ShiftHappens

Test Results

Pulley Test Stand


    Tests Performed

  • Belt Friction
  • Spring Force

    Tests Designed

  • Power Requirements
  • Comparable Efficiency
  • Belt Sizing

Belt Friction


Objective: Determine if the finish on the manufactured pulley sheaves and the belt tested will have adequate friction to withstand the max torque a human could apply

Procedure: Use a torque wrench to turn the pulley half's while applying friction to the plates until the belt slips on the surface of the pulley half's. Read the reading off the torque wrench and measure the belt contact area on the pulley sheaves to determine the force required to make the belt slip.

Results:



Analysis: The results of this test were unsettling because 15 ft*lb of torque is not very much. In a previous test on a stationary bike, we determined that an average human can produce about 40 ft*lb of torque to the pedals of a bicycle. This is a lot higher than the average torque the CVT took before the belt slipped. This shows that the CVT would have a hard time getting going from a complete stop.

Spring Force


Objective: Determine the requirements of the spring force needed pressing against the rear pulley half in order to maintain the correct amount of friction on the belt.

Procedure: Estimate the force needed to get adequate friction. Test a multitude of springs to find the right spring constant needed to maintain sufficient friction through the range of gears.

Results:





Analysis: The results with 6 springs pressing on the rear pulley halves at a lower gear ratio are the most encouraging data sets, but the average torque required to cause the belt to slip is still 10 ft*lb below the calculated torque an average human can produce. Although a step in the right direction, this data does not prove that a CVT on a bicycle will work.

Power Requirements


Objective: Determine what the requirements will be for a test motor to operate the belt pulley system.

Procedure: Test the drill we have acquired to see if it will spin our sheaves at the equivalent rate a human would pedal. If not, find a way to gear down the motor or find a new motor to run our test with.

Target Value: The target RPM of the drill will be consistent with a human pedaling speed at a range of roughly 50-100 RPM.

Comparable Efficiency


Objective: Determine the efficiency of the belt/pulley system compared to the standardized chain/sprocket drive system.

Procedure: Determine a friction force to use for both the belt/pulley and chain/sprocket system. Determine the same gear ratio for both systems so the extracted data will be comparable.

Target Value: The efficiency of the belt will be no more than 10% less efficient than a chain driven bicycle.

Belt Sizing


Objective: Determine if the width of a scooter belt will be able to meet our requirements for range of gear ratios.

Procedure: Test the belt in its lowest and highest gear ratio. See what the ratios are and determine if we want more adjustability then we might need to find/design a new, wider belt.

Target Value: Ideally we will find a common belt that will facilitate our requirements and will be easily purchasable.

Stepper Motor Test Stand


    Tests Performed

  • Power Consumption
  • Energy Storage
  • Stepper Reliability

Power Consumption


Objective: Determine the maximum instantaneous power drawn by the stepper motor to assure that a supply can sustain it.

Procedure: Connect a small resistor in series with stepper motor controller. Capture the voltage on either side while the motor is under load. Export scope data to Excel to compute current and Power.

Results:





Analysis: In the first test minimal resistance was used, results are for compression of the springs. Average power consumption from a steady state sample was 12.14W, while average over the entire shift was 9.94W. Maximum instantaneous power was 14.06W. The increase in current in the beginning seems to be based on the acceleration of the motor. It then seems to level out and spring force has little effect on power consumption. This concurs with the assumption that the controller attempts to draw constant current.

In the second test the load was not increased by much but the controller current was increased. Power consumption more than doubled to 28.16W over the entire shift. While reviewing the data I noticed that the power dissipated by the resistor was very high, 1.31 W over the shift, with a maximum of 2.59 W. the resistor was likely rated to only handle 1/4 watt, or possibly 1/2 watt. The higher power dissipation likely caused the resistor to burn up and the data to be invalid. After re-measuring the resistor that seems to be what has happened. This may be what caused the controller to draw the maximum amount of power set by the pot. The modified test stand should consume less power, but to ensure that this does not happen again the best option would be to repeat the test for the new screw and this time use a parallel combination of resistors to dissipate power, a bank of 10 10ohm resistors should be sufficient and provide an equivalent resistance to the 1ohm resistor used in the test.

Energy Storage


Objective: Determine energy required to shift the motor different distances. This can be used to determine the energy storage capacity and charge rate when combined with a typical ride shift profile

Procedure: Same as Power Consumption, use a lower resolution to allow a longer time interval. Integrate power curves for full, have and partial shifts in both directions.

Results:





Analysis: The data for decompression is not as linear as the results for compression. The motor is helped by the springs when decompressing, but this is not seen fully until the motor is running at full speed. This makes it so the energy is not reduced as much when the motor is running in shorter bursts. The transition seems to be somewhere around 2 and 4 but with the limited amount of data points it is difficult to estimate lengths in this area. To provide a better representation of this, it would be a good idea to take a larger collection of data points, including 3, 5 and 6 starts. Going further than 9 would also be a good idea if it is determined that typical shifts are much shorter. Adjusting the rate of acceleration could also affect the outcome of this test so the test should be repeated if that needs to be modified after further testing.

Stepper Reliability


Objective: Ensure that the stepper motor is not stepping without turning. This would result in unnecessary power consumption and extra stress on the motor.

Procedure: Measure the distance traveled by the spring plate. Record the number of steps taken and compare to compute expected value. Move the motor back to the original position and compare this to Zero. Do this for several different distances

Results:



Analysis: The results show that the average error was 8.17%, this is higher than expected. After further analysis, the variance between tests looks to be high. When the test was performed the dial indicator was not able to be properly anchored to the table. Movement while holding it in place may be the cause for much of the error. This test should be repeated once other modifications to the test stand have been completed.