Ultrasonic Probe SelectionInitially, the PRP stated that a 10 MHz probe would be provided for use on this project. However, due to logistical issues, the team was tasked with choosing purchasing a probe for use on this project. The team focused on probes offered by Olympus, whose website can be found here: http://www.olympus-ims.com/en/ultrasonic-transducers/. The team worked with Olympus representative Brendan Sullivan (firstname.lastname@example.org, phone number: (781) 419-3905 x3512 ).
It should be noted that after conversations with Dr. Helguera (RIT) and Dr. Phillips (RIT) it is believed that 10 MHz is not a high enough frequency to be able to detect the slight differences we are trying to see. This project focuses on measuring a depth of 3 - 5 mm, which (in theory) will require a much higher frequency. Unfortunately, due to time and budget constraints the team had to settle with the initial recommendation of 10 MHz.
A project similar to this one was performed by NASA and an immersion transducer was used. However, the team was uncertain whether this style of probe would work for our application because it was believed that these probes needed to be immersed in water in order to function. Upon contacting an Olympus representative the team learned that the immersion probes do not need to be immersed as the cone at the tip of the probe is made of a material with properties similar to those of water, simulating immersion. The Olympus representative recommended trying an immersion probe (2mm XMS-310) and a contact probe (0.125" M116) and sent the team both demo probes.
Upon receipt of the transducers the team quickly discovered that the immersion probe would not talk with our software, however the contact probe seemed to be providing different readings based upon the material being measured. Therefore, the team purchased the contact probe.
Future generations of this project should consider a few things during probe selection. First, it is likely that a higher frequency probe will be required to achieve the accuracy desired for this project. Unfortunately, the first generation of this project does not have the time or the budget to research higher frequency probes. Furthermore, various types of probes should be tested. The first generation team attempted to try an immersion transducer and a contact transducer, but simply did not have the time or personnel to perform a complete analysis of the transducers that would allow for a confident decision. Lastly, the physical size of the transducer should be considered. The contact transducer for the first generation of this project is simply too small to mount effectively. More importantly, the diameter through which it measures is too large considering the measurement being performed. Future generations should consider two options. The first (and likely cheapest and simplest) option would be to construct a tooth large enough to create the correct proportion between the transducer's measurement window and the tooth geometry of interest. The alternative option would be to purchase a transducer with a smaller measurement window so it does not gather more data than is necessary.
Tooth Phantom Design
The initial thoughts regarding the tooth phantom design can be found on the Systems Design page. This plan included finding materials resembling the dentin, enamel, gums and bone to construct the phantom. However, upon discussing the project with some experts (Dr. Caton of the University of Rochester and Dr. Helguera of Rochester Institute of Technology) the tooth design was reconsidered.
Dr. Caton pointed out that when periodontal disease is present the enamel has been worn away. Cementum is normally found covering the dentin (much like enamel does) below the gum line and meets up with the enamel where the gum separates from the tooth. However, this interface is eroded when periodontal disease has set in. Therefore, a patient with periodontal disease will not have enamel near the sulcus, however only cementum will be present on the tooth side of the sulcus.
Dr. Caton recommended neglecting enamel and concentrating on the dentin and cementum. However, given the time frame of the project Dr. Caton suggested using a pig mandible or extracted human teeth to serve as the phantom. He believes that finding materials that behave ultrasonically like human teeth (dentin) will be an unattainable task in the time allotted. However, he believes that finding materials that are similar to gums and bone will not be as insurmountable.
Dr. Helguera explained that due to the limitations of the equipment to be used (10 MHz transducer and oscilloscope) the difference between enamel and dentin would not be significant enough to recognize the change in material. Generally, a difference in material will result in a spike on the oscilloscope output, however the enamel is simply too thin of a layer to have the same effect. The same concept applies to the cementum. Therefore, Dr. Helguera recommended constructing the phantom to resemble just one material (presumably dentin) as the time frame of the project does not allow for extensive time/effort manipulating and interpreting data to find such a minute benchmark on the output.
Therefore, the tooth phantom will consist of materials representing dentin, gum tissue and bone.
That being said, the chart below was generated to represent the new material selection criteria and incorporate further research that has been performed. All of the materials researched for use are non biological, satisfying customer needs 8, 9 and 10 (specification 3.1). The baseline values for dentin, mandibular bone and gums present a difficulty for the tooth phantom design. One issue is that little research has been performed regarding ultrasonic behavior within the mouth. The major trouble, however, is that as people age the properties of their bones and tissues change. This is greatly due to the possibility that the density of the bones and tissues often decreases with age and varies with height and weight. Women often experience a larger decrease in these properties, partially due to menopause. That being said, the baseline values have an inherent uncertainty,as they represent a subject study of 42 adults (36 women, 6 men). According to the limited research that was available, the team researched countless common materials to see what ones exhibited properties similar to those used as the baseline. Concrete is the most viable option for representing dentin, brick for bone and paraffin or polyurethane for the gums.
One main concern with the materials chosen is the ability to manipulate them. The concrete, paraffin and polyurethane should be simple to form. Brick, however, could pose a challenge. Through many conversations, the team came to the agreement that when ultrasound attempts to pass through bone it is essentially "blocked", or the signal fails to pass through the material and simply bounces back. That being said, to mimic bone the acoustic properties do not necessarily need to be similar, rather a material must be found that will not allow ultrasonic waves to propagate through it. This mitigates any error associated with the values previously used as a baseline, as they are no longer relevant. Also, this opens up more options for the bone phantom material which will allow for easier selection and manipulation.
Therefore, it was agreed that a square aluminum tube could be used to represent the bone. The gums will still be modeled using paraffin wax. The wax will be applied to the aluminum tube through a dipping process that has yet to be refined. The dentin will be modeled using concrete and will be housed within the square bone.
The dentin is a simple cube, leaving the ability to vary the sulcus depth to the gums. Rather than altering the sulcus depth within one phantom, the team agreed to create multiple phantom teeth that each exhibit a different sulcus depth. Therefore, the angled portion of the gums (currently 0.04" on the drawing below) will be altered by changing dimensions L1 and L2 to provide varying sulcus depths, thus satisfying customer needs 1 and 4 (specification 3.3).
Interface With Turn Table
Initially, the phantom was to be placed directly into the turn table. This would be done by machining a square out of the turn table, creating a press fit with the phantom. This would ensure that each phantom was in the same position and orientation each time it was installed. Consistent positioning and orientation are necessary for satisfying customer needs 1, 3, 4 and 11 (specifications 3.2 and 3.3). However, this would require a hefty amount of tightly toleranced machining and would not necessarily allow the phantom to be removed and replaced in a reasonable amount of time, as is required by specification 3.2. Therefore, a new solution was developed.
A plate has been added to the bottom of the phantom. The addition of this plate allows for the phantom/turn table interface to be separate from the phantom itself. This makes for an easy assembly/dis-assembly process that does not involve the phantom materials, thus leaving them undisturbed. Subsequently, the assembly/dis-assembly process becomes simplified. Each plate will remain mated to the phantom it is initially paired with. The addition of the plate also reduces the amount of machining to be performed on the turn table. A simple bolt pattern (eight holes) will be machined on the turn table to allow for two screws to fasten each plate to the turn table.
A few options were available for attaching the phantom to the plate. The first option considered was a simple L-bracket. This would keep the phantom square while ensuring that the phantom was always in the same location (relative to the turn table). Ensuring consistent positioning is important so the probe does not need to compensate when phantoms are switched out. However, the L-bracket method would likely interfere with the turn table itself which is not ideal. The phantom and plate are intended to be a single unit that attach to the turn table with minimal interference.
Another option for attaching the phantom to the plate was through the use of PEMSERT Self-Clinching Flush Fasteners http://hitechfasteners.thomasnet.com/category/pemsert-self-clinching-flush-fasteners. These fasteners would go into the plate and sit flush with the plate on one or both sides. A screw would then be placed in the fastener from the bottom side of the plate, protruding through the aluminum tubing. However, as these fasteners were researched further it was determined that they are all too large to accommodate the wall thickness of the aluminum tubing.
The third option available for fastening the aluminum tube to the plate was Scotch Extreme Mounting Tape, discovered at Lowes. The tape would not interfere with the phantom design or the turn table and would provide sufficient fastening force (20 lbs). The tape would be placed on the plate and the aluminum tube pressed onto the tape. The tape would then be cut to size to avoid protruding beyond the aluminum tube and interfering with the holes for fastening the plate to the turn table.
The team agreed upon the third option, Scotch Extreme Mounting Tape, for fastening the aluminum tube to the plate.
The simplicity of the turn table / phantom interface is necessary to satisfy customer needs 3, 11 and 12 (specification 3.2). The phantom will be easily removable/replaceable due to the nature of the plate. This also mitigates the machining required on the turn table, as the turn table is being purchased from an outside vendor. A simple bolt pattern will need to be incorporated rather than significant, tightly toleranced material removal, as was necessary with the initial design.
Above: drawing for the phantom tooth.
Above: drawing for the dentin.
Above: drawing for the gums. The angled portion of the gum layer will be altered by varying the values of L1 and L2. This will change the angle and depth of the sulcus.
Above: drawing for the bone. The bone layer height changes depending on the amount of periodontal disease that is present. Normal bone crests are parrallel to the cemento-enamel layer. This is where most pockets start. For people with periodontitis the bone decreasees to about 1-2 mm below the pocket start. However with someone who has advanced periodontitis the bone can degrade to be further than 2mm from the start of the pocket.
Above: drawing for the plate.
Code Flow Chart
Above: flow chart of the overall code concept. Attached at the following link is a more broken down set of flow charts for the test code. Code Flowchart Link
Current wiring matrix of servo/stepper motors for electrical side.
Current wiring matrix of slide pot voltage divider implementation for stepper motor feedback system.
Overall wiring matrix of the system involving the Arduino Uno board, incorporating both the Motor design and the Voltage Divider design.
Block Diagram of Entire System
Five motors were required for the proposed system, one for each of the 5 directions of motion. Based on discussions with the team, the motor choices were decided upon once the fixture design was finalized. In the fixture, it was decided to utilize two servo motors and three stepper motors. The two servo motors control movement for pitch and yaw, while the stepper motors control movement for the X-Y-Z axes.
Originally, it was hoped that only servos would be required for the system to operate. This was due mostly because of the ease of control in operating a servo motor as opposed to a stepper motor. However, stepper motors are generally cheaper than servo motors and have a higher holding torque than servos. Thus, it was decided to utilize both types of motors in the design.
One of the more pressing concerns was whether the motors chosen would have enough torque to move the final assembled fixture. As such, it was decided to overestimate what the required torque may be and choose motors from there based on the budget allotted to the team.
The servo motor chosen to control the pitch and yaw motions was the HS-225MG Mighty Mini, shown above. It has a max torque rating of 67 oz-in and is able to operate within 180°.
The stepper motor chosen to control the motion across the three power screws was the Soyo SY57ST76-0686A Unipolar Stepper Motor. As a 6-wire stepper motor, it is able to be utilized for both high speed and high torque operations. The high speed operation is used when the BLK-GRN-RED-BLU wires are used, while the high torque operation is used with the BLK-YEL-RED-WHT configuration. With a holding torque of 125 oz-in, it is more than sufficient enough to operate the fixture based on the expected torque ratings of the system.
Stepper Motor Feedback System
The initial component choices for both a Tripwire and Potentiometer implementation of a feedback system can be found on the Systems Design page. The main concerns regarding the feedback system were cost, accuracy, and integration onto the fixture itself. Also, the pros and cons of choosing either a Tripwire implementation or a Potentiometer integration had to be determined.
Various types of sensors and pots were explored in order to decide what would be the best choice to use as a feedback system.
The Maxbotix LV-MaxSonar-EZ0 is an ultrasonic range finder. It sends out an ultrasonic ping which bounces off an object and returns to the device. The input it obtains changes based on the distance the ping travels. The benefits of utilizing this are that it could set off an alarm flag in the Arduino program based on an abnormal distance reading, related to the position of the probe along a particular stepper motor axis. However, it could potentially be harmful in the possibility of interference with the main ultrasound probe. Also, the size of the sensor and the requirements that it need an additional boundary to determine the distance based on the ultrasonic ping can make it difficult to implement in the main fixture. The cost of this component is within $27-30.
The Sharp GP2Y0A21YK is an IR proximity sensor. It shines a beam of IR light from an LED and measures the intensity of the light that bounces back to the phototransistor. The sensor outputs 0V when there is no object in front of it, and it ramps up as an object comes near the sensor. However, it doesn't sense distance reliably and the values begin to drop again when an object is in contact with the sensor. The cost of this component depends on both the version and range chosen; for this one, it would be $14.
The QRD1114 IR Emitted / Phototransistor is an IR transmitter and photoresistor chip. It can sense objects within a 0cm and 3cm distance, with a very fast detection rate. It only requires two resistors to run it, which also affects the sensitivity of values read by the Arduino board. However, it is susceptible to false positives.
The QRE1113 Line Sensor Breakout is a line sensor IR emitted / phototransistor chip. It is designed specifically for line sensing applications, and is widely used in industry with line following robots. This can potentially be adapted for use on the test fixture as a tripwire sensor, however the test fixture itself would need to be adapted for line sensing properties.
The Bourns 3048L-4-502 is a linear motion potentiometer. This type of pot would be ideal for the feedback system necessary for the stepper motors. However, the major drawback is the cost of the pot in relation to the length of travel it can record. For this particular model, it costs ~$30 and can only travel up to 0.5" length.
The ALPS RS60112A600N is a slide potentiometer. At a cost of $4.00, it is a very cheap option to consider. This is a 10K Ohm pot that is able to travel 60 mm, or 2.3622".
Tripwire vs. Potentiometer Selection
There were several factors to consider when deciding between utilizing tripwire sensors and utilizing pots for the stepper motor feedback system. The tripwire sensors would work for shutoff purposes, in order to ensure that there are no collisions when moving on a stepper motor axis. The potentiometer option would be utilized as an analog voltage reader that varies depending on the resistance of the pot, via voltage divider implementation.
With these component choices, it was decided that a potentiometer would be utilized for the stepper motor feedback system. This is due in part to the ease of integrating a pot with the fixture, along with the very low cost of doing so. As discussed in the preceding section, the tripwire sensors would require additional and potentially complicated means of adapting the fixture in order to integrate the sensors. Also, the reliability of utilizing the tripwire sensors as a software shutoff is not the best means of collision prevention due to the potential lag there would be with the Arduino in reading the alarm flags. The consistency of these sensors to correctly output values to the Arduino at all times is also suspect.
Between the Bourns linear motion pot and the ALPS slide pot, it was decided to use the slide pot for the feedback system. As discussed in the preceding section, the Bourns linear motion pot has very little travel length compared to the ALPS slide pot. The ALPS slide pot, at a travel length of approximately 2.36", would be a much better choice to utilize over the Bourns linear motion pot, which is at most 0.5". While the slide pot would require the test system to have a means of pushing the slide pot back and forth during each operation of the three stepper motor axes, the longer length of it made it a much better choice than the linear motion pot. Also, the cost of the slide pot is much cheaper than the linear motion pot.
Slide Pot Integration
The slide pot will be integrated into the assembly through a voltage divider implementation. Below are spec sheets of the slide pot.
Below is a cross-sectional image of the slide pot.
Below is a wiring schematic for one of the voltage divider circuits with slide pot implementation.
Mechanical Positioning System DesignThe initial concept design for the probe positioning system can be found on the System Design Page. As of the system design review, the overall structure of the positioning system was well known and seemed to be fairly robust. The 2 main axes controlled by the power screws were almost complete. The areas that needed to be defined the most were the pitch axis movement where the probe itself would be fastened, and the turntable interface that controls the yaw movement. Also, some material choices needed to be considered to try to limit budget constraints. The following pictures and descriptions illustrate how these challenges were overcome and how the final detailed design comes together.
Below is the 3-D model and assembly drawing of the completed system. This is a good overview of the design and specific subsystems will be looked at in further detail.
As previously stated, the movement on the 3 cartesian axes will be controlled with power screws and stepper motors. Each axis will have an almost identical design, since the screw is being ordered in a longer length and cut to size. The screws will be ordered from Hayden Kerk Motion Control and will be made from 303 Stainless Steel. The screws will be accompanied by a plastic anti-backlash nut from the manufacturer. The only difference between the three axes will be the length of the screw and screw threading. To deal with the axial loads transmitted by turning the screw, something to accept the thrust needed to be incorporated into the design. To accomplish this, a steel ball bearing will be placed on either side of the structure frame on the end of the screw opposite the motor. Also, set-screw style collars on each side of the frame will be used to axially locate the screw to the bearings.
Below is a spec sheet of the screw and nut and a detailed drawing of the nut.
Below is a cross-sectional drawing showing how the screw will be supported with the bearings, and connected to the motor.
The bearings chosen for this application are simple steel roller bearings. Since the axial loads that the system is expected to generate are so low, (see Torque Calculations on system design page) we are able to get by with simple bearings and collars instead of a more expensive thrust bearing. The bearing and collar spec sheets can be seen below:
The screws will also use simple guide rails to help take on loads and make for smooth travel to prevent binding up. The guide rails will be the same for each axis with only the length being varied, similar to the screws. The guide rails will be fixed to the outer frame of the system using small set screw clamps on each end of each rail.
The frame of the test system will be a 6-piece design, machined from .75" x 3" bars of extruded aluminium. The frame will have two longitudinal sections, left and right, spanning the longitudinal axis of motion, and two lateral sections mounted perpendicular to and joining the longitudinal sections at their inside faces. Two vertical sections will be mounted near the rear of and on the outside faces of the longitudinal sections, and will span the lateral axis of motion. The longitudinal power screw will align with the longitudinal sections and be supported by bearings mounted in the lateral frame sections, the lateral power screw will align with the lateral frame sections and be supported by bearings mounted in the vertical frame sections, and the vertical power screw will be aligned with the vertical frame sections and be supported by bearings mounted in the special U-Bracket, discussed later on this page. Drawings of the 6 frame components are shown below:
Two guide rail clamps needed to be designed; one for general use, and one for mounting on the U-Bracket bottom, which includes an extra milling operation to avoid interference with the frame. A spec sheet for the rails and detailed drawings of the clamps are shown below:
To mount the probe and turntable subsystems on the linear axes, there needs to be a platform that connects to the nut and slides on the guide rails for each axis. To simplify manufacturing, a common platform was designed with the only differences being the bolt pattern to accept necessary attachments, as well as the spacing of the guide rails through the platform. Drawings of the three platforms for each of the three axes are shown below:
In order for the sliding platforms to move along their respective axes as the lead screws are turned by the stepper motors, L brackets needed to be designed to bolt to the platforms and lead screw nuts, and in doing so mate the platforms to the lead screw nuts at a 90 degree angle. Thus, linear force transmitted to the lead screw nut from the lead screw as the screw turns will be further transmitted to the platforms, providing the force that actually pushes and pulls the platforms along their axes. Two L brackets were designed: one for the vertical axis, with a 1" platform mounting hole separation as necessary to avoid interference, and one for both the lateral and longitudinal axes, with a 1.25" platform mounting hole separation. Detailed drawings of the two L bracket designs are shown below:
Since the vertical axis of movement itself must move laterally, a special mounting system needed to be designed. A 3 piece U-shaped mounting bracket was chosen for its ease of manufacturing and the fact that it can be made of already purchased stock Aluminum. An isometric CAD view of this assembly and detailed drawings of its components are shown below:
The next subsystem of interest is the turntable. From our system design, the turntable will be driven by a simple servo motor and will be capable of controlling the yaw movement of the tooth phantom. One important limitation of this subsystem is that the turntable needs to be supported externally, since the servo can not accept any axial loads. To do this leftover Aluminum stock from the frame components will be used to build servo mounts and a mounting structure for the turntable. The turntable structure is comprised of two pieces, one being the table itself and the other being the shaft component that couples to the servo. The table is supported on a flanged sleeve bearing made of oil impregnated bronze to handle the axial loads and allow for low friction rotational movements. Below is a cross section of the entire turntable system, as well as a detailed drawings of the turntable and turntable shaft.
On top of just being able to rotate, the turntable needs to be able to provide an interface to mount the tooth phantoms. This interface is described in detail in section 1.3 above (Tooth Phantom Design/Interface with Turn Table). It is crucial that the tooth is placed in the same spot every time in reference to the probe, and that it is rigidly mounted. To accomplish this, the tooth phantom will have a flanged design to allow it to be bolted directly to the turntable.
Drawings of the servo mounts, horizontal table support, and vertical table supports are shown below, which comprise the support system for the servo motor and turntable assembly:
The next subsystem of interest is the pitch axis movement system. This system needs to fix the transducer probe to the vertical axis of movement and allow the probe to rotate through approximately 180 degrees using a servo motor. This is accomplished using three mounting plates; One being used to mount the servo, and the other to mount the probe holding fixture. A cross section of the entire system is shown below.
To hold the probe in place a 2-piece split fixture was designed. This component first has two holes drilled in it to match the profile of the t-shaped transducer, then split in half so that the probe can be installed. Once the probe is installed, the two halves are bolted back together. To protect the fragile probe, small lengths of rubber tubing are attached to each side of the probe before the fixture halves are bolted together. Detailed drawings of these fixture halves are shown below:
Once the probe is fixed, there needs to be probe holder housing to link it to the rotational axis and also something to allow the probe to ride on a spring within the probe holder housing. The purpose of the spring mechanism is to allow the probe to make contact with the tooth without damaging it, and to maintain contact while translating up and down. To do this a channel is milled into a rectangular extrusion of steel, which will comprise the probe holder housing. The probe clamping fixture will ride in the channel and there will be a spring between the two. The probe holder housing will then be pin connected to a shaft which is coupled directly to a servo. A detailed drawing of the probe holder housing is shown below:
The shaft that the holding fixture is riding on is made from 3/8" carbon steel bar stock. One end of the shaft is supported by a flanged bronze sleeve bearing that is inserted into the left mounting block. The middle of the shaft is supported by a simple bronze sleeve bearing in the middle mounting block, and the other end is coupled to the servo motor, which is seated in the right mounting block. Detailed drawings of the shaft, the three mounting blocks, and specs for all the necessary sleeve bearings/bushings are shown below:
The last important subsystem of the mechanical design of the fixture is the mounting of the positional feedback system, comprised of 3 linear potentiometers described above in the electrical section. The potentiometer (pot) for each axis is mounted to a frame component to hold the circuit board in place, while the arm of the pot, with a linear degree of freedom, is rigidly mounted to the sliding platform so that movement of the platform will result in an equal translation of the linear pot arm along its degree of freedom. In this way, motion of each platform is recorded in the motion of the pot arm, which enables the determination of a linear position of the platform along each axis. Mounting for the three axes is accomplished using a common mounting plate to "sandwich" the pot arms and provide a rigid connection through friction as the screws holding the "sandwich" in place are tightened. More detailed descriptions of the mounting scheme for each axis are shown below:
Longitudinal Axis: Linear Potentiometer Mounting
Vertical Axis: Linear Potentiometer Mounting
Lateral Axis: Linear Potentiometer Mounting
Detailed drawings for the linear pot mounting parts, including the common mounting plate, vertical and lateral L brackets, lateral mounting bar, and lateral extension plate are shown below:
Bill of Materials (BOM)The file maintained as the BOM can be found here:Budget/BOM
The options available for each of the three design components (phantom tooth, mechanical fixture and electrical components) were evaluated to determine whether or not they were feasible. The analysis and reasoning can be found here: Feasibility Analysis.
Test PlansThe current Test Plan is located here: Test Plan.
The projected schedule for when subsystem tests will be performed can be found on the Planning & Execution page under "Projected MSD II Schedule".
Detailed Design Review
Preliminary Detailed Design Review Documentation
The presentation for the preliminary detailed design review held on Friday, February 1, 2013 can be found here: Preliminary Detailed Design Review Presentation.
The meeting minutes from the preliminary detailed design review can be found here: Preliminary Detailed Design Review Meeting Minutes.
Detailed Design Review Documentation
The presentation for the detailed design review held on Friday, February 8, 2013 can be found here: Detailed Design Review Presentation.
The meeting minutes from the detailed design review can be found here: Detailed Design Review Meeting Minutes.