A USB controlled servo water gun. Great for firing at unsuspecting passers by, or keeping people with annoying questions at bay.

This project is a small water pump mounted on top of a servo for directional firing. The whole thing is driven by a microcontroller, and controlled from your keyboard over USB.

To see more of our projects and free video tutorials check out our website http://www.nerdkits.com

Step 1: Gather the Materials

This project is microcontroller based. Other than the ATmega168 microcontroller included in the USB NerdKit. For this project we used the following:

1 Hobby Servo, Hitec HS-50
1 Low voltage piston water pump
1 Small n-channel MOSFET, 2N7000

Step 2: Assemble the Circuit

The first part of our circuit just connects to the servo. This is simple here: one wire from the microcontroller to the servo. There are a few different color labelings depending on the manufacturer, so check before trying this.

Schematic Photo of the ServoSquirter circuit on the NerdKits breadboard

The second part of the circuit allows the microcontroller to switch the pump motor on and off. The ATmega168 chip itself only allows 40mA max into or out of any pin, but our pump requires closer to 1000mA! So in order to control this bigger load, we've chosen to use a bigger transistor, the 2N7000. First we explain the basics of using MOSFETs (Metal Oxide Semiconductor Field Effect Transistors) as switches: bringing the Gate voltage above the Source, we can allow current to flow from Drain to Source. From the 2N7000 datasheet, we've extracted Figure 1, which shows the relationship between drain current and drain-source voltage for different gate-source voltage settings. There are a few important things you can learn from this graph:

1. For VGS below about 3.0 volts, no current is allowed to flow. This is the off state, also called "cutoff".
2. For small VDS, the curve looks roughly linear through the origin -- which means it electrically "looks" like a resistor. The equivalent resistance is the inverse slope of the curve. This region of MOSFET operation is called "triode".
3. For larger VDS, some maximum level of current is reached. This is called "saturation".
4. As we increase VGS, more current is allowed to flow in both the triode and saturation modes.

And now you've actually learned about all three modes of MOSFET operation: cutoff, triode, and saturation.

Because our gate control is digital (+5 or 0), we're only concerned about the curve highlighted in yellow, for VGS=5V. Normally, using a MOSFET as a switch generally involves the triode mode of operation, because the MOSFET dissipates power PD=ID*VDS, and a good switch should dissipate little power in the switch itself. But in this case, we're dealing with a motor, and motors tend to require lots of current (with little voltage drop) when they're first starting up. So for the first second or two, the MOSFET will operate with high VDS, and will be limited by its maximum current -- about 800mA from the red dashed line we've drawn on the datasheet. We found that this wasn't enough to get the pump started, so we used a little trick and put two MOSFETs in parallel. This way, they share the current, and can effectively sink about 1600mA together.

Also due to the high power requirements of the pump, we used a wall transformer with higher current output. If you have a wall transformer with greater than 5V output -- maybe 9V or 12V -- then you ca

Step 3: Set Up the PWM on the MCU

PWM Registers and Calculations

In the video, we talk about two levels used by the timer/counter module: the top value, and the compare value. Both of these are important in generating the PWM signal you want.

But to activate your ATmega168's PWM output in the first place, we have to set up a few registers. First, we select Fast PWM mode with OCR1A as the top value, which lets us arbitrarily set how often to start a new pulse.

Then, we set the clock to run with a pre-division of 8, which means that the counter will increase by 1 every 8/(14745600 Hz) = 542 nanoseconds. Since we have 16-bit registers for this timer, this means we can set our overall signal period to be as high as 65536*542ns = 36 milliseconds. If we used a bigger division number, we could make our pulses farther apart (which doesn't help in this situation), and we'd lose resolution. If we used a smaller division number (such as 1), we wouldn't be able to make our pulses at least 16 milliseconds apart, as our servo expects.

Finally, we set the Compare Output mode for a "non-inverting" PWM output, which is described in our video. We also set the pin PB2 to be an output pin -- not shown here, but it's in the code.

Click to enlarge these shots from pages 132-134 of the ATmega168 datasheet, with our register value selections highlighted:

Step 4: Program the Microcontroller

Now it's time to actually program the MCU. The complete source code is provided on our website http://www.nerdkits.com/videos/servosquirter

The code first sets up the PWM to drive the servo. The code then just sits in a while loop waiting for user input. The characters 1 and 0 turn the MCU pin that is connected to the pump transistor on or off. This will turn the pump on and off giving us the ability to fire at will.

The code also responds to the '[' and ']' keys these keys will increase or decrease the compare value on the PWM pin, which will cause the servo motor to change position. This gives you the ability to aim before firing.

Step 5: Serial Port Communications

The last step is to set up the computer so you can send the commands to the Microcontroller. In the NerdKit, we use the serial cable to send commands, and information to the computer. It is possible to write simple programs in most programming languages that can communicate over the serial port to the NerdKit. However it is much simpler to use a terminal program to do the serial communication for us. This way you can just type on the keyboard, and see the response from the NerdKit.

Windows

If you're using Windows XP or earlier, HyperTerminal is included, and should be in your Start Menu under "Start -> Programs -> Accessories -> Communications". When you first open HyperTerminal, it asks you to set up a connection. Cancel out of those, untill you are at the main HyperTerminal sceen. You'll have to set up HyperTerminal, choosing the correct COM port, and setting the port settings appropriately to work with the NerdKit. Follow the screenshots below to get right HyperTerm setup.

If you're on Windows Vista, HyperTerminal is no longer included. In this case, go download PuTTY (Windows installer). Use the connection settings below to set up Putty, using the proper COM port.

Mac OS X

After entering the Terminal application, type "screen /dev/tty.PL* 115200" to start communicating over the serial port.

Linux

On Linux, we use "minicom" to talk to the serial port. To start, run "minicom -s" at the console to enter minicom's setup menu. Go to "Serial Port Setup". Set the parameters as follows:

Minicom configuration on Linux

Then, hit escape and use the "Save setup as dfl" to save settings as the default. You should now be able to hit "Exit" and use minicom for talking to the NerdKit.