RESNA 27th International Annual Confence

Technology & Disability: Research, Design, Practice & Policy

June 18 to June 22, 2004
Orlando, Florida


Design and Development of a Pneumatic Suspension Caster Wheelchair Fork

Jeremy Puhlman, Dr. Rory A. Cooper, Bill Ammer,
Mark McCartney, John Duncan
University of Pittsburgh,
Human Engineering Research Laboratories
Pittsburgh, Pennsylvania 15213

ABSTRACT

The objective of this project was to research, design, fabricate, test, and update the first pneumatic suspension caster fork to lessen the effects of shock that affect wheelchair users. The pneumatic suspension caster fork is a new innovative way of reducing these stresses experienced from wheelchair travel. This caster uses a radial design so that the two pieces of the caster can work in unison for the desired amount of shock absorption. The caster was designed for absorbing shock in both the vertical and horizontal directions simultaneously. This caster incorporates a pressure gage and a brass air tank valve for easy regulation of the air resistance within the piston casing.

Load deflection testing, both static and dynamic, was conducted at 60, 80, and 100 PSI. After analyzing the data, it was found that the caster initially absorbed the shock almost linearly and then leveled out at greater deflection during the dynamic trials. This same relationship was shown for the static trials as well. The results showed that the prototype was functional and could absorb reasonably high loads.

Figure A.
This picture shows a caster fork that utilizes spring suspension.  The spring is encapsulated where the caster fork joins the chair.  This caster fork is the Quickie R2.

BACKGROUND

The focus of this project was for active manual wheelchair users. There are four main parts of a manual wheelchair: the rear wheels, front caster wheels, the frame, and the seating system. The problem that researchers are facing is the reduction of shock that can cause further complications for the user. Some secondary complications include excessive pain and spasms. Shock affects both rear and front wheels of a wheelchair. The front caster is a significant contributor to vibration and pitching. This is apparent because the front casters are rigidly attached to the frame of the wheelchair.

As there are many types of chairs, there are many types of casters. There is the standard caster, which consists of a metal or plastic caster fork and a rubber wheel. It attaches to the chair via a bolt and nut. Suspension casters use springs and polyurethane shock absorbers to reduce vibration. The standard caster contributes to excessive pain and spasms as it transmits shock and vibration to the user. However, standard casters are inexpensive and are readily available on most manual wheelchairs. Suspension casters are more expensive as a result of the utilization of additional components and more complex designs.

Figure B
This picture shows a polyurethane suspension caster.  This caster was produced by froglegs.

The spring suspension systems that are on casters only prevent shock in the vertical direction, see figure A. The caster forks that use polyurethane has been designed to adsorb the shock in the vertical and horizontal directions, see figure B. The vibration that affects wheelchair users involves forces that are in both the vertical and horizontal directions following a radial pattern. This makes the suspension casters that are on the market today only 70% efficient. (7) Current suspension caster forks that use polyurethane shock absorbers are difficult to tune for each individual user and to changes in the user center of mass within the wheelchair. This results in a wide range of performances for individuals.

RESEARCH QUESTION

Figure 3. (Click image for larger view)
Figure 3: This picture shows the first preliminary sketch of the pneumatic suspension caster fork.  This sketch shows a radial design that utilizes a piston fixed with a roller for suspension.

Current suspension caster forks are severely limited in their ability to be tuned to the individual, the set-up of the wheelchair (e.g., center of mass location of the wheelchair and user), and the driving scenario (e.g., indoors, outdoor rough surface). The objective of this project was to research, design, fabricate, test, and update a pneumatic suspension caster fork to lessen the effects of shock that affect wheelchair users. The primary benefit to a successful pneumatic suspension caster fork is that it can be easily tuned by adding or releasing air from the system to stiffen or soften the suspension. Through using a simple Schrader style air valve a small hand air pump could be used to add air to the system by the user or an assistant to alter the air pressure to the users preferential value for the activity that he or she is performing.

Figure 4. (Click image for larger view)
Figure 4: This figure shows the second preliminary sketch of the pneumatic suspension caster fork.  This sketch shows dimensions and portrays a more clear vision of what the pneumatic suspension caster fork was going to look like.

METHOD

The design required precise geometry to absorb shock in the horizontal and vertical directions simultaneously. To assure accuracy in machining the parts, they were to be drawn in FeatureCam* before machining could take place, the materials of all the parts were to be chosen so that the caster would be lightweight and cost efficient. After the first generation was prototyped, it was to be analyzed and redesigned so that a second prototype could be built. Once the second generation was built, two more units were to be constructed for testing on a manual wheelchair.

The tests would show that the caster could withstand loads for 3 to 5 years, while not losing air pressure. Tests were also conducted to determine caster deflection when exposed to set loads and cyclic loads. The first sketches of the caster are visible in Figures 3 & 4. From these sketches a model was constructed using computer aided design software (FeatureCam & Solidworks**). An assembly view and exploded view of the pneumatic caster are pictured in Figure 5 & 6 respectively. A conceptual model was designed and the materials that were chosen were organized in a materials table, see Table 1. The material dimensions, features, and solids could be added to write the numerically controlled programs for manufacturing. All of the parts were cut from stock metal and cut to the desired geometry using the following machines: computer numerically controlled and manual mill and lathe, wire electrostatic discharge machine (EDM), band saw, and drill press.

Figure 5 (click image for larger view)
Figure 5: This picture is a SolidWorks modal of the pneumatic suspension caster fork.  Each part was designed separately and assembled using different colors to show contrast.
Figure 6 (Click image for larger view)
Figure 6: This is an exploded view of the pneumatic caster in SolidWorks.  All of the components are labeled for an easy interpretation of how this caster is assembled.





Table 1. Parts List
source material/Part #/Size operations # parts drawing in file
        ortho (dimentions) solid
HERL 1.50" - 6061 Aluminum Plate Cut, CNC Mill, and Mill 2 y y
HERL 1.0" Delrin Rod Lathe and Drill 2 y y
McMaster 9464K73   6 na na
HERL 1.50" - 6061 Aluminum Rod Lathe (Taper) 2 y y
McMaster 8063K33   2 na na
McMaster 38105K14   2 na na
HERL 1.5" - 304 Stainless Steel Rod Lathe, Drill, Hone   y y
           
HERL 1.50" - 6061 Aluminum Plate Cut, CNC Mill, and Mill 2 y y
Gatto Cycle Shop 76MM/82A   2 na na
           
HERL 10-32 X 1 3/4   2 na na
HERL 10-32 lock nut   2 na na
           
HERL 0.125" - 1018 Steel Plate Mill and Drill 2 y y
HERL 0.5-13 Grade 5 Steel Bolt NA 2 na na
           
McMaster 4591K11   1 na na
HERL Translucent   1 na na
Figure 7
Figure 7: This is a snapshot of the first prototype.  It consists of the air valve, roller piston, pneumatic housing, wheel, and caster wheel housing.

The first functional prototype is pictured in Figure 7. Through brainstorming, it was apparent that some changes needed to be made. The first change is that the first prototype did not have sufficient room to put a bolt to attach to the chair. The solution came by increasing the size of the top caster. Also a steel plate was designed so that the bolt could be welded onto the plate and then screwed into the top of the top portion of the caster fork. The next change that was made was to the piston. The problem with the piston was that the roller did not remain straight after repeated loading. It would turn and dig into the surface of the bottom caster fork. The new material that was chosen was delrin. By switch to Delrin TM and machining the end of the piston into a half sphere, it could slide along the surface of the bottom caster without damaging the aluminum fork that it was in contact with. The second generation was constructed with the newly designed parts and ideas. The second generation can be seen in Figure 8. The second prototype proved to hold air more effectively and could be tested for load deflection.

Figure 8
Figure 8: This is a snapshot of the second prototype.  This prototype is the design that was tested and assembled to a chair.  The differences from the first prototype are the caster-to-wheelchair attachment, mini air gauge, plastic piston, and a newly designed pneumatic housing.

The method of testing this unit required the use of computer numerically controlled press machine (Instron, model # 8872). The caster fork was bolted to a steel plate that attached to the base of the Instron machine. A hole was drilled in that plate so that the bolt of the caster fork would slide through and be secured by a locknut. The caster was positioned so that the pressure gauge and brass air tank valve were facing up. The w * heel was centered under the load cell. Testing of the pneumatic casters used the same protocol as that described by Blauch et al. (1). The data were compared to the theoretical analysis derived using the “Ideal Gas Law”

RESULTS

The data recorded from the static load/deflection test are presented on Table 2 and Figure 1. The data recorded from the dynamic load/deflection test are represented in Table 3 and Figure 2. The cyclic load was applied at 1.5 Hz to mimic the loading from an American National Standards Institute (ANSI) and RESNA double-drum fatigue test (3). The deflection and amplitude were measured using linear-variable displacement transformer and a load cell with an oscilloscope. The “Ideal Gas Law” was used to determine how many moles of air were in the caster for each of the different pressures tested. Using the following equation, derived from the ideal gas law and P=F/A, Force can be represented as a function of the travel of the piston expressed in Equation 1. The graph of the results derived using this equation is shown in Figure 3.

Figure 1g: This graph shows the Static Load/Deflection test data at 40, 60, 80, and 100 PSI.  The data shows the load increasing fast over small deflections at the start and then load increases steadily for 0.5” of deflection and then increases drastically again at the end of deflection. Figure 2g: This graph shows the Dynamic Load/Deflection test data at 60, 80, and 100 PSI.  The graph shows a similar relation between the pressures.  Roughly the data for each pressure is increasing linearly. Figure 3g: This graph is illustrating equation 1.  In this graph force is represented as a function of the distance the piston travels at any given time.
Table 2.
40 PSI
Load (lbs) Low Deflection (in)
5 18 0
10 38 0.0124
15 62 0.0286
20 86 0.0603
25 106 0.1274
30 126 0.2839
35 150 0.6859
40 170 0.8899
45 190 0.9609
60 PSI
Load (lbs) Low Deflection (in)
5 18 0
10 42 0.0114
15 62 0.0249
20 86 0.0397
25 106 0.0641
30 126 0.099
35 150 0.1452
40 170 0.3936
45 190 0.7086
50 214 0.9276
55 238 0.9886
80 PSI
Load (lbs) Low Deflection (in)
5 18 0
10 38 0.012
15 62 0.0242
20 86 0.0359
25 106 0.0499
30 126 0.0679
35 150 0.0959
40 174 0.1499
45 190 0.1619
50 218 0.4069
55 238 0.6469
60 258 0.8869
65 278 0.9949
70 302 1.0239
100 PSI
Load (lbs) Low Deflection (in)
5 22 0
10 42 0.0114
20 82 0.0318
30 126 0.0533
40 170 0.0864
50 218 0.1634
60 258 0.3494
70 306 0.8494
80 346 1.0394
Table 3.
60 PSI
Load (lbs) Deflection (in)
12.7 0.05
32.73 0.11
45 0.232
41.47 0.338
36.58 0.438
43.62 0.539
43.34 0.64
46.65 0.74
46.92 0.84
48.16 0.94
80 PSI
Load (lbs) Deflection (in)
26.21 0.052
37.04 0.102
58.8 0.202
64.76 0.337
61.98 0.437
58.2 0.539
61.34 0.638
63.85 0.739
61.8 0.839
63.7 0.939
100 PSI
Load (lbs) Deflection (in)
47.59 0.1
62.73 0.2
69.31 0.3376
63.5 0.438
69.89 0.533
69.43 0.638
72.33 0.739
75.342 0.839
75.6 0.939

DISCUSSION

In both the static load and deflection and the dynamic load and deflection tests, the pressure exhibited the desired response. The pressure would increase and decrease under deflection and return with little hysteresis. Pre-loading the chamber with a base pressure greater than of atmospheric pressure controlled for hysteresis. This is ideal for wheelchair users that are traveling over uneven terrain. The caster air chamber cannot be under filled because even when uncharged (i.e., when simply filled with air with zero relative air pressure) there is sufficient pressure built upon during compression of the piston within the chamber to prevent bottoming out. Hence, this prevents the effects of shock from being just as damaging as the shock produced by standard casters.

One explanation of why the data did not follow theory is because the piston required more force to start moving on the radial surface of the bottom portion of the caster fork. The static friction was greater at the start of movement than when the piston was displaced a certain distance on the surface of the caster fork. Another possible error is imperfections in the geometry of sliding surface of the caster fork during fabrication. As the piston follows the machined path, the curvature may be steeper and not at a consistent radius as compared the theoretical. This would consequentially cause the piston to displace unevenly along the radius. This may cause the load to increase more at smaller displacements and less at larger displacements. Also, the model assumed linear displacements at the piston, whereas the data collected on the Instron record linear displacements at the caster. Greater agreement would result if the geometry of the caster fork were accounted for in the displacement curves.

The force that is experienced by the caster wheel when in contact with a solid step is 400 lbf. The moment that the caster fork experiences from the caster wheel impact is 200 lbf-ft.(4) The testing suggested that at the pressures analyzed the highest pound force experienced was 80 lbf. This means that the pneumatic suspension caster fork did not apply enough resistance, via air pressure, to absorb the suggested loads that the caster fork experiences. Therefore, high air pressures may be necessary for heavier individuals who are traveling over rough surfaces. Future testing should include air pressures in the 200-300 PSI range.

REFERENCES

  1. Blauch C, Cooper RA , Guo S, McCartney M, Wolf E, Dvorznak MJ, Mechanical Properties of Polyurethane Elastomers in an Oblique Angled Suspension Caster Fork for Wheelchairs, 25 th Annual RESNA Conference, Minneapolis, MN, pp. 297-299, 2002.
  2. VanSickle DP, Cooper RA , Boninger ML, Road Loads Acting on Manual Wheelchairs, IEEE Transactions on Rehabilitation Engineering, Vol. 8, No. 3, pp. 385-393, 2000.
  3. http://www.vard.org/tts/avail/01-092.htm
  4. Dwight O. Johnson, MSME, Consultant to Wheeled Mobility Center San Francisco State University San Francisco, California 94132 “Simplified Strength Testing of Manual Wheelchairs in Developing Countries.

ACKNOWLEDGMENTS

This project was conducted at the University of Pittsburgh's Human Engineering Research Laboratories under the supervision of Dr. Rory A. Cooper

Jeremy R. Puhlman, BS,
University of Pittsburgh,
Pittsburgh, PA 15260,
Office Phone (412) 365-4836
EMAIL: jrpst46@pitt.edu

 

RESNA Conference Logo