I have completed a major rewrite/reformat of the CamAlign process. Here are the links:
Format text for Guiding by Right side of Rail
Format text for Guiding by Left side of Rail
Format text for Generic Instructions
I think that the new description will be more useful and easier to follow, but I am leaving this older version available.
This document describes a process for aligning the rear axles of a Pinewood Derby car. The axles may be either cambered or level. It also detects the presence of differential rear wheel frictions.
The process originated as an accurate method for aligning cambered rear axles, for which published methods were problemmatic. While working with the process, it became apparent that the process was equally effective for level axles. Moreover, it was easier to use than the shim alignment process that I introduced in my book, "Learn to Build a Winner," since its sensitivity did not require that the axles be exactly level. Consequently, it allowed direction adjustment by a wider variety of techniques, including twisting slightly bent axles, a historic method preferred by many.
The process is directly applicable to three wheel (4th wheel lifted) cars that guide using the center rail. With some thought the process can be adapted to "4-wheelers" (temporarily remove the 4th wheel) and to "straight runners" (temporarily replace the axle for the dominant front wheel with a slightly bent axle.)
The following procedure causes each rear wheel's contribution to direction to be most prominent in the observed test results. It is done by shifting the load temporarily from one axle to the other by adding a small cantilevered weight. A preliminary experiment showed that a rear wheel dominates directional control if it carries substantially more weight than the other rear wheel. The document shows alternatives for constructing a temporary cantilevered weight harness and how to use them in aligning the rear wheels.
The procedure says to "adjust the angle of the axle", i.e. alter the toe angle (the fore-aft angle) of the axle. You may do this using your choice of methods, such as shimming or twisting a bent axle. If the "bent axle" method is used, then the camber angle is changed simultaneously with the toe angle. If the axle does not stray far from the "up" or 12-o'clock direction, then the effect upon the camber angle is small enough that it can probably be ignored. I will describe using the "bent axle" method.
While following this procedure, you may uncover a source of "differential friction" in the rear axles. Differential friction can cause the racer to steer in directions other than that attempted by the front wheels. The process partially isolates a specific rear axle as the source. If such sources are found, correct that defect before trying to complete the alignment.
1. Prepare Test Ramp - Construct a test ramp with a smooth center rail, e.g. a common wooden yard stick, and an adjustable drop of about 1/2 inch to 2 inches over its 2 or 3 foot length. Using a straight edge and a fine soft tip pen (such as a Sharpie), draw a narrow line 1/16" away from the rail on the one side of the rail. Set your car on the ramp with the inner edge of one rear wheel aligned with the line. Check that wheels are gently at the ends of the axles and mark the ramp where the inner edge of the other rear wheel rests. Extend this mark equal distance from the first line for the length of the ramp. These will be the target wheel paths. Use care to not groove the wood by heavy pressure from a pen or pencil. A fine tip felt pen is probably best. Note that if you want to align a car with the DFW on the other side of the racer, just flip the ramp end for end!
2. Prepare Adjustable Weight - Construct an adjustable cantilevered weight assembly (or pair) which functions to alter the rear wheel weight distribution to 3:1 in favor of either rear wheel. If necessary, adapt the car to the weight assembly.
3. Setup Test Ramp - Identify the racer's dominant front wheel (DFW). Set the DFW for mild steer into the rail (rail guiding or Rail-Riding or RR). Orient the test ramp so that the DFW is on the same side of the ramp as the 1/16" line when the nose is pointed down-ramp.
4. Setup Adjustable Weight - Temporarily attach the cantilevered weight harness to the rear of the car. Iterate through steps 5 and 6, starting with the rear wheel that normally carries the greatest load, then alternating sides until no improvement is obtained. Each iteration should require progressively smaller adjustments. Note that if no special attention was given during design and construction, the wheel on the opposite side from the DFW probably carries the most weight!
5. Adjust Weight for Trial - Temporarily adjust the cantilevered weight so that the selected rear wheel carries 3/4 of the rear weight. When weighing, the car should be in the same orientation as the track, i.e. slightly nose down, and the front wheel weight should be unchanged from its original value. (The result is that this rear wheel will control the path followed by the racer's rear end. If misalignment causes binding and causes a wheel to slip from its own path, the extra load on this wheel should assure that this will not be the wheel that slides.)
6. Run Trial and Adjust Alignment - Place the car on the test ramp with the DFW gently against the rail and the target wheel aligned to its line. Release the car. If necessary increase the slope of the ramp. The DFW should stay on the rail and the rear wheel should stay on its line. If either strays from its line, act on the wheel that strayed first. If the DFW strays from its line first, increase its toe in and repeat the test. If the rear wheel strays from its line first, adjust its toe angle until it does stay on line.
7. Refine Test Ramp, If Necessary - By now, rear end tracking should be good enough and rolling should be free enough that the ramp slope can be reduced. Reduce the ramp slope to the smallest amount at which the car will run the ramp. Repeat the alignment tests in steps 4 through 6 until the wheels hold their lines.
This section attempts to tell whether the net or composite coefficients of friction of the rear wheels differ from each other. The idea of "net" or "composite coefficient of friction" is that this coefficient reflects the result of the various causes of friction. It does not reflect the effect of different amounts of weight carried by each wheel nor does it include misalignment, since that has already been corrected by the procedure above!
Some causes of differential frictions:
- difference in camber between the rear axles
- different smoothness of the axles or wheel bores or wheel hub faces
- different diameters of the rear wheels' bores
- different diameters of the rear wheels' treads
- different slopes of the axle head undersides
- different shapes of the tread contact patches with the track
1. Measure Unloaded. - On the drift test board, run the car several times to get the average drift.
2. Load Front - Install the "long arm" cantilever weight assembly and adjust it to load the dominant front wheel while leaving the rear wheel weights at their original value. The weight will be approximately on a line that passes through the car's original center of mass and the front wheel, but will be forward of the front wheels.
3. Measure Loaded - Again on the drift test board, run the car several times to get the average drift.
4. Analyze Losses - Compare the two drift measure averages. If they differ significantly, then one of the rear wheels rolls with a higher net coefficient of friction than the other. The deviation in step 3 tells which rear wheel is lossier: If the drift of the Unloaded runs is to the right of the drift of the Loaded runs, then the right rear wheel has a higher net coefficient of friction, and vice versa. This test does not tell the cause of the higher coefficient of friction, just that it is higher than on the other wheel. Since you did so well on the other wheel, you can probably improve this one! The good part is that "alignment" has been eliminated as a possible cause!)
1. Tune Front Wheel(s) - The next step in alignment is to adjust the dominant front wheel (DFW) toe-in for best performance. The toe-in and camber of the dominant front wheel is very sensitive to the car's weight distribution and to the track's roughness. A common toe-in for average conditions is about 2" of drift toward the guide rail in 4 feet of travel. Details on DFW alignment are outside the scope of this article. After setting the DFW toe-in, for best performance, it is prudent to rerun step 8 above to show that the DFW changes did not alter the rear wheel action. It is also prudent to watch the car carefully on a real track to verify that the rear wheels stay off the rail. If one of them does touch the rail, then its "line" on the alignment board needs to be adjusted outward by at least 1/16 inch, and the opposite side line needs to be adjusted inward by the same amount. Then run the entire alignment process again.
2. Lock Alignment - Finally, the rear wheels are aligned to their optimum. If you are inclined to glue axles in place, now is the time to glue in the rear axles. This is best done through a small hole drilled up from the bottom of the car and intersecting with the axle hole.
If two wheels roll together at slightly different toe angles, the wheel paths will converge or diverge. That is, the wheels will try to get closer together or farther apart. They are constrained by the car body and the ends of the axles from moving very far, but this constraint causes "pinching" and increases the friction between wheel and axle (an energy consuming brake!) and/or one of the wheels to slip on the track surface (an energy consuming skid) so that the racer can keep going forward. These frictions sap energy from the car's forward velocity and slow it down!
If two wheels roll together on parallel paths, but the paths don't match up with the path followed by the dominant front wheel as it follows its "straight and narrow" path, then the racer's rear end will flare to the left or to the right, mashing a rear wheel against the rail. That is more friction that uses up the energy of your car's forward velocity.
The procedure focuses on one rear wheel at a time. Weight is shifted to one wheel from the other to allow its friction to prevent the wheel from deviating far away from the path dictated by its axle orientation. Thus the path that the wheel follows is a strong indication of the axle's state of alignment.
Look at the typical weight carried by each wheel of a 5 ounce three-wheeler with a 4:1 weight distribution:
The dominant front wheel carries one ounce (including its own weight.)
The wheel following the dominant front wheel carries 1-1/2 ounces if the center of mass is on the midline of the racer.
The off-side rear wheel carries the other half of the racer's weight, 2-1/2 ounces.
Some folks will offset the ballast addition to more evenly divide the rear wheel weight between the rear wheels so that each carries about 2 ounces.
It is still possible that severe misalignment of the unloaded axle will allow it to affect the path of the loaded axle. The effect will be less than if both axles were unloaded, and this fact allows the process to converge by iterating the measurement and adjustment steps.
Intuitively, I suspect these relationships:
To prove these relationships, I propose Rear Wheel Dominance - Experiment 1.
In his 2008 article "The Art of Rail Riding", Jay Wiles described a "forward and reverse roll test" for measuring and adjusting rear wheel alignment. The process is applicable when rear axles have negative camber. The process iterates based on achieving a state in which the wheels "stay out" on both forward and reverse rolls. Then one can proclaim, "It's Aligned!"
Intuitively, I suspect these relationships:
To prove these relationships, I propose Cambered Rear Axles Alignment - Experiment 2.
Applying negative camber to the rear axles of a pinewood derby car has often shown to provide slightly better performance than if the axles are level. The improvement arises from the wheel rolling on its inner tread edge rather than on its full tread. Only a small amount of camber (3 degrees or less) is sufficient to provide the benefit. The camber increases the force between the wheel hub and the head of the axle. In most circumstances the increased friction between the wheel and axle head is more than offset by the reduction in friction between wheel and track.
A common problem encountered when using cambered axles is that it is more difficult to know when the alignment is optimum. The small advantage from cambered axles may be swamped by losses from slight misalignment.
I have published elsewhere a convergent process for achieving excellent alignment when axles are level to the track surface. The process in this document is an alternative when rear axles are cambered or when a front wheel in intentionally toed-in.
Every once in a while I encounter an advertisement that tries to convince us that the tools that they offer for sale eliminate the need for alignment. They might properly say that "some fine tuning of the front wheel to match the actual track, but" ... Well, you get the idea. That is a fine goal, but it doesn't require their expensive tools to accomplish. What it requires is the same knowledge that you need in order to use their expensive tools!
Most of us are still learning the techniques to produce perfect parts and perfect places to stick them!
The tool that mostly delivers on its promise is the inexpensive ProBody Tool for drilling true axle holes in a pinewood car block. With only modest care the tool should last for the racing career of a typical Cub Scout ... if not too many others drill their blocks with it.
Update: 2/13/2011 - Link to major rewrite as .php.
Update: 3/24/2010 - Add Unequal loss section.
Update: 3/10/2010 - Add instruction to Step 5.
Update: 3/6/2010 - Improve introduction, retitle.
Update: 3/1/2010 - Improve introduction and trailer.
Update: 2/28/2010 - Rewrite with experiments "proposed".
Update: 2/18/2010 - Rewrite with cantilevered weight and experiments.
Original page created: 3/9/2009
Copyright 2009, 2010 © by Stan Pope. All rights reserved.
Scouting organizations may print, duplicate and distribute copies of this document provided that this copyright notice remains intact and no fee, direct or indirect, is charged for the copies.