Design Progress
Nov. 17 - Oct. 12    |    Oct. 11 - Sept. 19

November 18 - The team finished the Detailed Design Specifications paper, and submitted it to Dr. Mockensturm.  Then, the team focused on a way to provide the ball with lateral stability for the initial attempts at levitation.  A metal cage was constructed, which restricted the ball's lateral movement, but allowed the ball to move vertically.  This provided the team the ability to precisely align the electromagnet, ball, and sensor.  The alignment is very important, since a small change in lateral movement would cause an inaccurate sensor reading.  Although levitation was not successful, members of the team were able to hold the ball near the equilibrium point and feel the ball shaking with instability.  This is a promising sign.  The team also spoke briefly with an EE professor about possible problems with the design.

November 17 - The team received the new op amp.  However, in response to the continuous problems with the op amps, the team developed a new circuit to eliminate the need for a large op amp.  The new design utilizes four 741 op amps in conjunction with four MOSFATs arranged in parallel.  When the signal from the controller comes through, the op amps amplify the voltage, and then the MOSFATs amplify the current.  When tested, this circuit passed the first test - the electromagnet held the ball when powered.  Now, the team must perfect the controller, to achieve levitation. 

Also, an equilibrium point for levitation was chosen.  This point was chosen arbitrarily, but was designed to provide approximately a 0.5 inch gap distance, as desired by the team.  At the equilibrium point chosen, the ball would be suspended 9.5 cm above the sensor.  This equates to a gap distance of 1.1 cm.  Using the sensor's characteristic equation of V= -0.5396 g + 4.607, where V is measured in volts and g is measured in cm, the team was able to obtain a relationship between the gap distance, x, and the sensor voltage, V.  As you can tell from the diagram to the left, D= g + d + x.  Combining these two equations, x = - 1.853 V + 5.2178 cm.

Therefore, when the system has a gap distance of 1.1 cm, the sensor reads a voltage of 2.22 V.  In a stable system, this would be the sensor voltage when the ball is at a constant distance of 1.1 cm below the magnet.  However, the instabilities will cause the voltage to deviate from this value.

November 16 - The team met with Dr. Mockensturm in the morning, hopefully that the new op amp would arrive within the day.  After contacting the company, it was clear the the new op amp would not arrive until 11/16.  The team prepared wires for the new op amp, as they would now be crimped together, not soldered.  Although it is not thought to be the reason for the failure, the team did not want to take any chances soldering the wires onto the new op amp.  Also, a small strip of metal was used to rigidly attach the motor to the roller, to optimize contact of the teeth.

November 11 - The team met to work on the control system with the new op amp.  A new circuit was made using the same resistance values as the initial circuit.  The team used a piece of aluminum as a heat sink for the op amp.  Holes were drilled in the metal to accommodate the op amp pins, and were then filled with thermally conductive putty.  The homemade metal heat sink was then screwed into a wooden board, which will just act as a base.  The circuit board was also attached to the wooden board.  Unfortunately, once assembled, the new op amp did not function.  All team members were confident of the design and calculations, so the company was contacted.  The company agreed to send a new op amp of the same type at no charge.  The old op amp will be sent back to the company to be tested for manufacturing defects.  If the component is found to be defective, the team will not be charged for the replacement op amp; however, if the component failed due to the team's actions, it must be paid for. 

The team also worked on the spinning mechanism.  The small rubber roller was disassembled, and the gear was attached along the same metal axis as the roller.  The fit was very tight, allowing the gear and roller to spin as one entity.  The gear then nicely messed with the gear to spin the whole assembly.  Although the ball is not yet levitating due to the issues with the op amp, the team wished to simulate levitation to see if the roller would spin the ball in the way anticipated.  A wooden stand was created to hold the ball closer to the magnet.  The magnet was then used to make the ball almost lift off of the platform.  Then, the roller was used to contact the ball.  This spinning method worked wonderfully!  The ball spun to everyone's satisfaction.  To estimate the RPM at which the ball had been spinning, the RPM of the roller was measured to be almost 700 RPM - far above the objective of spinning the ball at 200 RPM. 

A lot of progress was made today.  The team still faces the challenge of getting the controls to work once the new op amp comes in.  However, every other aspect of the project is running very smoothly, including the newly constructed spinning system.  The team is very optimistic that when the new op amp arrives, the ball will be levitated and spinning!

                             Preliminary Circuit Design                                              Motor and gear/roller assembly           

Construction of gear/roller assembly and op amp heat sink

Testing the spinning method

November 10 - The new op amp arrived.  A small rubber roller, usually used for crafts and such, was purchased for spinning the ball.  The team also obtained a small plastic gear and a small motor.  The idea is to attach the small gear to the roller as one piece.  The motor will mesh with the gear to make the gear/roller assembly spin.  The roller will then contact the ball, making the ball spin.

November 8 - Due to a miscalculation, the op amp was blown.  A new op amp was ordered.

November 2 - After meeting with Dr. Mockensturm, the team began work on the controls circuitry.  First, the inductance of the magnet was measured to be 16.9 ohms.  Therefore, the gain of the op amp of the control system needed to be designed to be 16.9 ohms.  This was done by selecting two resistors, R1 and R2, such that R2/R1 = 16.9.  After selecting the appropriate resistors, the basic circuit was assembled.  This circuit was designed to take the output from the power supply and feed it to the magnet.  The team decided that the method of spinning using just an air stream would not be reliable enough and that a roller would be used to contact the ball.

November 1 - The op amp and controls components arrived!

To the top

October 26 - The team presented the current design to the class, and answered questions posed by classmates.  Important issues discussed centered around the reasons for choosing the particular prototype design, as well as areas of refinement and future trouble.  The presentation was well-received and the team was given the possibility of borrowing a power source through another class member.  The amplifier and other parts ordered for the control system are expected no later than the beginning of next week, at which time the team can begin levitation.

October 24 - The electromagnet was successfully characterized.  The team also discussed the feasibility of adding a small protective pad to the underside of the electromagnet.  There is a fear that if the magnetic field becomes too strong and pulls the ball up into the magnet, that the force may crack or damage the magnet. 

October 19 - The team met with Dr. Mockensturm to discuss progress to date.  Thus far, the team has spent $460 of the $600 budget.  The team selected a design for addition of the sensor.  After reviewing the results from the characterization of the sensor, the best design would have the sensor about 1.3 inches below the ball.  The range surrounding 1.3 inches has the greatest voltage drop per change in distance, allowing optimum response and control of the system.  Several designs were considered, but the team's final decision was to add a cantilevered platform to hold the sensor approximately 1.3 inches below the bottom of the levitating ball.  Since the team is aiming for a 0.5 inch gap, this would put the preferred location of the sensor 3.8 inches below the bottom of the electromagnet.  The plan is to sit the sensor down in a U-shaped piece of aluminum, to protect the sensor from the ball, if it were to fall directly downwards.  Additionally, load analysis has been done, and the team has determined that the platform will be able to withstand the force of the ball falling on it. 

After the decision was made, the sensor apparatus was added to the prototype.  The figures below show the newest progress of the prototype.  Everyone decided that it would be a bad idea to rigidly connect the sensor to the platform, since having it centered is such a big concern.  So, the team decided to have the sensor be mounted over a slot-type hole, allowing for movement of the sensor in the xy-plane to perfectly center it under the ball.  Once centered, the placement will be locked using nuts on the underside of the platform. 

               The prototype thus far                                  The cantilevered arm used to mount the sensor

View from above of protective channel used to house the sensor

October 17 -  The electromagnet was characterized, much in the same way as the sensor.  As with the sensor, it is essential for the team to test and understand the relationship between the inductance of the magnet and the distance between the magnet and the ball.  An LCR meter was used to measure the inductance of the magnet as the ball was moved further away from it. 

October 13 - The team completed the characterization of the sensor.  In order to do this, a special apparatus was made to allow the ball to slide only along the line normal to the emitter of the sensor.  This was done by placing the ball in a set of rails for stabilization, preventing any lateral or vertical movements that could alter the results.  The figure below shows the apparatus used for sensor characterization.  Next, the team determined the exact height to place the sensor, using geometry to determine the height of the center of the ball from the top of the rails.  Then, with the sensor aimed directly at the center of the ball, the ball was moved further and further from the sensor in 0.25 inch increments.  This process was carried out for distances from 0 to 7.5 inches from the sensor.  A Digital Multi Meter was used to record he voltage of the sensor as a function of the distance to the ball.  The results of this test were then graphed.  (Sensor Characterization Graph)  This characterization of the sensor is the  determining factor of the placement of the sensor on the apparatus.  It is essential to place the ball in a position where small changes in position will be picked up by the sensor. 

Sensor Characterization Apparatus

October 12 - The team met to test the prototype with a power source.  While hooked up to the power source, the electromagnet successfully held the steel ball.  The prototype has passed the first test – the electromagnet is strong enough to counteract the weight of the ball (which is substantial).  Members even tested the system by placing a finger between the ball and the electromagnet and the ball was able to remain levitated!  These accomplishments were encouraging.  The team is looking into purchasing or borrowing a voltage-controlled voltage source. 

The sensor has yet to be attached to the prototype.  Placing the sensor will be a precise process.  First the team needs to characterize the sensor, which is planned for October 13, 2004. 

Looking to the future, the team anticipated that a fairly serious obstacle will be the lag of the electromagnet with respect to a change in power.  In other words, once the sensor detects the ball is dropping, the power source will increase current to re-center the ball.  However, this change in current will not cause an instantaneous change in the electromagnet.   The controller will have to overcome this obstacle; otherwise, control of the ball’s position will not be possible.

To the top

Nov. 17 - Oct. 12    |    Oct. 11 - Sept. 19