 Testing
Testing was done, using our design shown above, to ensure the sucess of our design. Our intial test run was a sample size of 20 parts. After the parts were staked we had to be sure they passed the quality controls GT has for roller finger follower staking. We tested each part for depth of the staking, how much torque was required to turn the axle, and the force needed to push the axle out of the housing.
The first test performed was the depth of staking. GT’s quality controls say the depth of staking must fall within 3mm to 5mm. This depth is meant to ensure the axle is set in place but is not over staked as to damage the roller finger follower housing. Also, staking depth has a correlation to the required amount of torque needed to turn the axle. The picture below shows the testing device for determining the staking depth. The roller finger follower is secured on its side by two V-Blocks. The depth micrometer is first zeroed out on the flat surface of the axle. We then moved the axle horizontally so the depth gauge fell into the staking ridges. In the test shown below the depth of the staking was determined to be 0.302 mm, which falls into the quality control specifications.
 
The next test performed was the torque to turn test. This is the most important test to determine if the staking was done properly. The quality of the part is mainly determined by the results of this test. This test measures the amount torque required to turn the axle. If staking was done properly the axle should require a large amount of torque to be turned free of the housing. The acceptable torque value is any value over 100 in.lbs.The test is performed by tightly clamping the axle ends and then turning the roller finger follower. The roller finger follower is pressed against the testing device to prevent it from twisting. A picture of the torque to turn test is shown below.
The final test performed was the axle push out test. This test measures the amount of force required to push the staked axle out of the housing. The passing parameters are a resulting force in the range of 2500-3250 lbs. The piece is secured on its side between two V-Blocks on a mounting block. The mounting block is then moved into the test machine and positioned so the axle is directly under a vertically positioned cylinder. When the test is started the cylinder pushes down on the axle until it drives it completely out of the housing. The pressure needed to expel the axle is shown on the digital display. The picture below shows the testing machine for the axle push out test.
Results
The most important test that the finger followers must pass is the torque to turn test. It requires that all of the axles fall within the range of 100-160 in-lbs. All of the parts passed this test and it might be tempting to dismiss them and deem them successful or “good parts.” This is not entirely true and will be revealed in the two other tests.
The push out test was operated with the first ten parts being pushed out from the right side and the other half pushed out from the left. This is an important fact because the data shows a clear difference between the two methods. On average the parts tested from the right side were pushed out with an average force of 2829 lbs while the one from the left were pushed out at 3641 lbs. The reason for the difference is due to the poor calibration and alignment of the staking tools. The material on the outside of the body (the left side in the diagram below) is softer and has no resistance to its expansion. On the contrary the material on the right is restricted by the body. This causes a small lip on the left side which makes the axle more difficult to remove from the left.
This discrepancy also agrees with the plot below which shows the increase in PO values during parts eleven through twenty. In comparison, the torque to turn values do not change from part to part. This fact proves that the parts were consistently staked uneven and again reinforces the concept above.
The final test shown above is the staking depth. This is the test that is of least concern. However, it gives great insight to the consistency of the staking process. In this case, this series of data is perhaps the most effective method in proving the uneven staking of the axles. The right side staking depth averaged a value of .36mm while the left side was considerably deeper at an average of .5mm. Again, because the axle was pushed out of the part on the left side before being staked the material had little opposition to expand while the right side had drastically more opposition. This allows the material on the left side to easily stake and flare out while the left stopped at a shallow depth. These values can be trusted based on their standard deviations and the graph below. The graph shows that most of the values hover around the average but without more samples this is not the best method of verification. However, the standard deviation for the right side staking depth is .08mm while the left side standard deviation is .05mm.
Although the staking depth needs adjusting (a simple fix) the main concern of this project has been deemed a success. The alignment cylinder which was in charge of keeping the parts aligned from top to bottom and left to right was flawless. None of the parts were improperly staked in either of the two directions. There was a 100% passing rate with respect to x and y axis alignment.
|
