It’s a Question of Balance-

Dynamic Propeller Balancing

All builders will face this problem as they transition from building to flight.

There are balancing services that justifiably charge several hundred dollars due to the time and equipment needed, but this solution only works until you change something.

There are complete turnkey balancing systems that cost thousands of dollars and get the job done.

However, if you want to invest the least and learn the most you may prefer this Do-it-Yourself approach.

An added benefit goes to those of you that are looking for an awesome science project to share with your children.

My method employs a low cost two-channel digital storage oscilloscope, an accelerometer, and an infrared Emitter/Receiver diode/transistor pair to produce the strobe signal. If you have to buy everything, it will cost less than $400. If you can borrow the oscilloscope it can be done for less than $40.

Objective:

  1. Reduce Mass Unbalance conditions with respect to Propeller, Spinner and Engine Combinations.
  2. The dynamic balancing procedure must provide accurate information about the relative mass unbalance and provide a phase angle reference with respect to a point on the rotor system.
  3. Accuracy and cost are important considerations.

Given:

  1. Vibration is destructive to all avionics, airframe parts and passengers.
  2. An unbalanced disk rotating concentric about a drive shaft will produce a disturbance that is synchronized with the fundamental forcing frequency. The forcing frequency of interest is revolutions per second or hertz. (RPM/60).
  3. The disturbance produces a sine wave when measured with an accelerometer on the X or Y-axes when the Z-axis is parallel to the drive shaft.
  4. The measured signal will be at its maximum as the heavy spot on the rotor passes the axes and it is proportional to the magnitude of the unbalance.
  5. A strobe signal is recorded simultaneously on the second channel of the oscilloscope and provides the reference point to calculate the phase angle of the disturbance.

Preparation:

There are some items that need to be taken care of prior to any balancing procedures.

  1. Tracking: Use an index pointer to check that your propeller blades are tracking within 1/16 th of an inch of each other and that they are indeed the same length.
  2. Pitch: Each blade of the propeller must be set to the exact same pitch. Use a blade protractor or possibly a laser level to check this.
  3. Alignment: The spinner must be perfectly aligned. On my William Wynne Corvair Engine, I use a 13-inch spinner that is 16 inches from base to tip. If this is off by 1/8 th inch in its mounting it will produce a ¾ inch wobble at the tip along with an unbalance that is increasing by the square of the RPM’s. Set up an index pointer to measure the alignment at the tip. Zero displacement is needed before trying to do any balancing.
  4. Static balance: The propeller should be static balanced prior to mounting it on the aircraft. Often an additional coat of paint on one blade tip is all that is needed.

System:

At first I thought it would be necessary to take instantaneous measurements of both the X and the Y-axes in sync with the strobe event. This would work but it is doing it the hard way. Instead we will record either the X-axis or the Y-axis on channel #1 of the oscilloscope and record the Strobe (tach) signal event on channel #2. When we examine the waveform of the X-axis, realize that any heavy spot on the rotor system will produce the maximum voltage in the X direction as it passes the positive X-axis and the minimum where it crosses the negative X-axis. Anywhere else the signal has a Y-axis component and the voltage will be lower. The same is true for the Y-axis except the maximum voltage will occur at the moment that it passes the positive Y-axis. Once I realized this principal, the problem can be simplified. We need only one axis to measure the magnitude of the unbalance andthe same axis to measure the time interval from the strobe event to the maximum or minimum voltages in order to calculate the phase angle.

A model of the system is a fun science fair project or kitchen table project. It is very instructive and gives you a chance to get some practice. I used a model airplane electric motor and a wooden prop with the same sensors that I used on the actual aircraft. I attached an aluminum disk to the face of the propeller. This allows me to quickly change the test weights.

Finding the heavy spot-

The time interval from one strobe signal to the next is the period. This is one revolution. The frequency in Hz is f=1/period. The time interval from the strobe signal to the peak voltage,divided by the period, and multiplied by 360 degrees/revolution gives us the phase angle after the strobe event. Or, the number of degrees after the strobe event before the heaviest part of the rotor system arrives at the Positive X-axis. In other words, placing the rotor at the strobe event and then advancing the rotor the calculated number of degrees will place the heaviest part of the rotor over the Positive X-axis. Corrective weights are added 180 degrees opposite to the calculated heavy spot.

Procedure:

  • Place the photo reflective marker onto the backside of one propeller blade or starter ring gear or the back of the spinner bulkhead. This will become the reference point of the rotor system andcreates the strobe event as this passes the receiver.
  • Mount the photo emitter/receiver to activate when the reflective marker makes its pass. With the propeller positioned at the strobe event place a small marker at 12 o’clock position on the spinner. This will be the ZEROdegree reference. When the strobe event occurs thispoint will always beat zero degrees and corresponds to the zero degree point of the EB-6 wind calculator.
  • Mount the circuit board holding the Accelerometer so that the Z-axis is parallel to the drive shaft, the X-axis is in the horizontal plane and the Y-axis in the vertical plane. This should be as close as possible to the front bearing. Observe the orientation of the positive direction of both axes. A right rotation Corvair engine when viewed from the pilot’s position will have the X & Y axes in the sameorientation as a standard Cartesian coordinate graph. This orientation will later allow you to directly use an EB-6 Wind Correction Computer to make field calculations.
  • Record the voltages of the X & Y axes while at rest. You will see that the Y-Axis reading is 300 mV lower than the X-Axis. The system is measuring gravity while it is at rest. The accelerometer has a resting voltage that is usually ½ of the supply voltage. With a supply voltage of 3.2 volts, the resting voltage for the X-Axis will be 1.6-volts. The Y-Axis will be (1.6 – 0.300) = 1.3 volts due the effect of gravity. A gravity of 1G isyields a 300 mV difference in output voltage with my transducer.
  • For initial testing I would take measurements at a reduced RPM since any unbalance is detectable even at lower frequencies and it is safer.
  • All comparisons should be made at the same RPM so that the measurement of the magnitude of the unbalance is directly comparable.
  • Take final measurements at RPM settings that are close to cruise settings in both the X and Y-axis. Watch for harmonics as you pass thru different RPM ranges, as these areas should be avoided if they exist.
  • Save the data for later analysis, comparison and documentation.
  • I have written an Excel Spreadsheet that allows you to import the Accelerometer and Strobe data. This program processes 4000 data samples taken over a 0.4 second interval and adjusts the data for the zero point reference, passes the data through a low pass filter which greatly improves the readability of the wave form, and it performs a first integration of the accelerometer wave form to convert the accelerometer data into industry standard vibration measured in inches/sec. Three graphs are used to document and analyze the data. The first is the raw accelerometer data, then the filtered accelerometer data, finally a velocity graph. These allow you to compare the results that you achieve between successive runs.
  • When working with the waveform, keep in mind that a propeller unbalance is a 1X frequency. It will produce the same disturbance each revolution. The pistons on the other hand in a 4-cycle engine produce a ½ X disturbance since it takes two revolutions for each piston to fire.
  • Work to eliminate the largest 1X wave first. The resultant wave may look quite different.
  • Using the cursors, take measurements of the waveforms at their peak voltages and record the difference above or below the resting state voltages. Measure the time interval from the channel 2 Strobe Event to the Peak Values seen on channel 1. Record the duration of a single period of the channel 2 waveform.
  • Calculate the Phase Angle as a ratio of Time from strobe to peak waveform voltage divided by Time from Strobe to Strobe event. Then multiply this by 360 degrees/revolution to get the degrees after the Strobe Event before the heaviest part of the rotor system arrives at the X-axis.
  • An EB-6 wind calculator is useful at this time to help visualize the solution. With zero degrees at the top, rotate the disk by the Phase Angle in the direction of rotation. The heavy spot will align with the positive or negative X or Y-axis and the counterweight will be 180 degrees opposite. The degree mark at the negative axis is the number of degrees from the zero degree tick mark to the counterweight.
  • Applysolution to the aircraft. After proper engine shutdown, move the propeller until the strobe is triggered and the zero degree mark is at the top, move it further, in the direction of rotation, by theamount of the calculated phase angle. The heavy spot will now align with the positive X-axis or the positive Y-axis depending upon which axis you are working with. Place the corrective weight 180 degrees opposite to the measured heavy spot.
  • Graph these measurements to scale on the Dynamic Balance Worksheet. This is a polar coordinate graphing of the data and looks like a paper version of the EB-6 wind calculator.
  • Rerun the same test and record the results for comparison. If your analysis of the previous run was correct, the waveform should be gone or reduced. It will take several runs to zero in on the best solution. Comparing the graphs is very helpful.
  • With a horizontally opposed engine the Y-axis may produce a cleaner waveform since each piston has more influence in the horizontal plane. Also windy days interfere with the testing.
  • Make X-Axis recordings for future reference. This will give you a signature of your pistons in a healthy running engine and can be useful if your engine develops a hard to diagnose problem in the future.

Materials List:

  1. Oscilloscope- 25 MHz, two channel, digital storage, data export and color LCD screen. ( , model INSTEK GDS-1022, $351.75) (Very useful for many electronic projects)
  2. San-Disk media storage card for Oscilloscope ($9)
  3. Triple-Axis Accelerometer Breakout board –ADXL330 or ADXL335 analog output from Sparkfun Electronics, $29.95 This accelerometer provides an option for on board low pass filtering which removes much of the higher frequency noise in a real life sample.
  4. Small perforated IC blank circuit board and IC socket ($3)
  5. Integrated Photo reflective LCD emitter/receiver ($34.00) and Photo reflective tape ($8.00) (range 10cm-2 Meters) (Caution: listed ranges tend to be required range due to focusing lens system.
  6. Make your own Photo-Tach from matched IR Emitter/Receiver ICs from www/sparkfun.com for $1.95 (range ½ to 1 inch). Paint the surface with flat black enamel and place a flat white tick mark at a single location to trigger the strobe event as it passes the emitter/receiver. I used the back of the ring gear.
  7. Balance Worksheets to graph results

Example:

The charts below show the type of information that you can collect in a real aircraft. The center chart has a large voltage spike that occurs prior to each strobe signal. This is a fundamental frequency since it occurs with each revolution. For example, if we say that the wave occurs 16ms after the strobe and the period happens to be 24 ms then 16/24 * 360 degrees/rev would give us the number of degrees after the strobe signal in the direction of rotation when the heavy spot on the rotor would arrive at the positive Y-axis and we would place the counterweight 180 degrees opposite this along the negative axis. Then we would rerun the test at the same speed to see the effect. In this case I placed a 10-gram weight at 6 inches from the axis of rotation and it yields the next set of graphs. I was able to eliminate approximately ½ of the wave, so I will add more weight at the same location and rerun the test again.

Notice the velocity graphs have been reduced to about ½ the amount that they were prior to the first weight. All of the graphs have improved. I do not have the data for the third run as I have made some changes and have to start over but I think you can see that this does indeed work.