the aluminum?

Activity 3.01   A 100 watt incandescent light bulb is attached to a stand. You hold your index card with liquid crystal thermometer glued to it 15 cm below bulb but facing the bulb. Do not have the index card touch the table. How will the liquid crystal thermometer change, if at all?

Activity 3.02   Place the index card against the table so that the liquid crystal thermometer turns black again.  Hold your index card with liquid crystal thermometer glued to it 15 cm below bulb but facing the bulb. Do not have the index card touch the table. Quickly place a pen above the card so that it casts a shadow on the card. Don’t have the pen touch the card. How will the liquid crystal thermometer change, if at all?

Activity 4.01   A 100 watt incandescent light bulb is attached to a stand. You hold your index card with liquid crystal thermometer glued to it 15 cm ABOVE the bulb but facing the bulb. How will the liquid crystal thermometer change, if at all? Will it be faster or slower than before?

Activity 4.02   A 100 watt incandescent light bulb is attached to a stand. You hold your index card with liquid crystal thermometer glued to it 15 cm ABOVE the bulb but facing the bulb. Place a piece of acrylic above between the card and the bulb. Don’t touch the card with the acrylic. How will the liquid crystal thermometer change, if at all? Will it be faster or slower than before?

Activity 5.01   You put on a rubber glove. You dunk your hand in ice water and determine roughly how long it takes before your hand gets to cold hold underwater. You cover your hand with Crisco and place it underwater again. Will you be able to hold your hand underwater for less time, the same time, or for more time?


Converting a Computer (ATX) Power Supply into a Power Supply You Can Use in Lab

Many PC desktop computers use a power supply known as ATX . These computer power supplies are surprisingly well made, nearly bulletproof, since computer components are notorious for being easily damaged by getting voltages over or under what they are expecting. Manufacturers go to great lengths to make them durable and close to specs.

And the specs are often fantastic. Most can supply 3.3 V up to 10 A, 5 V up to 20 A, and 12 V up to 10 A. Most have several output wires, so you could have several students use the same supply as long as they together stayed under the current limits. Best of all, most will automatically turn themselves off if your circuit tries to draw too much current. Unplug the power supply, wait 30 seconds or so, and plug it back in, and the supply will be as good as new. No more replacing fuses!

Since power supplies rarely go bad in home computers and they are easy to extract, you can have a ready source of power supplies for more or less free. Surplus houses often sell them for only a few bucks each, if you’d like to go for new.

Their only real downside is that they don’t supply variable current, only fixed levels, but if 3.3 V, 5 V, and 12 V work for you, then they are great.


  • ATX power supply
  • IEC power cord
  • scissors
  • soldering iron
  • solder
  • electrical tape or shrink wrap
  • wire stripper
  • two incandescent holiday lights
  • four alligator clips with vinyl jackets


  1. Cut off the connectors from the ends of the power supply using a pair of scissors. Leave the wires long, though. You can tie them into a knot to keep them out of the way.

  2. Find the green wire (there is usually only one) and a black wire (there are usually several). Strip the black and green wires and solder together. When the green wire is connected to ground – say by pressing a button – that signals the power supply to turn on. While you could connect a switch to the supply here, it is easier to turn it on permanently and turn it off by unplugging it.

  3. Some power supplies will not turn on unless they are powering something. That is, if nothing is connected to the wires, it won’t start. To solve that problem (and as bonus give a telltale that the power supply is working), strip the ends from two incandescent holiday lights and an orange wire and a black wire.

  4. Solder the two incandescent bulbs together in series.

  5. Solder the pair to the orange black wires. Cover the connections with electrical tape or shrink wrap. 
  6. Strip one orange, one red, one yellow, and one black wire.

  7. Solder an alligator clip to each wire. Remember to slide the jacket up the wire BEFORE soldering the wire to the clip. Hold the clip and wiggle the jacket over the clip.
  8. If you’d like, you can strip another set of black, orange, red, and yellow wires and solder on alligator clips. Most supplies can give enough current to run two stations. However, the wires are shortish this way, and if one group draws too much current, both groups will be shut off.


  • Orange to black = 3.3 V
  • Red to black = 5 V
  • Yellow to black = 12 V

Inside a baseball

Had a workshop teaching Exploratorium Explainers about how different balls worked. This is what's in a baseball. Cork core, two kinds of rubber, 150 yards of wool yarn wound around the core, a string mesh to hold everything together, and a leather cover. Well, that's what's in a MLB-style ball. There are also more boring balls that are made of solid ball of cork and rubber. Click for a full-size view.


Making Your Own Glow Paper

Phosphorescent paints are kind of horrible, but glow powders work great. The hard part is getting them to stick. Tammy Cook-Endres suggested Elmer's Glue, and it works great. Washes off with water but dries strong and most importantly, clear. Make a slurry of the pigment and the glue and let to dry. Glow, Inc on Amazon sells a variety of the powders. They seem expensive but they are not so bad as 1/4 ounce is actually a lot phosphorescent powder for our purposes.

Fluorescent paint works better and is cheaper. You can get it from many paint supply stores. Bring a handheld UV lamp if you'd got one to try them out first. These seem pretty good. 


Laser Pointers and Finger LEDS


Amazon more or less doens't sell laser pointers any more as they had issues with sellers violating FDA rules on power. To get them now, you will have to find your own way. Laserponiter Pro seems like an okay source, but they also sell crazy, dangerous laser pointers. Stick to 5 mW and less to be safe and realize that even then, no one is certifiying that the laser are low power. Keep the spots moving and make sure that the beam never goes into an eye. Watch for mirrored surfaces.

Small LEDs though have become easy and cheap. In addition to making your own, LEDs finger toys are pretty common and work fine. Amazon sells a color assortment of red, blue, green, and white lot of 100 for under $11 plus shipping. Check out



Rotational Inertia Demonstrator

Blue tape arms are filled with sand, hollow arms are on the tableA way to cheat in baseball is to cork a bat. That is, drill out the center of the bat to make it lighter and then fill the bat with cork so that it is less likely to crack when it hits the ball. You might imagine that a player would want as heavy a bat as possible to get the most momentum since momentum is mass times velocity. However, a lighter bat can be accelerated to higher velocity, evening out the lost mass, leaving the momentum close to the same. (See Nathan, 2011)

The key difference is that a lighter bat can turn more easily so that a player can wait just a fraction longer to evaluate the pitch before needing to move the bat. This turns out to be a big edge. With a metal or composite material bat, it is possible to create hollow bats that are much lighter than a wooden bat for the same length. So called "negative bats" have lengths that are larger than the weight of the bat in ounces. Many leagues limit these as to limit the offense.

This device demonstrates that idea. The ends of the arms of the device are filled with sand. They have a torque (a turning force) applied by the falling water bottle. This causes the arms to turn. However, the ends of the arms can be replaced by hollow tubes. When the bottle pulls down on the arms with the hollow tubes, the arms turn much faster.


  • about 8 feet 1/2 inch PVC
  • four 1/2  inch PVC endcaps
  • cross or four way 1/2 inch PVC connectors
  • two 1/2 inch PVC slip connectors
  • 6 inches steel rod, preferably powder coated in a diameter close but smaller than the inside diameter of 1/2 inch PVC
  • PVC ratchetting cutter or hacksaw
  • 30 inches of string
  • water bottle or other weight
  • 1 cup sand
  • funnel
  • tape measure
  • tall ring stand
  • two right angle bar connectors (sometimes called cheeseburoughs) for ring stands
  • blue or black tape, preferably gaffers' tape
  • glue gun and two sticks of glue
  • hacksaw (not shown)


  1. Use the hacksaw to cut the steel bar to about 10 inches long.
  2. Attach the steel bar to the ring stand with a right angle bar connectors.
  3. Cut two pieces of PVC so that when they are attached to the cross connector they almost cover the steel rod.

  4. Attach another clamp to the end of the rod, but be sure that the cross connector can turn freely. PVC has a relatively low coefficient of friction on powder coated steel, so even without bearings it should turn easily.
  5. Cut two pieces of PVC about 20 inches long. The exact length is less important than having them be the same length. Attach the slip connectors. Connect to the cross connector.
  6. Cut four shorter pieces of PVC about 10 inches long. Again the exact length is less important than having them be all the same lenght. Attach the endcaps to all the pieces.
  7. Using the funnel, fill two of the shorter pieces with sand until about 1 inch from the top. Try to have both pieces have the same amount of sand in them.
  8. Fill the rest of the space with hot glue so that the sand is trapped inside.
    Sorry about the focus.
  9. Wait for the glue to cool and set. Mark the pieces with tape.
  10. Attach to the slip connectors.
  11. Fill the bottle with water. Attach string to the mouth of the bottle and tape the string to the PVC. The string should be long enough that the bottle just rests against the ring stand when the string is fully elongated.
  12. Turn the arms to wind up the string. Let fall. Measure the time it takes for the bottle to reach the base of the ring stand. Replace the ends with the other two that don't have the sand. Repeat.

 A couple of terrible videos, but they get across how it works.

With sand (falls for about 30 seconds)

Without sand (falls for about 20 seconds)


Magnus Effect Demonstrator

From Wikipedia

When a spinning baseball, soccer ball, tennis ball, or table tennis ball move through the air, they feel a force that is caused by the spin. The force, called the Magnus Force or Effect, can be up, down, or side-to-side. You can see more about the effect at Wikipedia. You can watch a hopelessly perky video from SciGirl to get more information.

(Adapted a the demostration called Spin Doctor from Dr. Baseball also known as California State University Chico Professor David Kagan.)

Spinning top

The idea is that after you spin the ball in the PVC trough, the ball moves to the side. In this case, looking down, if you spin the ball clockwise, it will slide to the left. If you spin it counterclockwise, it will slide to the right.


Materials (hacksaw not shown)

  • Golf practice ball
  • Wooden disposable chopsticks
  • 9 inches of 1/2 inch PVC
  • Sandpaper (fine)
  • Drill
  • Hacksaw (not shown)


  1. Cut the PVC to 9 inches. A PVC cutter or a hacksaw will work.
  2. Cut the PVC tube the long direction to make a trough. I used a bandsaw for this, but a hacksaw and a vice will work fine.
  3. Put the ball on the chopstick to find where the chopstick fits snugly.
  4. Cut using the hacksaw the chopstick so that it is about 3 1/2 inches long. About 1/2 inch should extend from the bottom and 1 inch from the top.
  5. Take the wooden dowel and chuck it into an electric drill. Tack down the sandpaper. Hold at an angle to make a very shallow point on the chopstick. Much, much shallower than a pencil point. Then round over the top. 
  6. Push the peg into the practice golf ball. You will need to fiddle with the exact amount sticking out on each side. The lower the practice golf ball, the better the top will spin, but it will catch on the trough. Higher, the top will be unstable.
  7. Blow on the ball. You don't want to be too close, 10 to 12 inches is good. If you are too close, you will push the top over.




This doesn't really have anything to do with hands on science. Earlier today, I was replying to an email about magnets and I was searching different websites for products. I found two websites, Boreal and Ward's, to be eeriely similar and yet their prices were really different. Ward's was generally a lot less. You can see that here in this screen shot. Weird.


 Click to enlarge


I guess it pays to shop around. 





LEDs Demonstrate Quantum Mechanics

This is a strip of battery-powered LED Christmas Lights which are usually available at Walgreens at Christmas. I got them out to work on a project, turned them on, and was struck by something odd. I used to buy these by the bucket at the low price of three packs for $10 because it was a cheap way to get fairly bright LEDs that were already shrinkwrapped and easy to use. They've changed the lens design to spread out the light more, but they can still be useful. However, that's not what I'm interested in now.

Notice that the red LEDs are really bright, the yellows are dim and the rest are off. This suprised me and made be take a moment to consider what was going on and come up with an explanation.

A clue is that the batteries are mostly used up and aren't able to provide as much voltage as when they were fresh.

LEDs are weird beasts. Like filament light bulbs, if electrons get pushed through them, they start to glow, but unlike filament light bulbs, each color takes a minimum voltage before it will glow. Less than that voltage, you get nothing.

You might wonder why that is. Alhtough the story is really more complicated than this, essentially the LED turns the energy in one of those moving electrons into a bit of light. The energy that an electron has is related to its voltage.

And the energy of light is related to its frequency. The red part of the spectrum has the lowest frequencies and the violet part of the spectrum has the highest frequencies.

So low voltage electrons have low energy and can only make low energy light which is light with low frequencies which is red light. It takes higher energy electrons to make green and blue light. Since there aren't any high energy electrons, the green and blue LEDs can't make any light.

Perhaps you could put two electrons together to make blue light? That doesn't work in most cases, though. Quantum means more or less "the smallest possible amount."

With fresh batteries, everything started to work as you'd expect.



Getting a sense of precession

Playing with a spinning wheel is fun, but understanding where precession comes from is hard. Vertassium gives it a try, but it is still pretty tough going.

There are a bunch of mathematical ways to think about it using cross products and the like, but Lewis Carroll Epstein gives a good mechanical way of thinking about it in Thinking Physics, and I've expanded it a bit here.

Epstein's ideas is that instead of using a bicycle wheel, let's imagine a much simplified gyroscope. It is a square tube of water, and instead of a wheel turning, the water is flowing in the tube.  I've drawn a picture in SketchUp to show that. Email me if you'd like the 3D file.


Simplified version of a gyroscope

We tie a rope to the left side and let gravity do its thing. The left side will stay in place but the right side will fall.


The water running along the horizontal sides (top and bottom sides or east-west) are largely unchanged, but the water running vertically runs into the wall and bounces off. That creates a force. To be specific, in the west leg, the water runs into the south side and bounces off to the north. That creates a force on the west leg to the south.

from the other side

The same thing happens on the other side but the water bounces to the south so the force is to the north.

Maybe it is easier to see it from above. from above


Dancing Styrofoam Beads (Polystyrene)

Watch the video to see what's going on. 


The acrylic tube is 72 inches long x 2 1/2 inch (outside diameter). I bought mine at Tap Plastic, but you can get it online. The beads are expanded polystyrene (one brand is Styrofoam) for bean bag chairs. The minimum order from Tap Plastic was three cubic feet. You need a lot less, so if you live in the Bay Area, I'm happy to give you some.

I used plastic glue to finish the end away from the speaker.


Connecting the speaker to the tube in an airtight way was a challenge. The threads on the horn driver are different than those on PVC. Grr. Instead I made my own piece. I cut a short length of 1 1/4 inch PVC. I melted some InstaMorph in the microwave and put it inside the PVC. InstaMorph sticks to PVC and before it hardened, I screwed the driver on. Presto! A piece of PVC with the right threads. You don't have to be this elaborate though, it will still work with the PVC pipe just pushed into the threads of the driver.

The other side of the pipe is connected to a PVC 2" x 1.25" reducer bushing. Then I used a rubber pipe coupling to connect between the arcylic tubing to the PVC.


Driver plus PVC connector

You need a lot of amplitude (air motion) to move the beads. I use a horn driver since they are small and can fit into the tube. The horn driver is a Selenium D 250-X that I got from Parts Express for about $40. It is rated at 250 W at 1000 Hz and higher and 150 W at 500 Hz to 1000 Hz.

That last bit is important. As the frequency goes down the amount of power that the driver can handle goes down. Too much power will melt the wires in the driver, and it will stop working.


Burned out. Smells bad too. A nice thing about the Selenium drivers is that Parts Express sells replacements.

The key then is to get rid of any low freqencies from the signal, otherwise you'll melt the driver. For this, I used an electronic crossover which allows some frequencies through and blocks others. I used the Power Acoustik C3184 which has a variable crossover, and I set it to reject frequencies below 600 Hz. Makes the songs sound tinny but allows me to pump up the volume.

Variable crossover

The only problem with this crossover is that it is designed for a car which means that it wants 12 V power. I just use an old computer power supply but you can buy adapters. Instructables has lots of projects that turn computer power supplies into useful bench top supplies.

Computer power supply gives 12 V

Picking the right song can be a challenge. I found that songs that lots of mid and high frequencies work well, and songs with long sustains seem to work better.


Mini-Marshmallow Launcher

because "gun" sounds so violent.

This is a pump action device. The mini-marshmallow is placed in front fo the 1/2 inch pipe. The bottom pipe is pulled out, and the marshmallows get pushed by the outside air pressure (sucked, if you will) deep into the 1/2 inch pipe. The bottom pipe is then forced back in, and the marshmallows are launched forward.


  • 2 feet 1/2-inch PVC (schedule 40 or 80) pipe
  • 2 feet 2-inch PVC (schedule 40 or 80) pipe
  • 2 feet 1 1/2 inch PVC (schedule 40 or 80) pipe
  • 2 inch PVC connector sleave
  • 2 inch to 1/2 inch adapter
  • 2 1/2-inch right angle threaded bend
  • small disk of cardboard
  • duct tape
  • 5 inches of thread
  • yarn
  • glue
  • PVC cement
  • mini-marshmallows



  1. Cut out a piece of cardboard that is the same size as the 1 1/2 inch PVC's cross section. Use duct tape to firmly affix it to the end of the pipe. Wrap duct tape around the end until it is almost the same diameter as the 2 inch pipe. Apply glue to the tape (I used Gorilla Glue) and then wrap with yarn. It should make a snug fit into the pipe. Do the same a bit farther down. Repeat with yarn or use some self adhesive felt. I made snips in alternating sides so that it would lie flatter. Allow to dry.
  2. Dry fit the PVC pipes together using the model above. The small right angle fittings also come in slip fitting rather than thread. I prefer the thread because then I can take it apart.
  3. Place the thread across the 1/2 inch pipe between it and the right angle fitting to make sure that mini-marshmallows can't go all the way into the bent sections.
  4. Once everything is working, use PVC cement to connect the joints.

I didn't come up with this design, I first saw it here. I just adopted it from Nerf darts to mini-marshmallows.


Pressure Cooker Squished Cups

If you put expanded polystyrene (Styrofoam) cups in a pressure cooker and turn up the heat, they will shrink down to approximately 1/6 their original volume.


As I'll explain later, you don't actually need a pressure cooker, though they work the best. I use an electric pressure cooker for the show since it is easier to use. I don't need a hot plate, and the countdown timer shuts itself off, but a stove top one works too.

Even gentle heating is important if you want the cups to keep their shape while shrinking. So, you want the cups out of the water,  and on a flat surface that can take the heat but not conduct it too well to the cups. I use a Pyrex lid to a oven casserole. I lift it out of the water with three crumpled balls of aluminum foil.


Use permanent marker to write on the cups and place them in the pressure cooker. My cooker takes about 20 minutes at pressure to shrink the cups. You can crowd them in the container. They'll stick a little after you're done, but they easily pull apart.

How Does This Work?

A digression for a moment. When submersibles go down into the depths of the ocean, they often bring Styrofoam with them on the outside of the craft. It gets crushed, like in this picture from some cups taken down in the Alvin.

Does the pressure cooker do the same thing? The pressure cooker "only" has twice the pressure as usual. While that is enough to make a dangerous mess if the cooker were to rupture, that's not high compared to what the Alvin does. Two atmospheres is like going down to a depth of 10 meters. The cups get fully crushed on the Alvin at 300 m. Hmmm....

Perhaps it is the heat? Heat makes the Styrofoam soft and bendy. I'm imagining that the heat (and pressure) pops the foam, and the foam collapses back into its unexpanded state.



Siphon Model

A siphon is usually presented a tube that has one end higher than the other and is completely filled with liquid. A Google image search gives dozens of pictures along that line. 

The explanation that often goes with these pictures has to do with atmospheric pressure. That air pressure is slightly higher on the high side (A) and slightly lower on the down side (B). That's true but not the whole story. The difference in air pressure helps to hold the water together and does push it, but a siphon will work without air pressure. 

A better way to think of a siphon is probably good ol' conservation of energy. Though it takes energy to lift the water over the lip (from A to C) of the first side, the energy can be gained can be returned if the water on the other side ends up lower than where it started in the first side. 

The hard part is actually getting the energy back to the first side to lift the next bit of water. Atmospheric pressure helps to hold the water together so it doesn't break into two pieces at the bend. 

If the material can hold together on its own, then the siphon could work in a vacuum. The beaded chain in my video demonstrates this idea. If you pull it over the top and let it hang to the table it. 


Improved Elihu Thomson Coil (Jumping Ring)

My previous Elihu Thomson (I'm pretty sure that it isn't "Thompson") Coil -- worked okay, but my new one works like a dream, though it is a bit expensive to make, mostly because copper costs so much these days.


  • 500 ft spool of stranded (if you can find it) 12 or 14 gauge copper single conductor wire like this
  • sacrificial extension cord
  • heat shrink
  • 1 1/2 inch schedule 40 or 80 PVC pipe (about 14 inches long)
  • drop ceiling wire or iron welding rods (you are looking for thin iron or mild steel wires at least as long as you PVC pipe) I needed about 100 feet
  • lever action mini bolt cutters like one of these
  • polyurethane spray
  • aluminum and copper rings (see below)

Step 1: Find the right spool of copper wire


When the manufacturer coils a spool of wire, sometimes they tuck one end of the wire through a gap in the spool towards the center. This is called the tail, and you want to buy a spool that has this. I've outlined it in red. Most do, but you'll need to check. Sometimes it is tucked into the center of the spool; that will work but it's less convenient. 

Step 2: Expanding the hole.

The spool will have a small hole for hanging on a bar, but you want a bigger hole. Use a hacksaw blade with tape around one end as a handle. Cut the hole so that it matches the outside diameter of your piece of PVC. 

Step 3:

Cut the outlet end of the extension cord off. Strip some insulation off the cord. Do the same to the wire on the spool.

Slip heatshrink onto the extension cord and slide down. Attach one wire from the spool to one wire from the extension cord. This needs to be a strong attachment. I used the NASA standard for splicing since it strongly resists coming apart. Solder the one end of the coil to one wire of the extension cord. Do the same to the other wire. Pull the heatshrink down and use a heatgun to shrink. You may notice that I used several layers of heatshrink to get the insulation to the same thickness on both ends.

This is kind of important. You are going to plug this into the wall with 120 V available. You don't want to touch 120 V.

Note: Some people will want to connect a momentary or push button switch instead of just plugging the coil into the wall or using a power strip. I understand. It is possible that the launcher will overheat and there is live wall current. I didn't do that, but you might have a different standard of safety.

Step 4: Cutting the wires.

Cut the steel wires to the length of the tube. Standard wire cutters won't work. You'll need the mini bolt cutters. Even with a jig, I find it hard to cut them all to the same length.

Spray them with polyurethane to make them insulating. We want to minimize eddy currents in the wires. 

Push them as tightly as you can into the PVC tube. I needed a mallet. The wires aren't held in with glue. Friction seems to be enough.

Step 5: Buy some rings 

Online Metals seems to be a good place to order aluminum or copper tubing. If you go with 1.5 inch PVC then 2.25 inch outside diameter aluminum tubing works well. I found that the 0.125 wall thickness tubing fired the highest, but I also got some thicker 0.25 inch tubing. They will cut to length, and I got 2 inch, 1 inch, and 0.5 inch long pieces. I also bought equivalently sized copper to show that though it is a better conductor, it doesn't fly as high. The copper is expensive, though.

The rings will come with sharp edges. A knife blade will clean them up but a deburring tool works better.



Improved Ping Pong Ball Vacuum Cannon

I've been planning on improving my Ping Pong Ball Vacuum Cannon for several years, but never really gotten around to it. A couple of months ago, I started on the redesign. Most of the changes were small, but overall they worked out well in a slightly shorter device.


A Ping Pong Ball is placed at the end of aPVC pipe. The ends are covered with mylar. A soda can is placed near the other end. The pipe is evacuated. The far end is ruptured. The Ping Pong ball exits and rips a hole in the can.

Making the Cannon

First, I switched to transparent PVC. The stuff is expensive ($40+) and is not at your ordinary hardware store. Grainger sells it, though, and I picked up this 8 foot section at a local store in Arlington, VA. (I'm on sabbatical this year.) I also got a transparent connector. You can use opaque, but transparent is cool, especially for high speed cameras.

To connect to the vacuum pump, I drilled a hole 4 inches from one end using ???? which does a good job making a clean hole even in PVC. Using a pipe wrench I twisted in a short 3 inch long, 1 inch diameter brass pipe with tapered threaded ends. The pipe itself cut threads in the relatively soft PVC. This seals pretty well but not perfectly. After everything was done (and tested), I kneaded some plumbing epoxy and pressed it around the pipe to give the connection some stability and make a slightly better seal.

I screwed a 1" brass gas valve to the pipe using Teflon tape. I close the valve before firing to protect the pump from in-rushing air. I attached a brass nipple to the other side of the valve also with Teflon Tape.

Tubing and Pump

The red hose is vacuum tubing. I had some given to me, but you can buy it here from Flinn Scientific.

This gets connected to a vacuum pump. I've been using scientific pump but didn't bring it with me from SF. I bought this small pump from Harbor Freight Tools for $150, and it has worked like a champ so far.

I use one star (cheap) standard 40 mm Ping Pong Balls and pieces of emergency blanket (mylar) to seal the ends. I got the idea from folks at BYU, but I see the link is dead. Sigh. What I do is cut a 10 cm x 10 cm square and pull it over the end so it fits smoothly. Here's a picture.

Then I push a PVC connector over to seal. Love this method. It seals better than tape, is much faster to apply, rarely gets sucked into the pipe, and when it does, it doesn't get stuck. 

Can Holder

I used to use a bar clamp to secure a can to a lab jack, but I've upgraded and now have cut a holder.

The holder uses the fact that aluminum cans have a slanted top and bottom (a frustum) so that the can can be wedged and held in place.

To make the holder,


  1. Cut two squares of wood approximately 6 inches to a side and 3/4 inch thick. 
  2. Cut a two inch hole in the center of each square using a hole saw. 
  3. Stack the two boards together so that the holes line up as close to perfectly as possible. Clamp together.
  4. Drill two 1/2 inch holes through both boards on opposite corners. I drilled four holes, but I've since discovered that two work just fine and makes it easier to clamp. 
  5. Cut using a hacksaw two pieces of 3/8 inch threaded rod about 7 inches long. 
  6. Unclamp the boards. Insert and pound in two 3/8 inch t-nuts in the orientation in the picture. 
  7. Place the can so that the frustums slip into the slots. Tighten with two 3/8 inch wingnuts. 


The fired ball has a lot of energy but not much momentum, so it isn't probably going to knock the holder over or move it far, but metal and pieces of ball can fly after the collision. I put up an explosion shield. It's overkill, though. 

Some people make a holder that holds three cans. The ball tends to pass through the first two cans and get caught by the third. That works well, but I think that it's less dramatic than seeing a single obliterated can.