Whack-A-Cap: Miniature representation of a popular amusement game

by Gigi T Lam & Euborn Y Chiu

Introduction:

Our final project code calls for the implementation of an amusement game often bannered as "Test-Your-Strength" or less accurately (but more commonly) known as "Whack-a-Mole." Our machine is in essence a miniturized version of what can be found in most theme parks across the world. Its construction involves a spring load plastic cap to act as a surface for contact, a wooden mallet which the player uses to hit the plastic cap, and most importantly, a device, called an accelerometer, to measure the shock forces created by the impact of the mallet and the cap surface. A potential player would proceed to hit the plastic cap with the mallet provided, and depending on the level of force registered by the accelerator, a winning or losing result will be annouced visually through a liquid crystal display and light emitting diodes, as well as musically through the use of a speaker.


An 8535 development board was choosen for this project because we require an analog to digital conversion process to capture voltage values (as a function of g forces experienced) as they are registered by the accelerometer. Having decided on a chip to use for our project, we proceeded to the programming stage. Our approach was to design a state-machine with three states. The following paragraphs will explain the functions and reasoning behind each state.

Machine States:

Upon power up, our machine enters a welcome state. In this state, the following events occur: First, a series of tones are played through the speaker, and each note has an associated LED that lights up accordingly. The LCD displays a welcoming message, and the keypad is sensed for any button presses. At this point, the game operator has the option of setting game input sensitivity for a male or female player. A force of 49g and 40g is required to win in the male and female modes respectively.

Our program exits the welcome state when a keypad button is depressed, and proceed to the wait state where it loops until an acceptable input is received from the accelerometer. In this waiting state, a loop is used to monitor the output from the analog-to-digital conversion complete interrupt. We reject all motions of 20g or less to minimize noise, environmental vibrations, but most importantly mis-hits. As soon as a force greater than 20g is experienced, our program compares the value to the one set forth by the operator (i.e., male or female) and goes into a display state. In the display state, depending on results of the player's attempts, we either flash the LEDs rapidly (to indicate a win) or hold the LEDs at a certain position to represent relative effort (to indicate a loss). The display state does not end until the operator presses a soft-reset button, and when this happens we revert back to the welcome state.

Music:

Music is generated by creating square like waveforms to an output pin of the 8535 microprocessor. To create a note at, for example, 500Hz, we calculate that an output inversion each millisecond is needed. A output inversion at that rate produces a periodic signal of 2 milliseconds each cycle, which when fed to the speaker produces an audible note at 500 Hz. Using timer zero with a prescaler value of 64, we are able to create the range of sound desired by switching output values at each timer overflow interrupt. To obtain different notes we simply reloaded various values into TCNT0 at each overflow interrupt. This causes the interrupt to be called at different time frequencies. We then came up with a scheme to store the frequency and duration desired into memory. Each note is represented as a frequency-duration pair. For example, a value of (192,248) corresponds to a 500 Hz played for a quarter of a second. Using this convention, we programmed each note in our melody as frequency-duration pairs in memory. Whenever music is desired, timer zero is initialized, our machine then reads each value from memory sequentially, plays it for the duration required, and loads the next one. This process is repeated until a ending zero is reached. To acheive periods of silence (mute) in speaker output, it is decided that a value of (1,don't care) be used.

Frequency [Hz] Reload Value Duration Value for .25 Seconds

194 96 98
218 113 136
260 136 131
386 175 195
434 184 220
500 192 248
512 195 255

292 149 100
326 160 110
343 165 130

SILENCE 1 DONT CARE


Light Emitting Diodes:

Sychronized with the beats of music are sixteen light emitting diodes driven by two Radio Shack 74HCT138 demultiplexors. Seeing that the number of output pins on our microprocess are quickly being taken up by switches, LCD, accelerometer, and speaker, we see no way of driving sixteen LEDs without the use of a bus output to some sort of decoder. The Radio Shack decoder was choosen for its simplicity and ability to drive 20mA of current at each output pin. A pair of 74HCT138 takes output from pins b1 to b3 and lights up LEDs as required. Output to LEDs are always latched until the next timer zero overflow interrupt. This allows for easy sychronization with music when needed. If a LED is to maintain its value, then the same value can always be relatched into output register (LED).


Accelerometer:

The Analog Devices ADXL150AQC accelerometer is capable of recording shocks of up to 50g in both directions of one axis. The output of the accelerometer is dicated by the equation 2.5-(.038g) [Volts] where g is the force in g's experienced. We phyiscally mounted the accelerometer in such a way that the larger the g force, the small Vout becomes. In other words, a 49g impact will register as approximately 0.64 [Volts] whereas a 20g impact will cause 1.74 [Volts] at the output.

When input from the ADXL150AQC is needed, the ADC is allowed to cycle in free-run mode without noise cancelling, and the ADC complete interrupt is used to capture values from the ten-bit capture register. We used an Aref setting of 4.096 [Volts] so that, similiar to lab five, the least significant bit in the capture register accounts for about 4 [mV]. This becomes helpful when we scale the output value into a score. Since are mostly interesting in the range of activity between zero to 50g, we can simply translate the voltage received into a score between zero to one hundred by manipulating the ADC results by a factor of two.

Our accelerometer is mounted to the inside of a plastic cap with silicon rubber so that is aligned with the verticle axis. The plastic cap is then spring loaded with energy absorbing foam layers to help disperse excess energy as the cap is depressed with a mallet. The fact that our ADXL150AQC came in a surface mount package made it difficult to soldier but greatly simplified the mounting process by virtue of its small size.


Conclusion:

"Whack-A-Cap" is in effect a culumation of what we have learned this semester. It contains many bits and pieces from various labs, but we have improved on many things to make it a complete product as opposed to a congregation of modular code (that approach would have not worked anyhow). For example, we used the instruction lpm (something learned while studying LCD) to retrieve music material from memory, restructed the LCD code to use more relative branches and calls to save program memory, and used external circuitry like the decoder to ease internal programming requirments. We believe these improvements have made the final program possible in a very efficient, manageable way. The ADXL150AQC also works remarkably well, although with a better budget we would have opted for a version of larger range (perhaps on the scale of 250g) so that we may take harder hits before setting off a game-winner.

Link: http://instruct1.cit.cornell.edu/courses/e.../gigi/INDEX.HTM