Rochester Institute of Technology
Project Number: P15665GLEASON GEAR JAW METRICSEvan MolonyMechanical EngineeringDoug PerryMechanical EngineeringJosh SmithMechanical EngineeringKatie BaldwinIndustrial EngineeringABSTRACTThe Gleason Works is a Rochester based company whose “principal activity is the design, manufacture and sale of machinery and equipment for the production of bevel and cylindrical gears” [1]. The manufacture of these gears often involves a gear blank (gear prior to machining) being lowered onto an arbor by a pair of steel jaws. If these jaws are machined improperly or damaged in some way they have the potential to tilt and damage the entire gear manufacturing machine costing Gleason and its customers thousands of dollars in repairs. Gleason approached RIT to design and produce a gear jaw inspection station capable of identifying defective jaws for three of Gleason’s machine models before they are allowed into the field in order to prevent future damage and costs. The requirements for this station state that it must be able to accurately characterize any tilt or deflection in jaw surfaces with all measurements being within 0.001” of their true values. Our group’s solution to this involved the creation of an adjustable jaw mounting structure with full control over the pitch and roll of a measurement plane. This structure allows an operator to ensure a parallel measurement plane to a granite surface plate. This system is used in conjunction with an adjustable height drop test indicator to obtain accurate measurements of all relevant jaw surfaces.INTRODUCTIONThis project was conceived due to several malfunctions Gleason machines have experienced in the field. When a defective jaw or pair of jaws is loaded into a Gleason machine it can have the effect of tilting the gear blank. The gear blank is typically lowered onto a circular metal arbor. When a tilted blank is lowered onto this arbor it can cause serious damage to the arbor or become lodged in the machine. When this occurs it can require a large rework operation to repair. Gleason roughly estimates that catching a single non-conforming jaw could save the company up to $10,000 in repair costs and lead to higher customer satisfaction. This is the motivation behind this project. Gleason officially requested a measurement station capable of measuring all jaws used in three of their machines, the 400H, 130/210, and the 200/300 with the potential to accommodate future machine models as well. Gleason also requested a system accuracy of 0.001” meaning that all measurement error in the system must add up to no more than 0.001”. Gleason estimates that if a discrepancy of more than .030” is detected from one end of a horizontal gear blank contacting surface of a jaw to the other, the jaw will be unacceptable for use in the field.Gleason has never designed a system for this purpose before, and a review of prior RIT projects showed no similar work. Based on this, all background research into this project was largely based on prior group member experience, and through the use of subject matter experts both at RIT and in the field. The system was broken down into several subsystems, measurement and data collection, jaw securing and structure, feedback, and calibration. While some aspects of this such as the securing of the jaws had straight forward solutions, others such as measurement and data collection did not. Several group members have previously worked with various metrology equipment in the field and this was the starting point for this project. The group brainstormed all of the various gauges and technologies we have worked with or heard of in the past and conducted extensive Internet searches to discover new technologies. It was immediately clear that some form of accurate calibrated gauge would be required to characterize the jaws. The initial research turned up several promising technologies; Drop test and test indicators: Small calibrated linear or radial displacement gauges (both digital and mechanical) commercially available in a wide variety of lengths, accuracies, and prices. These gages are ideal due to their relatively low prices, versatility, and accuracy.Height Gages: Mobile tabletop gages which measure the height of surfaces through the use of highly accurate calibrated linear encoders. Though ideal for their accuracy and ability to measure at varying heights, digital height gauges are more costly and their ability to measure height discrepancies over a single surface is questionable.Linear and rotary encoders: Devices which turn linear and rotary motion into voltages which can be used to accurately record motion. Though incredibly accurate, such devices are typically expensive and would require the construction of complex systems to use which may already be commercially available (see Height Gage).1D and 2D Lasers: Laser measurement devices which either give linear distances from gauge to surface, or provide a map of an entire two dimensional surface. These devices were found to be the most accurate option available but were found to be well outside of the project budget.Coordinate Measurement Machines: Commercially available machine which consists of a flat granite table and accurate probes which can manually or automatically characterize the geometry of hardware. Though accurate and made specifically for applications similar to this project, these machines are expensive and would require modifications to accommodate this project. With these ideas in mind, our group approached several members of the RIT staff as well as representatives from various metrology companies to refine these basic concepts into a functional measurement solution. Research into the jaw securing subsystem was split into two further subsystems, a table to hold the weight of the device, and a jaw positioning system to position and secure the jaws to the various machines. The table research was straightforward and involved Internet searches to find a table which was large and sturdy enough to support all other hardware. The jaw positioning system research involved extensive communication between our group and Gleason to determine the jaw mounting geometry of the various machines involved in our project, as well as communication with professional toolmakers and subject matter experts to design a system which can be adjusted and calibrated to ensure accurate measurements.Benchmarking and research into the feedback and calibration systems was unnecessary as the research mentioned above provided solutions for these without requiring additional work. Based on all of this research and learning our group was able to provide a functional design despite our lack of extensive experience in the metrology field.PROCESSThe most critical step in the design process of this project was gaining a proper understanding of the customer requirements. The initial guidelines were vague, but after numerous meetings with the customer we were able to come up with a suitable set of requirements which enabled our design process. Some of the more critical requirements which impacted our final design are listed below.Provides accurate measurementsSecurely positions the assemblyConforms to predefined budgetAccommodates three specific machines (GP400H, GP200/300, GP130/210)Measures all critical surfacesThese customer requirements enabled the creation of objective measurable engineering requirements. Several of the design critical requirements are listed below.Gage Error From Ideal Over Full Measurement Envelope ≤ 0.0004” (1.016e-5 meters)Functional Prototype Cost ≤ $10,000Jaw deflection when subjected to 20 Lb load ≤ 0.0004” (1.016e-5 meters)Measurement System & Output Error ≤ 0.001” (2.54e-5 meters)These requirements provided a concrete framework for our design efforts. By combining our aforementioned research and benchmarking with these requirements we were able to begin developing our ideas into physical concepts.The jaw securing and structural system was the first to be designed. Several key assumptions were made based on subject matter expert input on machining processes, as well as the general engineering comprehension of our group;No part can be machined perfectly. Every part created will deviate from the specified dimensions by some amount no matter how carefully it is machined. Similarly, error in one part can stack with errors in subsequent parts. In order to guarantee accurate and reliable measurements the system must be adjustable and capable of some form of calibration to remove these inherent uncertainties.All materials deform. Even the strongest materials will deflect when subjected to a load or moment from a jaw (sometimes weighing up to 15 lbs or 67 Newtons). Due to the strict accuracy requirements of this system, efforts to maximize the rigidity and structural integrity of the system must be taken, and calculated deflections must be added to the final measurement uncertainties.Knowing this, the group decided that the system had to be adjustable and capable of some form of calibration in order to remove the stacking of tolerances between parts. The critical parameter which we needed to establish in order to eliminate inaccuracies was a plane of measurement which was completely parallel to the plane in which our gages would travel. In order to do so, we took examples from various preexisting metrology equipment. After several iterations of ideas, the simplest method of adjusting the angle of the plane of measurement was established. We decided to use a series of bearings and thumb screws which can be turned to adjust the pitch and roll of the system. This enables complete control of the plane of measurement, and allows for periodic calibration of the system to ensure measurement accuracy. In addition to the adjustment screw and bearing system, a machine spring was placed opposite each screw to ensure system stability and reduce vibrations. This spring was sized to provide a moment to counter 80 lbs of force placed on any point in the system (this number was the result of anecdotal experience of a subject matter expert). The customer requested all parts be made from a specific steel alloy (AISI 8620) which has superior strength to low grade steels such as 1018. Based on the strength of the material and relatively low weight of the system only basic stress analyses were performed and shown to be largely unnecessary. Figure 1 below shows the final form of the adjustable jaw securing and structural system.Figure 1: Adjustable Jaw Securing SystemThe next critical design step taken was to minimize the deflection of the system when subjected to the load of the jaws while also ensuring a proper jaw height, and minimizing the cost of the materials required to make the system. We designed a vertical structural member known as the “central column” to support the jaws and ensure they are at the proper height. To calculate the system deflection, a conservative bending analysis was performed on this central column. The customer gave a maximum jaw weight of 15 lbs and stated they could be up to 18 inches in length and height (67 newtons and .457 meters respectively). Using these parameters, we assumed the full 15 lb load was applied at the full 18 inches from the central column. The resulting moment was applied over the full 15 inch (.381 meters) length of the column. This analysis returned a bending angle which was then converted into a deflection over an 18 inch jaw. This process was iterated with thicker column widths until an acceptable 18 inch deflection was reached. With a 2 inch thick column a maximum deflection of approximately .00027 (6.780e-6 m) inches could be reached (with only one jaw mounted). While this number is nearly one third of our allotted .001 inches, it is the most significant source of error in the system that cannot be calibrated out and thus is acceptable. The full bending analysis may be found on our group website [2]. In order to mount the jaws to the central column, a versatile series of machine interfaces had to be designed. In order for our system to accommodate more than one Gleason machine, interchangeable machine interfaces were required. Our solution to this problem was to design a series of three 6.5” x 11” plates. These plates can be bolted directly to the top of the central column and mimic the geometry of the three specific Gleason machines we were asked to accommodate. These plates are exact replicas of the Gleason geometry, but the tolerances of our plates are tighter, and require surface grinding of all jaw locating features to ensure minimal machining error. The dimensions of these parts are among the most critical dimensions in the entire system. If these dimensions were out of spec, the system would be difficult if not impossible to calibrate due to improper interfaces between the plates and jaw or calibration hardware. Additionally, angular mismatches between left and right jaw mounting surfaces would negatively impact system accuracy. Special care was taken with these parts during the machining process as described below. Figure 2 shows the three completed interface plates. Figure 2: Machine Interface PlatesFinally a table was selected based on its ability to hold the weight of our system, and a base plate was designed to place the central column and adjustment system on to prevent unnecessary damage to the wooden table.The second system to be designed was the measurement and data collection system. The primary assumptions needed to enable the design of this system were as follows.No gauge is completely accurate. All gages and metrology equipment will have some error. Though this error is unavoidable, it can be minimized through the use of properly calibrated equipment.Error in metrology equipment stacks with other metrology as well as other mechanical tolerances and must be tracked to ensure the total system accuracy is within the acceptable limits.These assumptions guided our group towards picking the most accurate measurement equipment possible, starting with our granite surface plate. Granite is a standard device used in the metrology field based on its ability to hold extremely tight flatness and surface profile tolerances. Our group selected a granite plate based on the size of our measurement envelope (approximately an 18 inch square), a customer request for additional storage space, and its overall tolerances. We selected an 18” x 24” grade AA granite surface plate meaning that the overall surface flatness tolerance is within 0.000075” over the full surface of the plate. This is less than 10% of our total system accuracy allowance. After choosing the granite plate, we selected a gage. In order to measure several points on a surface and determine its profile and angularity, we knew we needed a device with the ability to accurately measure vertical distances. Several of the devices mentioned above were considered, however it was found that laser devices and CMM machines were outside of our budget, linear encoders are expensive as well and would require pairing with a second indicator to get useful measurements, and finally, normal test indicators are rarely accurate enough, and provide additional complexity due to their angular deflections. Based on these facts, a simple digital drop test indicator was selected as our gauge of choice. After consulting with a representative from the Starrett company, we decided that a Starrett 2900-2 drop test indicator was the best fit for our project. With an accuracy of 0.00012” (0.003 mm) this indicator takes up around 10% of our allotted accuracy requirement. As this gauge does not come with any fixturing, our group also had to design a vertical support structure to hold the measurement device steady during the measurement process. By coincidence we were able to salvage a modified Starrett 755-24 height gauge from RIT storage, and modify it to hold the indicator. This eliminated the need to design further support. The full measurement system can be seen in figure 4 below.The feedback system design went hand in hand with the selection of a measurement device. Most devices we looked at had electronic outputs which were capable of sending the gage readings directly to an excel document. We purchased an output cable for this very purpose with an additional foot pedal to prevent an operator from shifting the system by pressing a button over the table. The output cable plugs into a computer which Gleason will provide via a USB port.The calibration system was the final system to be designed. The purpose of this system is to remove all unnecessary system error. This system works in conjunction with our screw and bearing adjustment system described above to ensure the plane of measurement is perfectly parallel to the plane created by the interface plates. In order to accomplish this it was clear that flat plates similar to gage blocks would need to be designed to mimic the locating features of Gleason jaws, but with tighter tolerances. The basic idea was to create flat surfaces perpendicular to the interface plates and ensure their squareness to jaw mounting surfaces with careful surface grinding and machining techniques. These “ideal jaws” if machined properly would allow an operator to confidently calibrate the system to ensure that if a completely conforming set of jaws were placed in the system, zero deviation in height would be detected over the entire system measurement envelope. In order to use the calibration plates, an operator must align the locating features of the calibration plates with the interface plates, bolt the two together, and tram the drop test indicator across the top of the calibration plate until zero deflection is measured across the entire surface. It should be noted that the surface grinder used to create the finished surfaces of the parts has a flatness tolerance of approximately 0.0002” and that this number must be added to the overall system accuracy. An example of a calibration plate can be seen below in figure 3.Figure 3: 400H Calibration PlateFigure 4: Final AssemblyNumerous other considerations were taken into account during the machining stage of this project. For a full description of the tolerances and requirements of each part in our system, detailed drawings may be found on our edge site [2]. The key processes to the success of this project however, involve extensive use of surface grinding as it is the most accurate process RIT is capable of producing. All machined parts have multiple ground surfaces to ensure maximum dimensional conformance, and minimize system error. Additionally, each interface plate and the central column were each heat treated to provide a minimum hardness of 60 Rockwell C to a minimum depth of 0.030” (7.620E-6 meters). This prevents damage and wear over time as these parts will be consistently contacting other hardened surfaces. Even with these precautions operators will need to use extreme caution when replacing interface plates due to their importance for system accuracy.RESULTS AND DISCUSSIONA comprehensive test plan was used to evaluate the capabilities and success of this project. Each engineering requirement was addressed with a numerical test the full details and results of which may be found on our edge site [2]. The details of our more critical requirements however will be discussed below.The final cost of the prototype fell well below the $10,000 set aside for this project. The final estimated cost on our bill of materials is $6,766. Due to additional costs associated with shipping, and labor from Gleason to heat treat and grind several parts, it is hard to give an exact figure on the final cost of the prototype. Our group conservatively estimates that the project cost no more than $8,000 to the customer.The prototype meets all sizing requirements including a total footprint of 8.975 sq ft, less than half of the 20.25 sq ft allotted as measured by tape measure. The device can accommodate a measurement envelope of approximately 18”x24”x24” above the 18” cube requested. The table height was measured with a tape measure and is approximately 39”, well within standard ergonomic ranges.The process time was tested by giving 10 RIT students work instructions, and then timing them as they completed a full calibration and jaw measurement cycle. These trials produced an average time of approximately 19:42 which is lower than the 20:00 requirement. Additionally this was the first time any subject had used the machine, we predict that these times will decrease drastically with practice.The device repeatability was measured by repeating the calibration and measurement process for the same jaw several times and noting the largest discrepancies in measurements in each location. The final repeatability was found to be approximately 0.0012”. This number falls short of the 0.001” allowed, however Gleason failed to provide proper steel jaws for this test, and instead 3D printed replicas were used. These replicas were found to change measurement results based on the tightness of their bolts, and even slight movement of the indicator tip would produce unacceptable deflections that would not be found on steel jaws. We have recommended that Gleason repeat this test when they are able to procure a real set of steel jaws.The most important test in this project was the final accuracy test. This test consisted of a tolerance stack of all error in the measurement and calibration system. This accuracy can be broken down into the following key tolerances:Angular mismatch between locating surfaces of the calibration plates and the interface plates. This angle must be turned into an inch value 15 inches from the central column.Calibration plate surface flatnessGranite surface plate flatness. This tolerance must be added twice, once during calibration and again during measurement.Gauge error during measurement.Jaw deflection due to column bending.The angular mismatch values have been tested for the 200/300 interface plate using calibrated metrology equipment. The other values in the above list were obtained from calibration certificates and standard manufacturer assurances. Adding each of these values for the 200/300 calibration plate yields a full system error of approximately 0.00085” at a distance of 18” from the central column. This value is the best estimate our group is able to provide as to the accuracy of our system. For the 200/300 our machine meets our system accuracy requirements and to that extent the project has been a success. Final accuracy measurements for the other two machined plates will be obtained in the following weeks as they require complex grinding which our group lacks the experience to do ourselves. The system accuracy will be met for all three eventually, however, the remaining plates may require rework to ensure this .CONCLUSIONS AND RECOMMENDATIONSWhen we started this project, we were far from metrology experts. The first thing we would do differently if we were to repeat this project is to meet with subject matter experts (SME’s) much sooner. For the first month or so of MSDI, we were brainstorming amongst ourselves and with our customer. None of us knew what the best solution was, and when we met with SME’s down the road, we learned that there were a plethora of solutions we had never considered because we did not know they existed. SME’s were also quick to point out parts of our working concept that were infeasible or too complicated to work properly. Another thing we would do differently is to push our customer to help us fully define the problem early on. At one point, we thought there was a need for two or three inspection stations, but once we understood the problem more, we realized that one station would be sufficient, which greatly shortened our fabrication time. At another point in time, we were under the impression that we needed to measure points on the jaw surfaces in three dimensions. That is, we would need to define the location of each point measured, not just the deflection at said point. This confusion made us investigate and get hung up on many more complicated solutions. Had we known that we did not need to measure the locations of the points, we would have ignored that component of the design and focused on the relevant components, reaching a final solution much faster. Having now completed the fabrication of all components, we would be much more efficient if we had to make them again. We were all pretty inexperienced at machining when we started, especially machining steel. There are a lot of basic questions we would not need to wait to ask the full time machine shop staff before we started machining a part. This knowledge would have also been beneficial while we were designing the parts. Some of the geometries we came up with were difficult to machine, but could have been designed another way to make machining easier. Some of these changes were identified and changed before we machined them, but there are still some changes we could make to ease with the fabrication of a second inspection station. The calibration plates were made almost as an afterthought. While they are going to be installed and removed frequently, and are likely to be scratched or dented by the indicator tip after much use, they have not been heat treated. This is partially due to time constraints, but is also because we used generic angle stock rather than 8620 steel. If the calibration plates wear out, it is recommend that they are redesigned before being replaced. The new design could be based closer on the design of the actual jaws, which will make it easier for the operators to mount them. Also, 8620 steel should be used so they can be heat treated. There are two main weak links in our final inspection station. The first is the wooden table surface and supporting structure. While the table is rated for the combined weight of the calibration fixture, it is not entirely stable. When moving the height gage along the granite, the table rocks back and forth, and the number on the indicator can move several ten-thousandths of an inch while the table rocks, even if the indicator tip does move along the calibration plate or jaw surface. When the station is delivered to our customer, they plan to bolt it to the floor and make any other accommodations to reduce this problem, but to what extent they can fix it is yet to be determined. A new, stronger table may need to be purchased if the issue persists. The wooden tabletop itself may be the main problem, as it seems that the granite and calibration fixture are making the surface flex a small amount when the table is bumped. Both components are individually solid and resist deflections as required, but since wood is soft, it is possible that each component moves and settles differently each time the table is bumped. The ideal solution would be to purchase a larger granite surface plate, and include threaded inserts for the calibration fixture to screw into. With this set up, any movement in the table should not change how the fixture is positioned relative to the granite, since the contacting surfaces are much harder than wood and are much less likely to deflect. This solution should be explored if a stronger table does not mitigate the deflections caused by the table moving. The other weak link is the indicator tip. Because of the geometry of some jaws, we are not able to reach the surface of interest from directly above it. An angled tip was purchased and allows us to reach all required jaw surfaces. The issue is that, since the angled tip is a thin piece of metal, it seems to deflect as the indicator gage is loaded, and the deflection varies based on how much to gage is loaded. To solve this, a more robust tip should be used. Our customer may be able to make one in house, but a metrology company is also capable of making a custom tip. This is a relatively inexpensive fix, and will remove a large amount of the uncertainty in the measurements. It has only not been implemented due to time constraints. Ultimately, we were able to deliver an inspection station that accommodates jaws for the three required machines, that can be calibrated to create a horizontal plane parallel to the granite surface plate, and measure gross deflections across horizontal jaw surfaces within 0.001”.REFERENCES"The Gleason Works." Gleason Corporation, 1 May 2014. Web. 10 May 2015.Molony, E., Smith, J., Perry, D., Baldwin, K. 2015, “Gleason Gear Metrics.” , Ferdinand P., and Russell Johnston. "Chapter 5." Mechanics of Materials. 6th ed. S.l.: Mcgraw-Hill, 2011. 314-17. Print.ACKNOWLEDGMENTSWe would like to thank the following people without whom this project would have failed. Richard Russell (customer and primary Gleason customer contact), Gerald Garavuso (project advisor and mentor), Rob Kraynik (primary subject matter expert and mentor), Jan Maneti (tooling expert and machinist), ................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.
Related searches
- seton institute of reconstructive surgery
- michigan institute of real estate classes
- sports institute of tucson
- institute of physics iop
- institute of physics uk
- american institute of physics
- american institute of physics citation
- american institute of physics inc
- chicago institute of plastic surgery
- indian institute of public health
- nigerian institute of international affairs
- eric institute of education science