211Sd~~fi . ~/ - / ....

No. 13475PISTON AND RING ASSvY FRICTION STUDIES

IN CU1JTINS 903 ENGINECoNmTACT DAAE07-84-C-R134

JUNE 1989DONALD J. PATTERSONIKEVIN M. MORRISON

GEORGE B, ScHwAmRZ

DEPARTMENT OF MECHANICALENGINEERING AND APPLIED MECHANICS

TH UNIVERSITY OF MICHIGANANN ARBOR, MICHIGAN 48109-2121AND

U.S. ARw rANK-AUOMOTI VE CO•vAND

By WARREN, MICHIGAN 48397-50

APPROVED FOR PUBLIC RELEASE:

DISTRIBUTION IS UNLIMITED

U.S. ARMY TANK-AUTOMOTIVE COMMANDRESEARCH, DEVELOPMENT & ENGINEERING CENTERWarren, Michigan 48397-5000

REPRODUCTION QUALITY NOTICE

This document is the best quality available. The copy furnished

to DTIC contained pages that may have the following qualityproblems:

* Pages smaller or larger than normal.

0 Pages with background color or light colored printing.

* Pages with small type or poor printing; and or

* Pages with continuous tone material or colorphotographs.

Due to various output media available these conditions may or

may not cause poor legibility in the microfiche or hardcopy outputyou receive.

7 If this block is checked, the copy furnished to DTICcontained pages with color printing, that when reproduced inBlack and White, may change detail of the original copy.

I"ýNOTICES

This report is not to be construed as an official Department of the Army

position.

Mention of any trade names or manufacturers in this report shall not beconstrued as an official endorsement or approval of such products or companies

by the U.S. Government.

Destroy this report when it is no longer needed. Do not return it to the

originator.

Unclassified

SECURITY CLASSIFICATION OF THIS PAGE

Form ApprovedREPORT DOCUMENTATION PAGE OMB No. 0704-0188

la. REPORT SECURITY CLASSIFICATION lb. RESTRICTIVE MARKINGSUnclassified

2a. SECURITY CLASSIFICATION AUTHORITY 3. DISTRIBUTION/AVAILABILITY OF REPORT

Approved for public release;2b. DECLASSIFICATION / DOWNGRADING SCEDULE distribution is unlimited.

4. PERFORMING ORGANIZATION REPORT NUMBER(S) 5. MONITORING ORGANIZATION REPORT NUMBER(S)13475

6a. NAME OF PERFORMING ORGANIZATION 6b. OFFICE SYMBOL 7a. NAME OF MONITORING ORGANIZATIONDept. Mech. Engr. & Appl. Mech, (if applicable) AOUniversity of Michigan TACOM

6c. ADDRESS (City, State, and ZIP Code) 7b. ADDRESS (City, State, and ZIP Code)2250 G.G. Brown LaboratoryAnn Arbor, Michigan 48109-2125 Warren, Michigan 48397-5000

8a. NAME OF FUNDING/SPONSORING lb. OFFICE SYMBOL 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBERORGANIZATION (If applicable)

U.S. Army TACOM DAAE07-84-C-R134

Sc. ADDRESS (City, State, and ZIP Code) 10. SOURCE OF FUNDING NUMBERSPROGRAM I PROJECT TASK WORK UNITWarren, Michigan 48397-5000 ELEMENT NO. NO. NO. ACCESSION NO.

11. TITLE (Include Security Classification)

Piston and Ring Assembly Friction Studies in Cummins 903 Engine (u)

12. PERSONAL AUTHOR(S)Patterson, Donald J., Morrison, Kevin M., and Schwartz, George B.

13a. TYPE OF REPORT 13b. TIME COVERED. 14. DATE OF RE;ORT (Year, Month, Day) 15. PAGE COUNTFIAL .FROM Oct. 84TOOct. 8&. Jun58

16. SUPPLEMENTARY NOTATION

17. COSATI CODES 18. SUBJECT TERMS (Continue on reverse if necessary and identify by block number)FIELD GROUP SUB-GROUP Friction, Piston, Piston rings, Uncooled Diesel

19. ABSTRACT (Continue on reverse if necessary and identify by block number)

Piston and ring assembly friction has been measured in a single cylinder Cummins 903 diesel.The engine was motored and fired, cooled and uncooled. For uncooled operation, plasmasprayed chromium oxide ring and liner coatings were employed with a synthetic lubricant.The Instantaneous IMEP and Fixed Sleeve methods were used for the friction measurement ona crank angle by crank angle basis, with a cycle average of about 1.5 to 4 psi fmep. Con-siderable effort was expended in developing the measurement techniques, with a major efforton the Instantaneous IMEP method. In a parallel effort a bench type, heated, piston ringand liner wear simulator was developed. This permitted rapid screening of promisingcandidate materials for uncooled engine tests. The plasma sprayed chromium oxide coatingswere found to be best.

20. DISTRIBUTION/AVAILABILITY OF ABSTRACT 21. ABSTRACT SECURITY CLASSIFICATION0 UNCLASSIFIED/UNLIMITED 0 SAME AS RPT. [3 DTIC USERS Unclassified

22a. NAME OF RESPONSIBLE INDIVIDUAL 22b. TELEPHONE (Include Area Code) 22c. OFFICE SYMBOLE. Schwartz (313)574-5656 AMSTA-RGRD

DD Form 1473, JUN 86 Previous editions are obsolete. SECURITY CLASSIFICATION OF THIS PAGEUnclassified

2

PREFACE

The very important role of the Cummins Engine Company,especially Ralph Slone and Malcolm Naylor is gratefullyacknowledged. They supplied virtually all of the enginehardware, some experimental equipment, and the basic pistonand ring laboratory fixture.

Furthermore, the authors thank Professor Zu-Wei Cui ofZhejiang University, China, and Professor Kenneth Ludema ofThe University of Michigan, for their assistance withdevelopment of the piston and ring simulator and theanalysis of the results. Also, we wish to acknowledge thecontributions of Guy Babbitt, graduate student, for his workin designing, constructing, and evaluating the Fixed-Sleevemethod for the Cummins 903 engine.

We also thank Jack Brigham and Fred Rowe for their expertassistance in design and construction of the test equipment.Most importantly, we thank the Army-Tank Automotive Commandfor their financial and technical support. Special note ismade of the important contributions of Dr Walter Bryzik, andErnest Schwarz. They put the program together andcoordinated the Cummins and University of Michigan efforts.

3

4

TABLE OF CONTENTS

Section Page

1.0. INTRODUCTION .................................... 18

2.0. OBJECTIVE ............................... ....... 19

3.0. CONCLUSIONS AND OBSERVATIONS .................... 203.1. Instantaneous IMEPMethod ....................... 203.2. Fixed-Sleeve Method ............................. 213.3. Ring and Liner Beinch Test ....................... 21

4.0. RECOMMENDATIONS .................................. 224.1. Instantaneous IMEP Method ....................... 224.2. Fixed-Sleeve Method ......... .................... 224.3. Ring and Liner Bench Test ....................... 23

5.0. DISCUSSION ...................................... 235.1. Background ................................... 235.2. Design .......................................... 245.2.1. Instantaneous IMEP Method ..................... 245.2.2. Fixed-Sleeve Method ........................... 325.2.3. Ring and Liner Bench Test ..................... 325.3. Implementation .................................. 385.3.1. Implementation of the Instantaneous IMEP

Method ........................................ 38

5.3.1.1. Engine Installation ...................... 385.3.1.2. Grasshopper Linkage Design.................. 405.3.1.3. Cylinder Pressure-Transducer............. 495.3.1.4. Strain-Gauged Connecting Rod.......... 495.3.1.5. Data Acquisition ......................... 555.3.2. Implementation of Fixed-Sleeve Method......... 585.3.2.1. Implementation on the 4.1-Litre Engine ...... 585.3.2.2. Additional Details ...................... 615.3.2.3. Implementation on the Cummins 903 Engine .... 615.3.3. Ring and Liner Bench Test ..................... 705.4. Testing ......................................... 705.4.1. Testing With the Instantaneous IMEP Method .... 705.4.1.1. Procedures ....................... 705.4.1.2. Results on the Cummins 903 Engine ........... 725.4.2. Testing With the Fixed-Sleeve Method .......... 895.4.2.1. Procedures .................................. 895.4.2.2. Calibration ................................. 895.4.2.3. Fixed-Sleeve Results on 4.1-Litre Engine .... 905.4.2.4. Testing With the Cummins 903 Engine ......... 1055.4.3. Ring and Liner Bench Test ..................... 1125.4.3.1. Baseline Ring/Liner Results ................. 114

5

TABLE OF CONTENTS (Continued)

5.4.3.2. Titanium Carbide Ring and Chrome CarbideLiner Coatings .............................. 115

5.4.3.3. Plasma-Sprayed Chromium Oxide Ring and LinerCoatings ........... .... .................... 120

5.4.3.4. Comparison With Cameron Plint ............... 1205.4.3.5. Comparison With Engine Results .............. 1275.4.3.6. Interface Oil Quantity and Quality

Considerations .............................. 1275.4.3.7. Simulator Repeatability ..................... 1335.4.3.8. Photographic Results from Engine Tests ...... 133

LIST OF REFERENCES .................... ........ .......... 154

DISTRIBUTION LIST ................................... Dist-6

6

7

LIST OF ILLUSTRATIONS

Figure Title Page

5-1. Single-Cylinder Cummins 903 Engine Equippedfor Instantaneous IMEP Method .................. 25

5-2. Stain-Gauged Connecting Rod .................... 28

5-3. Grasshopper Linkage ............................ 29

5-4. Typical Cylinder Pressure, Connecting Rod,and Inertia Forces ............................. 30

5-5. Piston and Ring Assembly Friction Force,Velocity and Power Dissipation ................. 31

5-6. Fixed-Sleeve Design for the 4.1-LitreGasoline Engine ............. ... ............... 34

5-7. Cylinder Pressure (upper), Strain-GaugeVoltage (center), and Resulting FrictionForce (lower), WOT, 1000 RPM, Motoring ......... 35

5-8. Schematic Showing Essential Features ofBench-Test Simulator for Ring and Liner Wearand Friction ............................. ....... 37

5-9. Photograph of Layout Board for DesigningGrasshopper Linkage ......................... 41

5-10. Photograph of Completed Grasshopper Linkage .... 42

5-11. Lower-End Bracket, Side View ................... 43

5-12. Lower-End Bracket, End View .................... 44

5-13. Oil-Pan Modifications .......................... 45

5-14. Oil-Pan Spacer .................................. 46

5-15. Photographs of Lower-End Bracket (upper)and Oil-Pan Modifications (lower) .............. 47

5-16. Grasshopper to Connecting Rod AttachmentBracket ........................................ 48

5-17. Pressure-Transducer Sleeve ..................... 50

8

5-18. Sleeve and Head Assembly ....................... 51

5-19. Pressure-Transducer Location in Head ........... 52

5-20. Photographs of Pressure-Transducer ............. 53

5-21. Photograph of Machined Connecting Rod .......... 54

5-22. Data Acquisition System and Display ............ 56

5-23. Crankangle Encoder Installed on Front of Engine 57

5-24. Schematic of Data Acquisition System ........... 59

5-25. Block Diagram for Data Acquisition System ...... 60

5-26. Photograph of Fixed-Sleeve Assembly for 4.1-Litre Gasoline Engine .......................... 62

5-27. Components of the Fixed-Sleeve Method for theCummins 903 Engine ............................. 64

5-28. Original Strain-Gauge Bridge Noise, No BridgeExcitation ........................... 64

5-29. Same as Figure 5-28 Except with Excitation,10 lbm Weight Applied Intermittently ........... 65

5-30. Same as Figure 5-28, Except 2 lbm Weight Added. 65

5-31. Same as Figure 5-30, Except 10 kHz Filter ...... 66

5-32. Same as Figure 5-30, Except 30 kHz Filter ...... 66

5-33. Same as Figure 5-32, Except No Load Applied .... 67

5-34. Fourier Transform of Figure 5-33 ............... 67

5-35. Fourier Transform of Figure 5-33, HigherFrequencies .................................... 68

5-36. Original Signal, Except IBM Monitor Unplugged.. 68

5-37. Bridge Calibration, Volts Versus Weight Appliedto Strain-Gauged Collar ........................ 69

5-38. Brake Specific-Fuel Consumption DuringBreak-In: BSFC units, lbm/bhp-hr ............... 73

5-39. Motoring Torque During Break-In ................ 74

9

5-40. Fuel Injector Inlet Pressure During Break-In... 75

5-41. Friction Force with Low (102 F) and High(200 F) Oil and Water Temperatures, SAE 40Oil, 200 RPM, Normally Aspirated ............... 77

5-42. Friction Force Trace Including Error fromCrankshaft Magnetism - 1200 RPM, 200 Waterand Oil, Naturally Aspirated ................... 79

5-43. Strain-Gauge Signal Without BridgeExcitation Before and After Reduction ofMagnetic Interference, 1200 RPM ................ 80

5-44. Motoring Piston and Ring Friction, 1200RPM, 29.17 PSIA Manifold Pressure, 200 FWater and Oil, Friction MEP = 3.03 PSI ......... 82

5-45. Firing Piston and Ring Friction, 1200 RPM,29.17 PSIA Manifold Pressure, 200 F Waterand Oil, Friction MEP = 2.25 PSI ............... 84

5-46. Instantaneous Crank Speed at 1200 RPM,Firing, 30 In. Hg. Boost ....................... 85

5-47. Results Shown in Figure 5-45 BeforeCorrected for Instantaneous Speed Variation .... 86

5-48. Motoring Piston and Ring Friction, 1200RPM, 30.01 PSIA Manifold Pressure, UncooledOperation, Friction MEP = 2.97 PSI,Synthetic Lubricant ............................. 87

5-49. Firing Piston and Ring Friction, 1200 RPM,29.91 PSIA Manifold Pressure, UncooledOperation, Friction MEP = 1.55 PSI,Synthetic Lubricant .................... 88

5-50. Calibration Curve Showing Increasing andDecreasing Pressure ............................ 91

5-51. Strain-Gauge Signals Overlapped for FourConsecutive Traces, 1000 RPM, Motoring, WOT .... 92

5-52. Friction Force Versus Crank Angle Motoringand Firing, 1000 RPM, Part Load: Note MajorDifference After TDC Firing .................... 93

5-53. The Influence of Engine Speed Under FiringOperation, 1000 RPM, WOT: M.V.P. IndicatesMaximum Piston Velocity Crank Angle ............ 95

10

5-54. The Effect of Water and Oil TemperaturesUnder Motoring Operation, 1000 RPM, WOT:T.P. Indicates Transition Point ............... 95

5-55. Comparison Between Honed and Machined Liners,1000 RPM, Motoring, WOT ........................ 97

5-56. Friction Without Oil Ring, Various CompressionRings Removed, 500 RPM, Motoring, WOT .......... 98

5-57. Friction with Oil Ring, Various CompressionRings Removed, 500 RPM, Motoring, WOT .......... 99

5-58. Histogram of Friction Mean EffectivePressure for Various Ring Combinations,From Data of Figures 5-55 and 5-56 .............. 101

5-59. Comparison of Fixed-Sleeve (upper) andInstantaneous IMEP (lower) Methods,Motoring, 1000 RPM, SAE 50 Oil, 14.39 PSIAManifold Pressure, 104 F Water, 130 F Oil ...... 103

5-60. Comparison of Fixed-Sleeve (upper) andInstantaneous IMEP (lower) Methods, Firing,1000 RPM, SAE 30 Oil, 8.39 PSIA ManifoldPressure, 176 F Water, 220 F Oil ............... 104

5-61. Comparison of PR FMEP Results for the TwoMethods: (The Fixed-Sleeve Results areConsistently Higher) ........................... 106

5-62. Piston and Ring Friction by Fixed-SleeveMethod, Cummins 903 Engine, 100 RPM, SAE 40Oil, Motoring, 175 F Water, 195 F Oil,Friction MEP = 4.08 PSI ........................ 107

5-63. Noise Generated by Magnetic.Fields, NoBridge Excitation, 100 RPM ..................... 108

5-64. Data of Figure 5-62 Corrected for MagneticInterference ..................................... 109

5-65. Noise Generated by Magnetic Fields, No BridgeExcitation, 1000 RPM ............ . ...... .. 110

5-66. Piston and Ring Friction by Fixed-SleeveMethod, Cummins 903 Engine, 1000 RPM, SAE 40Oil, Motoring, 175 F Water, 195 F Oil,Friction MEP = 4.85 PSI, Corrected forMagnetic Noise..................-.. ... ...... .i i

5-67. Friction Coefficient Versus Crank Angle forBase Engine Materials, Normal Force 85.4 N,268 RPM, 35 C Test Temperature ................. 115

5-68. Proficorder Trace of Worn Baseline LinerTested Against Chromium Plated Ring,Mineral Oil: Scale Vert = 0.51 um/div,Horiz = 2.5 mm/div, Sample Interval of 7.4 um.. 115

5-69. Average Friction Coefficient During Test,Baseline Engine Materials, Lubricated .......... 116

5-70. Average Friction Coefficient Over Test,Titanium Carbide Ring on Chromium CarbideLiner, No Lubrication .......................... 117

5-71. Average Friction Coefficient Over Test,Titanium Carbide Ring on Chromium CarbideLiner, Synthetic Lubricant ..................... 118

5-72. Proficorder Roughness Traces of Plasma-SprayedChromium Carbide Liners, Before and After Two-Hour Test Against Titanium Carbide PlasmaSprayed Ring, Scale: vert = 0.25 um/div, Horiz= 0.1 mm/div, Sample Interval of 1.0 um ........ 119

5-73. Proficorder Traces Showing Wear ScarGeometry on Worn Chromium Carbide LinerSamples, Two-Hour Test, Scale: vert = 5.0um/div, horiz = 1.0 mm/div, Sample Intervalof 7.0 um ...................................... 121

5-74. Average Friction Coefficient Over Test,Chromium Oxide Ring on Chromium OxideLiner, No Lubricant ............................ 122

5-75. Average Friction Coefficient Over Test,Chromium Oxide Ring on Chromium OxideLiner, Synthetic Lubricant ..................... 123

5-76. Proficorder Roughness Traces of PlasmaSprayed Chromium Oxide Liners, Before andAfter Two-Hour Test Against Chromium OxidePlasma-Sprayed Ring, Scale: vert = 0.25um/div, Horiz = 0.1 mm/div, Sample Intervalof 1.0 um ...................................... 124

12

5-77. Proficorder Traces Showing Wear ScarGeometry on Worn Chromium Oxide LinerSamples, Two-Hour Test, Scale: vert = 5.0um/div, horiz = 1.0 mm/div, Sample Intervalof 7.0 um ...................................... 125

5-78. Ring and Liner Wear Coefficients fromCameron Plint Tests: Solid and Dashed LinesShow Variations of Ring and Liner Wear withTemperature for Conventional Materials (CrPlated Ring, Gray Iron Liner) with aCommercial CE/SF 15W40 \Mineral Oil BasedLubricant; Points are for Various CeramicCoatings; Captions Refer to Ring Material,Liner Material, and Lubricant Combination;Cr203, and Cr3C2 are Plasma-Sprayed;TiC is Plasma-Sprayed with CaF2, Ni, and Cr;SCAis Silica-Chromia-/Alumina Slury Coating .... 126

5-79. Profilometer Traces From Engine TestedPiston Rings .......................... ......... 129

5-80. Profilometer Traces From Engine Tested Liners.. 130

5-81. Right and Left (+) Side FrictionCoefficient Over Test Before Side-to-SideLubrication Balance. CE/SF 15W40 MineralOil, Chrome Plated Ring, Gray Iron Liner,149 C, 180 N Load, 500 RPM ..................... 131

5-82. Same as Figure 5-81, Except After Partial-Lubrication Balance ............................ 132

5-83. Same as Figure 5-81, Except After Full-Lubrication Balance ............................ 134

5-84. Same as Figure 5-83, Except SDL-1 SyntheticLubricant ...................................... 135

5-85. Cr203 Coated, Unworn ........................... 137

5-86. Cr2C3 Coated, Unworn ........................... 137

5-87. Cr203 Coated, 2 Kg Load Knoop Indenter ......... 139

5-88. Cr2C3 Coated, Worn, 1 Kg Load Knoop ............ 139

5-89. Cr203 Coated, 2 Kg Load Knoop Indenter ......... 140

5-90. Cr2C3 Coated, 2 Kg Load Applied ................ 140

13

5-91. Cr203 Coated, Many Axial and Radial Cracks ..... 141

5-92. Cr203 Coated, Many Axial and Radial Cracks ..... 141

5-93. Cr2C3 Coated, Light Areas are Worn ............. 143

5-94. Cr2C3 Coated, Top of Peaks Show Scratches ...... 143

5-95. Cr203 Coated, Worn, More Bumps Still Exist ..... 144

5-96. Cr203 Coated, Slightly Worn .................... 144

5-97. Photographs of Wear and Friction BenchSimulator ..................................... 145

5-98. Cr203-Dipped Coating, Unworn ................... 147

5-99. Cr203-Dipped Coating, Unworn, 0.1 Kg Load ...... 147

5-100. Cr2C3 Coating, Unworn ......................... 148

5-101. Cr2C3 Plasma-Sprayed Coating, Unworn .......... 148

5-102. Cr2C3 Plasma-Sprayed Coating, Unworn .......... 149

5-103. Cr203-Dipped Coating, Run Dry, 266 RPM ........ 149

5-104. Cr2C3 Plasma-Sprayed Coating, Run Dry ......... 150

5-105. Cr203-Dipped Coating, Synthetic Oil ........... 150

5-106. Cr203 Plasma-Sprayed Coating, Syntheticoil ............................................ 151

5-107. Cr2C3 Plasma-Sprayed Coating, Syntheticoil ........................................... 153

14

15

LIST OF TABLES

Table Title Page

5-1. Cummins 903 V-8 Engine Information .............. 26

5-2. Cadillac 4.1-Litre V-8 Engine Information ....... 33

5-3. Automated Features of Simulator ................. 71

5-4. Summary of Average Break-In Data ................ 76

5-5. Steps Taken to Minimize Magnetic Noise .......... 81

5-6. Simulator Material Couples and Data ............. 113

5-7. Uncooled NTC-250 Ring and Liner Coatings ........ 128

5-8. Side-to-Side Repeatability Comparison ........... 136

16

17

1.0. INTRODUCTION

This research at the University of Michigan was aimed atsupporting the advanced technology diesel program of theU.S. Army Tank Automotive Command, working with the CumminsEngine Company. Piston and ring assembly friction,scuffing, and wear are expected to be stumbling blocks tothe development of commercial, and even laboratory advancedtechnology engines, versions of which are expected to beminimally cooled. In such engines, liners may reachtemperatures in the range of 500-600 C. At thesetemperatures, liquid lubricants decompose rapidly. Wear,scuffing, excessive friction, and general destruction ofpiston and ring rubbing surfaces may be expected.Consequently, a measure of piston and ring assembly (PRA)friction, together with assessment of the wear and scuffingof rubbing surfaces, are needed engine-development tools.

The problem is that there are no well-established techniquesthat give the piston and ring assembly friction, as theengine runs under loaded conditions. Furthermore,conventional bench-type wear testers do not lend themselvesto total screening of advanced technology engine ring andliner components because of geometrical and environmentaldifferences between bench and engine tests. The chiefproblem is duplication of the top-ring reversal temperature,load and lubrication environment.

At the University of Michigan, a PRA friction measurementtechnique, termed the Instantaneous IMEP method, had beendeveloped prior to this program and reported by Uras inReferences 1-6. In that earlier work, it had produceduseful results in running gasoline engines. The initialthrust of the present program was to apply that technique toa minimally cooled, advanced-technology engine. The Cummins903 engine was selected. A single-cylinder version wasused.

Because of possible experimental limitations of theInstantaneous IMEP method and the absolute necessity ofdetermining reliable PRA friction data, it was decided topursue an alternate method as well. That method, aderivative of so-called moveable bore methods, is termed theFixed-Sleeve method. Its development had started in anearlier program at Michigan using a gasoline engine. Thatearlier work was reported by Ku in References 7 and 8. Inthe present program, the Fixed-Sleeve method was applied toa second Cummins 903 single-cylinder engine, and some ofthat work was reported in Reference 9.

To support their work on low-heat rejection engines, Cummins

18

had begun development of a bench-test simulator to permitrapid screening of materials with respect to piston and ringfriction and wear. Rapid screening is desirable, sinceengine test are relatively complicated, time consuming, andexpensive. That work was transferred to Michigan to supportthe present research program. Further development of thatfixture has occurred and is reported herein, as well inReferences 13 and 15.

A good deal of the material in this report was included inProgress Reports to the U.S. Army, References 10 - 12.These reports contain more detail on some topics.

In the sections that follow, the overall program discussionis organized into three segments. They are:

1. Instantaneous IMEP Method2. Fixed-Sleeve Method3. Ring and Liner Bench Test

Motoring and firing friction results are reported on the 903engine with the Instantaneous IMEP method, including resultswith ceramic-coated rings and liner, synthetic lubrication,and uncooled engine block. Motoring results are reportedwith the Fixed-Sleeve Method. A variety of results from thebench-test fixture are reported, including various ceramiccoatings and lubricants. A comparison with Cameron Plintwear data, as well as engine worn samples is made. Micro-scopic examinations of fresh and worn surfaces are included.The piston and ring assembly friction has been reported asthe quotient of the friction work per 720 crank angle degreeengine cycle divided by the cylinder displacement. This istermed the PR MEP and is reported in units of psi.

2.0. OBJECTIVE

The objective of this research was to quantify the pistonand ring friction of advanced technology, low-heat rejectionengines. This mission involved:

1. Development of measuring techniques.

2. Evaluation of advanced materials including ceramics.

3. Evaluation of lubricant effects includingsynthetics.

4. Development of experimental screening methods forfriction and wear of piston and ring-rubbingsurfaces.

19

3.0. CONCLUSIONS AND OBSERVATIONS

The measurement of instantaneous piston and ring friction ina turbocharged, firing engine is a difficult challenge.Only near the end of the program were successful frictionmeasurements made in a firing engine with the InstantaneousIMEP method and in a motored engine with the Fixed-Sleevemethod. On the other hand, the bench-test simulator provedrelatively easy to operate, and much data was obtained onceramic-coatings, synthetic lubrication and high-temperatureservice.

Below are conclusions and observations for the various

experimental methods within this program.

3.1. Instantaneous IMEP Method

The Instantaneous IMEP method appeared to give reasonablevalues for PRA friction in the 903 engine with intake airboost. The technique is promising for revealing significantchanges in friction, or the appearance of unusual frictionforces, such as those associated with an engine failure. Onthe other hand, it may be difficult to quantify smallfrictional changes because of the lack of sensitivity of themethod. A major advantage of the method is that nosignificant engine modifications are required. Thus thedata obtained correctly reflects the engine design. Noother technique has this advantage. Some specificobservations were:

1. Firing and motoring PRA friction were similar undermotoring and firing conditions. Friction was in theboundary or mixed regime near the dead centers andappeared relatively hydrodynamic midstrokes. With cooloil (high viscosity), the friction was higher, but morehydrodynamic. Warm engine PR MEP values ranged fromabout 1.5 to 4 psi.

2. The synthetic lubricant gave lower friction than theconventional SAE 40 oil in conventional, cooled-engineoperation.

3. Friction with an uncooled block and syntheticlubricant was similar in character to that of the baseengine with cooling and conventional motor oil, whenmotoring. Firing results looked similar, but the PR MEPvalues were unreliable.

20

3.2. Fixed-Sleeve Method

Some difficulties were experienced in adapting this methodto the Cummins 903 engine. As a result, limited data wereobtained. PR MEP was 2.34 psi at 1000 rpm, motoring underwarmed-up engine conditions. Results looked similar tothose obtained with the Instantaneous IMEP method.

Below are some observations and conclusions.

1. Repeatability and the ability to differentiate smallfriction differences appear excellent. On the otherhand, a special liner is required; thus enginemodifications are necessary which themselves may modifybase engine friction characteristics. The methodappears to be readily adapted to engines with liners,such as the Cummins 903 engine.

2. The Fixed-Sleeve method permits relatively rapidchange of piston, ring, and liner variables, andcalibration is relatively straight forward. Designvariables can be investigated relatively quickly.

3. This method is projected to be very useful forevaluating bore roundness, straightness, and surfacefinish effects. It will also be useful for evaluatingnew piston or ring designs.

4. The location of the strain-gauges, away from theheat and cylinder stresses, suggests that this methodmight be more easily used to measure friction inuncooled engines than the Instantaneous IMEP method.

3.3. Ring and Liner Bench Test

Below are some conclusions and observations on the benchtest simulator for piston and ring wear, and friction.

1. In unlubricated tests, the titanium carbide,chromium carbide pair was a poor combination, whereasthe chromium oxide pair was exceptional. The simulatortests correlated directionally with the Cameron Plintresults and were in qualitative agreement with engineresults.

2. Lubricated tests demonstrated lower wear andfriction, as expected. Careful control of lubricantfeed rate, targeting, and spray pattern are required forgood repeatability and minimization of side-to-sidevariations. Air atomization of the lubricant is veryhelpful.

21

3. Additional work is needed to establish correlationbetween bench-testers and running engines. Such workincludes development of standard test procedures,parametric evaluations, and more sophisticated surfaceanalysis of wear samples. A better knowledge of thequantity and quality of the lubricant at the wearinterface of running engines would be helpful inachieving a more correct laboratory simulation.

4. Overall, it is concluded that the simulator is apromising bench-test device for evaluating ring, liner,and lubricant combinations for advanced technologyengines. While it will never be possible to totallyeliminate engine testing, this efficient benchsimulation should make the job of screening newmaterials, new lubricants, and new designs moreefficient and cost effective.

4.0. RECOMMENDATIONS

4.1. Instantaneous IMEP Method

This method of measuring piston and ring assembly frictionworked reasonably well on the boosted Cummins 903 engine.Problems were encountered developing the instrumentation,especially solving the problem of magnetic noise. Finally,results looked reasonable. This method is projected to bevery useful for evaluating friction and friction-relatedproblems in uncooled engines.

4.2. Fixed-Sleeve Method

The progress of this method was slow, since the major effortwas on the Instantaneous IMEP method. By program end, afirst-generation design had been implemented on a motoringengine. Some problems with strain-gauge lead wiring weresuspected. The signal to noise ratio was poor. It wasconcluded that a new design was needed which would:

1. Eliminate the threaded liner hold-down in favor of aclamped design.

2. Use a thinner section on which the strain-gaugeswere applied.

3. Route the strain-gauge wires to produce less of anantenna effect.

4. Demagnetize all the moving parts in the crankcase.

22

4.3. Ring and Liner Bench Test

The ring and liner bench-test fixture worked quite well asdeveloped within this program. It should prove useful as ascreening tool for rings, liners, materials and lubricants.The number of engine tests required should therefore, bediminished considerably. A number of organizations haveexpressed interest in obtaining the simulator.

5.0. DISCUSSION

5.1. Background

The Instantaneous IMEP technique had been developed at theUniversity of Michigan prior to the start of the presentprogram. It had produced useful results on running gasolineengines. The initial thrust of the present program was toapply that technique to the minimally cooled, advancedtechnology diesel engine.

The first step was to apply the technique to the productionCummins 903, and a one-cylinder engine was set-up andinstrumented for this. Once a successful result wasobtained, the approach was applied to the same engine with"low-heat rejection components" and run under "low-heatrejection operation." That effort is described in par.5.2.1., 5.3.1, and 5.4.1.

Early in the program, it was found that the instrumentationsignals needed for the Instantaneous IMEP method were quitenoisy with the 903 engine. In addition, it was anticipatedthat the differencing of the large numbers derived from thisturbo-charged engine would make precise frictionquantification quite difficult. Prior to the start of thisprogram, the Fixed-Sleeve method had been initiated on anautomotive-type gasoline engine. That technique workedextremely well on the Cadillac 4.1-litre V-8 engine, whichhad a die cast aluminum block and iron liner. That work wasreported in references 7 and 8.

The success on the gasoline engine, coupled with concernsfor precision of the Instantaneous IMEP technique, led us toapply the Fixed-Sleeve method to a second Cummins 903single-cylinder engine. The design implementation isdescribed in this report, but the test results, obtainedprior to the end of this program, were disappointing. Nouseful results were obtained, although the approach stillappears attractive. A summary of the design and results ofthe earlier work on the gasoline engine, together withextension to the 903 engine are included in this report.

23

(See par. 5.2.2., 5.3.2, and 5.4.2.)

Parallel to the above-engine program, a laboratory benchtest friction and wear simulator was developed to assesssurface damage, wear, and friction between liner segmentsand reciprocating ring. This simulator allowed rapidscreening of candidate materials. The basic reciprocatorwas designed by Cummins. It was extensively modified andcomputer controlled at Michigan. The progress on thisportion of the program has been rapid, and considerable datahas been taken with this rig. The simulator has become auseful screening device for selecting promising, high-temperature engine materials and lubricants. This portionof the program is summarized in par. 5.2.3., 5.3.3. and5.4.3. The simulator is discussed in references 13 and 15.

5.2. Design

Discussed below is the design of the various experimentalmethods explored within this program. These are:

1. Instantaneous IMEP Method2. Fixed-Sleeve Method3. Ring and Liner Bench Test

5.2.1. Instantaneous IMEP Method. The Instantaneous IMEPmethod employs a force balance between the measured cylinderpressure force, the compressive force in the connecting rod,the inertia force and the friction force. Considering thecomponents of these forces acting along the cylinder axis,the sum of the other three equal the friction force of thepiston, rings and piston pin assembly. Figure 5-1 shows thesingle-cylinder version of the Cummins 903 V-8 enginemodified for the Instantaneous IMEP method. The large pipeemanating from the cylinder-head region is a vent for thecrankcase. This provides atmospheric pressure on thebackside of the piston. A crankangle encoder is mounted onthe front of the engine. Simultaneous measurements ofcylinder pressure and rod compressive force are made at eachcrank angle over the entire cycle. Table 5-1 gives relevantengine information about the production Cummins 903 engine.

In applying this method, inertia forces are commonlycalculated from the well-known theoretical equation,assuming a constant crank-speed. As discussed later, forthis one-cylinder engine, and where inertia forces are quitelarge, crank-speed variation must be taken into account inthe calculation of the inertia forces. In the determinationof the inertia forces, it is necessary to consider thereciprocating mass of the piston, rings, the pin, and thatportion of the connecting rod which reciprocates with the

24

Figure 5-1. Single-Cylinder Cummins 903 Engine Equipped for

Instantaneous IMEP Method

25

Table 5-1. Cummins V8-903 Engine Information

Bore 139 mm

Stroke 121 mm

Connecting Rod Length 208 mm

Displacement 903 cu in

Compression Ratio 15.5:1

Closed-Deck Iron Block

Wet-Cast Iron Liner (Removable)

26

piston. As reported in reference 1, a distributed model ofthe connecting rod inertia is necessary, since it was foundthat the error associated with the use of a lumped model wasexcessive.

Figure 5-2 shows the strain-gauged connecting rod used withthe Instantaneous IMEP method. Pairs of 90 degree rosettesare installed on either side of a machined section locatedat approximately one-third of the connecting rod web lengthnear the small end. Wires from the strain-gauges, powersupply, and thermocouple, to measure the bridge temperature,are attached to the connecting rod. From the connectingrod, the wires are routed exterior to the engine by the useof a grasshopper linkage shown in Figure 5-3. More detailof the implementation of the Instantaneous IMEP technique isgiven in Section 5.3.1.

For purposes of illustration of the method, Figure 5-4 showssignals for Instantaneous IMEP method measurements made on atwo-litre gasoline engine at 800 rpm motoring, partthrottle. The top trace is the cylinder pressure, thecenter trace is the compressive force in the co-necting rod,and the lower trace is the inertia force. Note that thestrain-gauge signal includes abrupt jumps at thi deadcenters, indicated by the circles in the figure Thesejumps arise from the change in static friction as the pistonreverses direction. Differencing these forces to obtainpiston ring assembly friction results in the top curve ofFigure 5-5. This curve shows piston-ring assembly frictionforce as a function of crank angle. The static frictionreversals at the dead centers are very apparent.

It is difficult to determine the factors that contribute tothe shape of the friction force curve. This is because thecurve reflects the skirt, the three piston rings, and thepiston pin friction. Each component is operating in asomewhat different lubrication regime. Near the deadcenters, the piston rings dominate the friction results.There, the rings operate in the mixed or boundary•ubrication regime. In midstroke, probably all componentsire operating in the hydrodynamic regime.

The frictional work associated with these forces resultsfrom the product of friction force times piston velocity.The velocity is shown in the center portion of the figure.The power is shown at the bottom of Figure 5-5. The areaunder the curve of power versus crankangle is the totalenergy loss due to friction. The frictional work, dividedby the displacement of the cylinder, has been termed thefriction mean effective pressure (fmep) of the piston andring assembly, (PR FMEP).

27

Figure 5-2. Stain-Gauged Connecting Rod

28

Figure 5-3. Grasshopper Linkage

29

I- Intake --4 Compression .-Expansion4-- Exhaust --I

2.000

.0,500 .05 .000-

"• 1.500"0 70

0.000[

2.0

06!-• 90

S 180

0.

270

I. .. ± . J L..... I ,1 .. , I0 7" 2w 37r 4w

Crank Angle (rod.)

Figure 5-4. Typical Cylinder Pressure, Connecting Rod, andInertia Forces

30

1-Intake ---- Compression "Expansion - Exhaust -160

" 80OU~

COC

~ 0z U160

4

Z' 222>- 0

C 00o• 2

0_

4

240

3: 200

12a..9160-

U0.

80 80

.o 40

0 I I I I

0 7r 27r 37r 4,

Crank Angle (rod.)

Figure 5-5. Piston and Ring Assembly Friction Force,Velocity and Power Dissipation

31

5.2.2. Fixed-Sleeve Method. The Fixed-Sleeve method is aderivative of a class of what is termed "movable bore"methods. It was implemented on a 4.1-litre gasoline engine(7), and considerable data were obtained. Table 5-2 givesrelevant engine information for that engine. A briefcomparison of the two methods, simultaneously implemented onthe 4.1-litre engine, may be found on page 64 under"Comparison of the Fixed-Sleeve and Instantaneous IMEPmethods on the 4.1-Litre engine."

Figure 5-6 shows a cross section of the Fixed-Sleeve design.The original 4.1-litre engine liner, which had an 88 mm boreis fitted with a smaller 80 mm sleeve which is attached tothe outer liner by a pair of lock nuts located at the bottomof the assembly. Piano wire in tension is used near the topof the assembly to pilot the inner sleeve so that it doesnot touch the outer liner. O-rings are used to sealcombustion gas, water and oil. Cooling water is permittedto circulate between the two sleeves. Strain-gauges aremounted on both inner and outer surfaces of the necked-downsection near the bottom of the outer liner.

The outer liner, which is essentially identical to theproduction liner, is supported in the engine in theconventional way between the top of that liner and thesupport lip. Thus the inner liner is positioned like astandpipe, the top of which is shortened slightly so that itcannot contact the cylinder-head. Acting on the innersleeve are the sum of the cylinder pressure force acting onthe rim and the friction force. An example of these forcesis shown in Figure 5-7. The upper figure shows cylinderpressure while motoring at 1000 rpm, wide open throttle.The center figure shows the strain-gauge signals, and thelower figure shows the friction force as a function ofcrankangle. This is found by differencing the two uppersignals.

5.2.3. Ring and Liner Bench Test. The ring and liner benchtest simulator provides a rapid means of screening candidatering, liner, and lubricant combinations under simulatedengine conditions. The rig simulates the most severe wearand friction conditions in the engine, namely directly aftertop dead-center, where piston speed is low, and loads andtemperatures are high. The simulator incorporates acomplete piston ring positioned in a disk- shaped holder.The ring ends are held together by pins and clamped to thenominal gap clearance. The holder with ring reciprocateswith a stroke of 2.54 cm, and portions of the ring areloaded by two liner segments placed 180 degrees apart. Eachstationary liner segment is about 3.8 cm long and 0.7 cmwide. It is cut from a finished liner. The ring holder is

32

Table 5-2. Cadillac V8-4.l-Litre Engine Information

Bore 88 mm

Stroke 84 mm

Connecting Rod Length 134 mm

Displacement 4100 cc

Compression Ratio 8.5:1

Closed-Deck Iron Block

Wet-Cast Iron Liner (Removable)

33

88.8"

---- 80.00

UPPER O-RING

WIRE CLAMP

SUPPORTING WIRE

0"CYLINDER LINER

SUPPORTINGSLEEVE

SUPPORT LIP

LOWER O-RING

STRAIN GAGES

COPPER WASHER

LOCK NUT

Figure 5-6. Fixed-Sleeve Design for the 4.1-Litre GasolineEngine

34

"u 1200- MOTORING AT 1000 RPM(WOT)

,., 900-a..

La 600-zZ-

300-C")

- 100

600-

z450-

i-

3000LL

i-,

400-z

200,0

, -------------I0U0

200-

400 INTAKE I COMPRESSION I EXPANSION I EXHAUST180 360 540 720

CRANK ANGLE (DEG)

Figure 5-7. Cylinder Pressure (upper), Strain-Gauge Voltage(center), and Resulting Friction Force (lower),

WOT, 1000 RPM, Motoring

35

powered by a 1.34 kw d.c. electric motor that is speedcontrolled by a SCR controller. Load is applied to thering, liner interface by means of an air cylinder and leverarm arrangement. Provision is made for lubricant spray atthe wear interface, and the entire unit may be heated to ashigh as 550 C, in order to simulate low-heat rejectionengine top-ring reversal temperatures.

Data obtained from the simulator include friction force as afunction of stroke and average friction coefficient as afunction of time. As is often done after such tests, linersamples may be analyzed for wear volume, surface finish, andother meaningful wear parameters. Simulator results areavailable for both right and left wear interfaces.

Figure 5-8 shows a schematic of the simulator. Bypressurizing the air cylinder, ring loads are applied whichcan simulate the high pressures experienced in highlyturbocharged engines under peak pressure conditions.Maximum reciprocating speed is limited to 700 rpm by inertiaforces. In effect the simulator attempts to duplicate themost severe ring and liner condition; namely low speed, highload at TDC firing. Most testing to date has been at speedsbelow 500 rpm. Very high speed operation, if desired wouldrequire addition of a reciprocating balance mechanism.

The friction force is transduced by strain-gauges mounted onboth right and left arm pivots. This not only permits anindependent friction measurement at both wear locations, butalso eliminates unwanted bearing friction, inertia force,and noise from vibration which were characteristic ofearlier designs.

As desired, lubrication is provided by means of water cooledstainless steel tubes through which a pressurized air andoil mixture pass. The oil is provided as a spray mistdirected to the rubbing surfaces. A peristaltic pump,capable of precise, variable feed rate, is used to controloil quantity. A double wall, insulated, stainless steeloven surrounds the liner samples and ring. Electric heatersare used to produce high temperatures. Heater control is bymeans of the computer (IBM XT compatible), which cycles theheaters. Thermocouples are employed at several pointswithin the oven to indicate temperature. An angle encoderwith 250 pulses per revolution is used to tell the computerwhen to sample the strain-gauge force.

36

VARIABLE SPEED a STROKE

MACHINE FRAME NEEDLE BEARING

AIR CYLINDER

THRUST BEARINGSTRAIN GAGES

SPSON ING

iLINER SECTION•

Figure 5-8. Schematic Showing Essential Features of BenchTest Simulator for Ring and Liner Wear andFriction

37

5.3. Implementation

Below is discussed the implementation of the variousexperimental methods conducted within this program. Theseare:

1. Instantaneous IMEP Method2. Fixed-Sleeve Method3. Ring and Liner Bench Test

5.3.1. Implementation of the Instantaneous IMEP Method.

5.3.1.1. Engine Installation. A single-cylinder Cummins 903engine was obtained from Cummins Engine Company for thisprogram. The engine was one of several made by Cummins forresearch purposes. It was constructed from a V-8 by cuttingthe block in such a manner as to leave two cylinders at theflywheel end of the engine. The crank, front end, cylinder-head, and fuel injection system were appropriately modified.Figure 1 showed the engine installed in Room 249 of the W.E. Lay Automotive Laboratory at the University of Michigan.

In this modified engine, only one cylinder was working. Theother did not compress, due to a large hole placed in thepiston crown. Additional weight was placed on the pistonpin to preserve engine balance. By this means, a fairlysmooth running single-cylinder engine is obtained.

The engine was installed and coupled to a 600 HP G.E.Electric Dynamometer. The dynamometer was improved with theaddition of a 500 lbm maximum capacity load cell and aDaytronics Model 3270 strain-gauge conditioner/indicator toreplace an existing "LINK" pneumatic system. One of thestrain-gauge conditioner filter selections was modified tohave a cut off frequency of about 0.2 Hz so that thedynamometer output torque was averaged over a longer periodof time. This reduced low-frequency torque fluctuationswhich had made it difficult to get steady torque readings.The dynamometer speed controls were tuned thoroughly by G.E.personnel.

A system to convert the existing test cell engine coolantsystem from a total loss tap water type system to a closed50/50 antifreeze system was designed and installed. Thissystem used a 2 HP coolant pump which was provided byCummins. A BCF 5-030-03-024-004, 24", 2-pass heat exchangerwas purchased and installed to separate the coolant solutionfrom the tap water coolant. An identical unit was installedfor the oil system. Pneumatic temperature controllers wereused for control of the coolant, oil, and inlet airtemperatures. These were available in the laboratory.

38

Control valve bodies, with the proper flow coefficients,were installed. The smallest capacity units (Cv=0.5) gavebest temperature control.

Oil-pressure supply to the engine was provided by a 2 HPpump. An adjustable pressure regulator was added in aby-pass configuration, to allow accurate control of oilpressure. It was found that the piping from the pump outletto the heat exchanger, and on to the engine, had to be aslarge as possible (1" was used); otherwise, the pressureloss though the system dropped oil pressure below anacceptable level at the highest flow rate with the highestoil temperature. A dry sump oil reservoir was needed toavoid dragging the grasshopper linkage through the oil.

An intercooler unit modified from a multicylinder engine wasused to preheat the pressurized intake air. This was donein order to simulate air temperatures encountered in theturbocharged/intercooled multicylinder engine (240oF). Heatwas provided by steam at a pressure of about 30 psig. Thissteam was provided by a 40 kW electric mini-boiler in theequipment room above the test cell. The steam outlettemperature was monitored by the pneumatic controllertemperature sensor in the intake pipe, and the controllerregulated the flow of steam through the control valve. Bothoil and coolant system heat exchangers were equipped withsteam lines and traps, so that heat could be added whenrunning motoring tests, and prior to starting.

Pressure in the intake manifold of up to 50 inches ofmercury gauge pressure was provided by two screw-type ai,compressors. One was 40 hp. This was the regular builc •'gsupply (110 psig). The second was an auxiliary unit of :0hp placed above the test cell and was used only when theengine was run with boosted intake pressure. Air flow tothe engine was measured by a critical flow orifice systemconsisting of five orifices in parallel. The air-floworifices were calibrated with a 10 cu. ft. air bell. Alarge Heise gauge was installed upstream to monitor pressurethere. A 35 psi relief valve was installed in the line tothe intake manifold for safety reasons. Two-inch compressedair lines ducted the air from the compressors. Two low-resistance rayon air filters were installed to filter theair.

A variable speed fuel pump system was provided by Cummins.The pump was equipped with a small needle valve bypassingthe pump output for fine pressure adjustment. Coarseadjustment was provided by controlling the speed. A CoxInstrument Model 402 weigh scale was used to measure flow.

39

Exhaust back pressure was controlled by a manually operatedvalve. An 80 in. Hg. well type manometer was installed tomonitor the back pressure.

5.3.1.2. Grasshopper Linkage Design. The purpose of thistwo-bar linkage is to fully support a wiring harnesscontaining a number of wires from their point of attachmenton the connecting rod to a stationary point on the engine.From there, the wires are connected to external circuitry.Below are design criteria:

1. Minimize swing angles al, a2, a3 for minimum wire

twist.

2. Minimize link lengths for better clearance.

3. Minimize acceleration of the joints to minimizestresses.

4. Optimize geometry that will allow motion through allangles without interference with block and minimizemodifications of the oil-pan.

The specific design was made with the assistance of a full-scale layout board, Figure 5-9 This was constructed ofcardboard upon a full-scale drawing of the block. Pins andguides permitted rotary and reciprocating motion. Linklengths and attachment points were selected to avoidinterferences and to minimize the swing angles. The topswing angle al, is usually the largest, with some occasionalwire breakage experienced with angles as large as 90degrees. Angles less than 70 degrees are much preferred.Computer analysis of these linkages showed that the maximumacceleration is at the pivot point a2, which is the "elbow."This occurs just before the piston reaches bottom center.Figure 5-10 shows the completed linkage.

Figures 5-11 and 5-12 show the bracket design for holdingthe bottom (fixed) end of the grasshopper linkage in themodified oil-pan. The modified oil-pan is shown in Figures5-13 and 5-14. A photograph of the bracket holding thefixed pivot of the grasshopper linkage is shown in Figure5-15, upper. Figure 5-15, lower, shows a photograph of themodified oil-pan. Figure 5-16 shows the bracket attachingthe grasshopper linkage to the big end of the connectingrod.

40

Ný

Figure 5-9. Photograph of Layout Board for Designing

Grasshopper Linkage

41

?17

Figure 5-10. Photograph of Completed Grasshopper Linkage

42

• � �0 7, 0U K a 0pPAI

L -L I L

""14P FORIid.2S 111AP Eý I

xI I. so4M0zs".0s 9I j

AWADS SOL NO

,J ,

0,6z " BRACK..•ET SSE" O .CioAtLULt( 6. Er- H) A D- ,' o,,u' kj I : ", ,,

Fi-gur'e 5-11. L~ower'-End Br'acket., Side View

43

_,E FT 5 E. .0_0" CU MM INS

5.00 0 BRACKET

.0.004

r1 r

0.80 "O.5""L

II0.710 0~.1 0 1 .754

Figure 5-12. Lower-End Bracket, End View

44

CU/ktH/M5 0ziL- PPW 5PIPCE5-

I"

CUMMI[NS PAIN MODIFICATIONS

ExT &PAW

-4-- o i.5

4.6

* I . . . .t• !

Figure 5-15. Photographs of Lower-End Bracket (upper) andOil-Pan Modifications (lower)

47

CONNECTING ROD. BRACKET

O. WI4o"I . ........ ... 4.00" ..... .. .. )..J o. €b40" Cl•

.. . . -0.30.N l, I I ', ! ,. ,i , f,.oo" .. .. .... r - --....... - •:"- - "+:• .

T _ I... .....,, - I. '° "0'

S.T, 0. . . . . 5

0. 4OO "

-1

0.1501

2.10 RAO I D

' 3i" 75 - - T"

Figure 5-16. Grasshopper to Connecting Rod AttachmentBracket

48

The final Grasshopper linkage design parameters were:

Top member length 6.90"Bottom member length 6.00"Swing angles al 65 deg

a2 57 dega3 56 deg

Upper pivot pin relative X = +3.188"to connecting rod big end Y = -1.375"centerline (rod vertical)

Bottom pivot pin relative X = +2.15"to crank center Y = -11.96"

This linkage was satisfactory in that it never broke, andthe wires it carried never broke.

5.3.1.3. Cylinder Pressure-Transducer. A key aspect of theInstantaneous IMEP method is the correct measurement of thecylinder pressure at each crank angle. Our experienceindicates that the Kistler 7061 transducer is the bestavailable. This is a water-cooled unit, requiring a 14 mminstallation hole. The transducer was installed in avirtually flush mounted manner in the Cummins 903 cylinderhead. The transducer was contained in a 303 stainless steelsleeve. This prevented engine oil and coolant fromcontacting the transducer and cable connections. A coppercompression washer provided a seal against cylinderpressure. The transducer installation, including sleeve,are shown in Figures 5-17 to 5-19. Figure 5-20 presentsphotographs showing the installation of the pressuretransducer into the cylinder-head.

A Ruska Model 2400 HL dead weight tester was used tocalibrate the pressure-transducer. Six calibrationpressures up to 300 psi were applied. The procedure wasrepeated several times. Finally, the curve of calibrationvoltage versus pressure was fit by the least squares method.Subsequently, the pressure-transducer was used to calibratethe strain-gauges applied to the connecting rod. This gavean internally consistent calibration between the cylinderpressure and strain-gauge signals. The Kistler Model 504E10charge amplifier was used with the Kistler transducer.

5.3.1.4. Strain-Gauged Connecting Rod. An integral part ofthe Instantaneous IMEP method is to measure the compressivestress in the connecting rod. This requires machining flatson the connecting rod. Figure 5-21 shows the machined rodprior to application of the gauges. The intent of themachining is to provide smooth surfaces which are symmetricabout the neutral axis of the rod. Micromeasurements strain

49

SII0

1.041L

I I g •t

I 4.508,

8.9 71 *[1 OI~ -RINiG GROOVE

8.175"og

x ,0

I I€ I :'

_._,______AT___________-___.NLo5•Sr

HEADTRANUCER SLEEVE0,"FOR CJJMMINS V-90OUNIVERSITY OF MICH IGA. DIMENSIONS

Figure 5-17. Pressure-Transducer sleeve

50

SLEEVE - HEAC A M5'MBLY .... /FOR Ciji\IN5 V-90- /-/- -. VALVE COVER

UNtVERSITT OF MICHIGAN/

6-13-8r,- FULL SCALE

I Iil I SLEEVE

SEAT FOR iO-RIN6'"

, .•,. - - -... . /,./ I

/ .. ".- I /,' - - .II

/. , /"~ I ""1.0.

1.750"- MILLED FLAT FORLOPPER /ASHER

TAPPED FOR 0.5 20

Figure 5-18. Sleeve and Head Assembly

51

-205

.~ .010

Figure 5-19. Pressure-Transducer Location in Head

52

Figure 5-20. Photographs of Pressure-Transducer

Installation

53

Figure 5-2 1. Photograph of Machined Connecting Rod

54

gauges Model WK-06-062TT-350 with resistance of 350 ohmswere used. These are pairs of gauges mounted at rightangles.

After the gauges were applied and the rod installed, theengine was motored for a few hours, in order to age thegauges. Then the strain-gauges were calibrated twice asfollows. A one-inch thick flat plate with two 14 mm tappedholes and one-inch spacer was bolted to the cylinder blockdeck to allow introduction of calibration pressure andmounting of the pressure-transducer. Steam heated coolantwas circulated to control temperature during calibration.The crankshaft was locked at two different crank anglesusing brackets on the flywheel, and a nitrogen bottle wasused to apply continuously increasing and decreasingpressure to the piston. Stain-gauge and pressure-transduceroutputs were recorded simultaneously, and then calibrationconstants were calculated from this data. These constantsare the relationship between applied force and voltageoutput, which together with bending and temperaturecorrections are described in References 1 and 2. AMicromeasurements Model 2310 strain-gauge amplifier wacaused.

5.3.1.5. Data Acquisition. The Instantaneous IMEP methodrequires simultaneous measurement of cylinder pressuretransducer and connecting rod strain-gauge voltages at onecrank angle degree intervals. A data acquisitic: system wasdeveloped based on the IBM PC-XT, which enabled thesemeasurements at engine speeds up to 3000 rpm, a computerdata acquisition rate of 18 kHz. Ultimately, the dataacquisition rate was 20 kHz. A 12-bit resolution system wasused. At this resolution the error caused by a variation ofone bit is 0.25 percent, well within the objective of 1percent accuracy. That is one part in 4096. Data for nineconsecutive engine cycles were stored in memory for transferto a floppy disk. At the same time, the friction force as afunction of crank angle, was calculated and periodicallydisplayed on the a high-resolution (720 point up by 450down) monochrome monitor display screen (Figure 5-22). Inthis manner, it was possible to determine if the data"looked right" as it was being acquired. A dot matrixprinter with graphics capability allowed print out of testconditions, as well as the monitor display.

Two Techmar Labmaster 2009 option TM-40-PGH analog todigital conversion boards were installed in the IBM XT forthe data acquisition. These were triggered by the opticalencoder mounted on the engine crankshaft (Figure 5-23). Insome engine tests, a commercial crank angle encoder, theLedex Model RC23-DM-360-5SE-1B, was substituted. Sample and

55

a-Ho

Figure 5-22. Data Acquisition System and Display

56

00

Figure 5-23. Crankangle Encoder Installed on Front ofEngine

57

hold circuits on the A/D boards remember and hold thevoltage until the next trigger signal, giving the computertime to acquire the data. The circuit of this dataacquisition system is in Figure 5-24. Every other top dead-center pulse is removed, so that the crank angle pulses canbe gated through, starting with TDC on the intake stroke. Ablock diagram of the system software is shown in Figure5-25. The main programs were written in Fortran with anassembly language subroutine linked to the data collectionprogram for fast data acquisition. This data acquisitionsystem was further used for measuring the time intervalbetween crank angles, for determining the speed variation ofthe crank, and in the data acquisition of the Fixed-Sleevemethod.

5.3.2. Implementation of the Fixed-Sleeve Method.

5.3.2.1. Implementation on the 4.1-Litre Engine. Thisgasoline engine is unique in that it is comprised of a diecast-aluminum block into which are inserted cast-ironliners. It presented a very convenient vehicle to implementthe Fixed-Sleeve method. In the Fixed-Sleeve method thecylinder-head is not modified, except that a flush mounted,water cooled, Kistler 7061 pressure-transducer was installedin a manner similar to that described in 5.3.1.3., exceptthat it was installed through the front of the cylinder-headin number 1 cylinder.

The cylinder block was modified in that a sleeve wasinserted within a modified steel replica of the removableproduction liner. Acting on this sleeve are both thefriction force of the piston and rings and the combustiongas-pressure force acting on the upper rim of the sleeve.That gas force, acting on about 6.5 square cm of rim area,is then subtracted from the total, leaving the frictionforce as a function of crank angle.

The combined gas and friction forces are measured by full-strain-gauge bridge circuits uniformly spaced on both insideand outside of the slender section (1.5 mm thick) of thetailpiece of the outer liner. The gauges and method ofapplication are similar to that described in 5.3.1.4. forthe Instantaneous IMEP method. In this design, a total of16 active and 16 dummy gauges were used. The arrangementprovided temperature and bending compensation and enoughsensitivity to resolve better than one-pound force. Theproduction cylinder-head and gasket seal against the outerliner as in the production engine. A smaller 80 mm piston,provided by Cadillac engineering, and specially built castiron liners, provided by SPX Corporation, were used with anotherwise production engine, except that a carburetor and

58

2 133 12 CMOS Chips.4 11 (Starter Circuit)5 '1l 04024 - 7 Stage Ripple Counter7 V a 4013 - Dual D-type Flip-Flop

4001 - Quad 2-input Nor Gates

Power Supplys1 - 5 (Starter Circuit &2 1 Photo Diode power)

Sola Solids 84-05-210(5 volt, 1 amp)

0710)+5st Pressure

5..,10 Daughter

60t Mother Board .- #1

S~StrainFtý!ý0710H1 Starting Bus Address

IMother Board .

470 o h J2 pin 3

Frm]~ lo .o SYSTEM

oPhoto Diode Position 0coder

_- SCHEMATIC

ENGINE FRICTION DIAGRAMDATA ACQUISITION SYSTEM

Figure 5-24. Schematic of Data Acquisition System

59

Fortran Program

"GETDATA.FOR"

-Input Test Conditions

Fortran subroutine"UTecmar,,

Fortran Subroutine -Plots pressure and straingage inputs on screen-Stores raw pressure

and

strain gage data on disk 1

Assembly language subroutino

"ADC.ASW'

-Operates A/D converter

and loads data to memory

Floppy________________

Disk

Storage

Fortran Program

"FRA2. FOR"-Take data from disk, calculates

and plots friction on screen.

-Stores friction, pistonvelocity and test conditionson disk

Figure 5-25. Block Diagram for Data Acquisition System

60

mechanical distributor were used for experimentalconvenience. All cylinders were fired for the tests.

5.3.2.2. Additional Details. Referring to Figure 5-6,other features of the design are: tensioned piano wire (0.5mm dia) near the maximum side thrust region, to pilot theinner sleeve and keep it from contacting the outer liner(which gives no axial restriction.); soft silicon O-ringsfor isolating the cooling water and combustion gases; andseveral holes between inner and outer sleeves, to admitcooling water. The top of the inner sleeve was shavedslightly to avoid any contact with the head. The twosleeves were held together by a pair of lock nuts at thebase of the assembly. With this design, both piston andliner can be changed without disturbing the instrumentation.Figure 5-26 shows a photograph of the Fixed-Sleeve assemblycomponents.

5.3.2.3. Implementation on the Cummins 903 Engine.

Basic Design and Instrumentation. A second single-cylinderCummins 903 engine was set up in Room 246, for theimplementation of the Fixed-Sleeve method. In all respectsthe instrumentation and procedures were identical to thosedescribed for the 4.1-litre engine. Each engine presents adifferent challenge insofar as designing and applying theFixed-Sleeve method. Figure 5-27 shows a photograph givingan exploded view of the design created for the Cummins 903.At the left is a collar which is rigidly attached to theunderside of the block. That collar has a necked-downsection with strain-gauges mounted inside and outside. Thecollar and liner bottom are both threaded. The liner is theninserted into the block and screwed into the collar, whichsupports it rigidly at the bottom. The top of the liner isspecially prepared as a separate piece, shown on the rightin the figure. This piece is clamped to the block by thecylinder-head in the same manner as the production engine.The liner and the top piece have a concentric step and areallowed to slip relative to each other. Sealing isaccomplished by two O-rings. Combustion gases are sealed bya metal O-ring. A rubber O-ring is located just below themetal ring.

Similar to the 4.1-litre engine, the collar was instrumentedwith 16 pairs of high-temperature strain-gauges. Eachbridge consisted of eight gauges, four of which were equallyspaced on the inside surface and the other four on theoutside. The bridges were wired so that the gaugesmeasuring the circumferential, or hoop stresses, providetemperature compensation. The gauges were wired to minimizethe effects of bending loads. The gauges and wires were

61

Figure 5-26. Photograph of Fixed-Sleeve Assembly for 4.1-

Litre Gasoline Engine

62

covered with a thin layer of epoxy to protect them fromdamage and temperature variations due to hot oil.

Calibration. The purpose of calibrating the fixture was todetermine the following information:

1. Signal to noise ratio2. Causes of noise3. Repeatability4. Linearity5. Gauge factor6. Magnitude of hysteresis caused by O-rings

Signal to Noise Ratio. Figure 5-28 shows the noise presentwith no excitation. Figures 5-29 and 5-30 show the sameconditions, except that the excitation is applied, and10-(Figure 5-29) and 2-lb weights (Figure 5-30) arealternatively placed on the collar. It can be seen inFigure 5-30 that with no filtering, the noise level is about1-2 lbf. Figures 5-31 and 5-32 show the same signal (2-lbweight) when 10 kHz and 30 kHz low-pass filters are used.

Causes of Noise. Figure 5-33 is the same as Figure 5-32except, no load is applied. Figure 5-34 shows a spectraldistribution plot of Figure 5-33 obtained with a fastFourier transform (FFT). This plot is used to detect any 60Hz component. It is seen that 60 Hz and its harmonicsrepresents a sizable portion of the noise. Figure 5-35 isan spectral distribution of the same signal, but with thesampling rate adjusted so that the FFT would include muchhigher frequencies. The figure shows the presence of a 1700Hz signal and its harmonics. After some investigation, itwas determined that the IBM monitor was the source. Figure5-36 shows the signal with the monitor unplugged.

Repeatability and Linearity. Repeatability is excellent,with no observable difference from run to run. Linearity wasassessed by putting successively heavier weights on thecollar. Figure 5-37 shows the result. The strain-gaugesare quite linear, and no correction is anticipated.

Hysteresis. The hysteresis of the O-rings at theliner/top-ring interface has been investigated atatmospheric pressure and room temperature only. Nohysteresis was observed, a very desirable result.Subsequently, it was found that the sleeve could not beinserted with the lower cavity seals in place. It was thendecided to cool the running engine with engine oil, andeliminate the seals. Thus water was eliminated and norubber seals used.

63

Figure 5-27. Components of the Fixed-Sleeve Method for theCummins 903 Engine

VOLTS

I *0.0 051.0

SECONDS

Figure 5-28. Original Strain-Gauge Bridge Noise, No BridgeExcitation

64

0.02

-0 02-

-0.04-

-0.06-

SECONDS

Figure 5-29. Same as Figure 5-28 Except with Excitation,10 lbm Weight Applied Intermittently

VOLTS ~ ''P

I ~I '-- ; oil

I I I0 20 40

SECONDS

Figure 5-30. Same as Figure 5-28, Except 2 ibm WeightAdded

65

SECONDS

Figure 5-31. same as Figure 5-30, Except 10 kHz Filter

VOLTS

-C. 0.

SECONDS

Figure 5-32. Same as Figure 5-30, Except 30 kHz Filter

66

0.021.

VOLTS

Sao 20•

--. 04'

- 3. O.S $

SECONDS

Figure 5-33. Same as Figure 5-32, Except No Load Applied

VOLg-6

I * I

HERTZ

Figure 5-34. Fourier Transform of Figure 5-33

67

VOLTS

I I I * *

HERTZ

Figure 5-35. Fourier Transform of Figure 5-33, HigherFrequencies

VOLTS

0.00

-0.0

S I * I * I ..0.0 0. 0.4 0.3e 0.3

SECON-DS

Figure 5-36. Original Signal, Except IBM Monitor Unplugged

68

1.4-

1.2-

1.0-

+.8- + ++

Volts.6-

.4-

.2-

0 1020 40 60 80 100 120 140 160 180Weight, Lbm

Figure 5-37. Bridge Calibration, Volts Versus WeightApplied to Strain-Gauged Collar

69

5.3.3. Ring and Liner Bench Test. The implementation ofthe ring and liner bench-test involved not only thecalculation of friction coefficient as a function of crankangle, as well as average coefficient for the cycle, butalso implementing various process control tasks. Frictioncoefficient is determined by dividing the friction forcetransduced by the strain-gauges, by the normal load providedby the air cylinder. Friction coefficient is calculated andstored every 1.4 degrees of crank rotation. The frictioncoefficient is also averaged over one revolution. Theaverage coefficient, as a function of crank angle, isdisplayed on the monitor screen for both right and leftsides and may be recorded on a floppy disk at predeterminedintervals. Three cycles are averaged prior to plotting andstoring. Upon completion of the test, ring and linerprofiles may be measured to determine wear volume, weightloss, and surface roughness, as part of an overallassessment program. In the present study, liner and ringroughness profiles were taken in the axial direction of theliner and liner weight loss determined. Ring weight loss isdifficult to determine, since only a small portion of thering is worn.

For process control, as well as data acquisition andanalysis, a computer system is employed. Speed, load andtest temperature, as a function of time, are input for theprocess control function. The simulator is capable ofunattended operation over the prescribed cycle. While theresults presented herein are for tests of 2 and 4 hoursduration, tests of up to 20 hours have been conducted.Automatic shut-down occurs if variables exceed presetlimits. Periodically, friction data is taken, displayed onthe screen, and stored on a flexible disk. Table 5.3 givesa list of the computer controlled features.

5.4. Testing

Below is discussed the testing of the various experimentalmethods conducted within this program. These are:

1. Instantaneous IMEP Method2. Fixed-Sleeve Method3. Ring and Liner Bench Test

5.4.1. Testing With the Instantaneous IMEP Method.

5.4.1.1. Procedures. In testing with the InstantaneousIMEP method both motoring and firing tests were run. Somemotoring tests were run without compression, by installing aone inch thick steel plate with a hole the size of thecylinder bore. Typically two data runs were taken at each

70

Table 5-3. Automated Features of Simulator

Measurement and Control Features

Strain Gauge Force - Friction force stored digitally as afunction of angular crank position

- Unit stops if friction coefficientexceeds specified value(Important for scuffing tests)

Motor Speed - Speed monitored by angular positionsensor and recorded

- Speed Controlled from 50 to 700 rpm- Independent measurement.foroverspeed protection

Loading Force - Air cylinder pressure monitoredand recorded

- Pressure regulator motorized- Load constant or ramped up to

about 650N

Oven Temperatures - Sample block temperaturemonitored for control, and recorded

- Heaters cycled on and off based oncontrol equations

- Oven air sensed independently for

over-temperature protection- Temperature up to 6500C

System Integrity - Automatic shut-down after 10seconds if no signals received fromcontroller

71

engine speed. Each run contained two sets of nineconsecutive cycles and each set was taken about 1 minuteapart. Prior to tests the engine was stabilize at the testcondition. Pressure and strain-gauge signals were recordedsimultaneously using the data acquisition system describedin 5.3.1.5. Results are presented in Figure 5-41 through5-48, and are the average of 18 cycles. The use of an 18cycle average arose from the characteristics of the dataacquisition system, and did not reflect additionalconsiderations. Eighteen cycles did provide very repeatablefriction results, and there appeared to be no reason toacquire additional data.

5.4.1.2. Results on the Cummins 903 Engine. Below is adiscussion of the test results and some problem that wereencountered in conventionally cooled engine tests and withSAE 40 lubricant. This is followed by results from a cooledtest with ceramic-coating on rings and liners, and SDL-1lubricant.

Baseline Engine Data. Data taken during the break-in of thesingle-cylinder Cummins 903 engine are shown in Figures5-38, 5-39 and 5-40, and summarized in Table 5-4 togetherwith the test conditions. Looking at Figures 5-38 and 5-39it is seen that the bsfc increased about 0.01 units (2.5percent), and the motoring torque increased about 4 lbf. ft.(13 percent) during the break-in. This indicated a problemresulting in friction increase. Subsequently, it was foundthat a large steel ring bolted to the top of the dummybalance piston had worked its way against the cylinder wall.This caused some light scuffing and friction increase. Thesteel ring was modified and reinstalled, which solved theproblem. A further problem is evident in Figure 5-40. Fuelpressure increased during break-in. Ultimately, this wastraced to swelling of the rubber fuel return line, which wasincased in a flexible stainless braided covering. The linewas replaced. After the initial problems were corrected, nonew problems arose. The engine installation was judgedsatisfactory, since the performance was similar to other 903single-cylinder engines built by Cummins previously.

Piston Assembly Friction Without Compression. Figure 5-41shows friction force, as a function of crankangle, under alow-speed engine condition of about 200 rpm withoutcompression. There is no pressure force across the pistonand, thus, the curves reflect the strain in the connectingrod which gives the friction forces almost directly, sincethe inertia forces are small at this speed. Results arepresented for a relatively cool oil and water temperature of102 F and also a relatively hot oil and water temperature ofapproximately 200 F. An SAE 40-grade oil was used.

72

CUMMINS DIESEL

.5 ', •s• d•--' -- .... "~m

.3

.2

.I

1 is 15 21 2S 30 5 413 45 53 55 6soENGINE HOURS

Figure 5-38. Brake Specific-Fuel Consumption DuringBreak-In: BSFC units, lbm/bhp-hr

73

CUMMINS Di-SEL

0

18 1i i 20 25 30 35 46 4S Sm SS soENGINE HOURS

Figure 5-39. Motoring Torque During Break-In

74

CUHMINS DIESEL

1243 -. __ .--L ______,_ _-_ _- -

I120 - -

*

ins --- ----- 'A

I~nB• e

-,J40

26.

I i s3 15 21 25 33 35 41 4S So SS soENGINE HOURS

Figure 5-40. Fuel Injector Inlet Pressure During Break-In

75

Table 5-4. Summary of Average Break-In Data

SCE V903 performance Data Baseline

Speed (rpm) 2600

Torque (ft-lbs) 135

Brake Hp 66.83

BSFC (lb/bhp-hr) 0.400

Fuel Rate (lb/hr) 26.85

A/F Ratio 30.50

In/Ex Mfld Press (in Hg) 50.5/50.5

Ex Mfld Press/In Mfld Press 1.00

Fuel Press (psi) 100-125

Oil Rifle Press (psi) 35

Oil Temp (-F) *240

Intake Mfld Temp (-F) 140

Ex Mfld Temp (-F) 810

Water Out of Cyl Head (-F) 185

Water Into Cyl Head (-F) 176

Blowby-in H20 3.0

Fuel Out of Cyl Head (-F) 135

Fuel Into Cyl Head (-F) 75

76

INTAKE COMP. POWER EXHAUST240

180

120

U-m

-J 60LUUJ

LL o

S60__ _ _

120

180

240 90 180 270 360 450 540 630 720

THETA (DEG)

Figure 5-41. Friction Force with Low (102 F) and High(200 F) Oil and Water Temperatures, SAE 40Oil, 200 RPM, Normally Aspirated

77

It is evident that the greater viscosity in the lowertemperature test produced a much more hydrodynamic characterto the resulting friction. With the higher water and oiltemperatures, the friction trace exhibits considerable mixedlubrication and may be viewed as essentially a square waveof amplitude approximately plus or minus 20 lb force. Thepiston and ring assembly friction mean effective pressurewas was 4.8 psi cold and 2.37 hot. The nearly two-to-onedifference in friction results mainly from lower frictionforces during the midstroke where the product of frictionforce times velocity is especially high.

Occurrence of Magnetic Interference. At the University ofMichigan, we have had considerable experience applying theInstantaneous IMEP method to various gasoline engines. Theapplication to the diesel engine was not expected to presenta large problem, at least under motoring conditions.Consequently, friction traces, such as that shown in Figure5-42, were entirely unexpected. Figure 5-42 shows frictionforce as a function of crankangle at 1200 rpm, motoring forthe 903 engine. In spite of our best efforts, the dataappeared to be nonsense. After making a number of changesin engine hardware and instrumentation, it was finallyrealized that the strange result was coming from magneticfields of the rotating crankshaft. The crank of this enginehad been magnetized either during manufacture or inspection.Strain-gauges and lead wires, including the wires on thegrasshopper linkage, constitute conductors moving in thatmagnetic field.

Figure 5-43 shows the strain-gauge bridge output signalbefore and after reduction of the magnetic field, with nobridge input. The voltages generated by the magnetic fieldwere dominating the signal. A number of steps were taken tominimize this magnetic field effect, which are summarized inTable 5-5. The extraneous signals were minimized primarilythrough a combination of degaussing of the crankshaft andmagnetic shielding. As a result, the extraneous voltagebecame a relatively minor portion of the signal, which couldbe measured at each engine speed without bridge excitation,and the resulting noise subtracted from the strain-gaugesignal. The relatively smooth curve shows the noise afterall reduction methods were implemented.

Results After Elimination of Magnetic Noise. Havingresolved the magnetic field problem, both motoring andfiring data were taken on the Cummins 903 engine. Figure5-44 shows motoring results at 1200 rpm with one atmosphereboost pressure. In part, the remaining noise on the traceresulted from the relatively high amplifier gain required toamplify the strain-gauge signals from the rather massive

78

INTAKE COM-P. POWER EXHAUST

240 -- m l180

120

La.

60

w

10 •

L)vvLL 60

120

24090 180 270 360 450 540 630 720

THETA MEG)

Figure 5-42. Friction Force Trace Including Error FromCrankshaft Magnetism - 1200 RPM, 200 F Waterand Oil, Naturally Aspirated

79

INTAKE COmP. POWER EXHAUST

240

180

120- 20 ______

L-mdw 60

w

cc 60w cc7

120

180

24090 180 20 360 450 540 60 720

THETA MEG)

Figure 5-43. Strain-Gauge Signal Without Bridge ExcitationBefore and After Reduction of MagneticInterference, 1200 RPM

80

Table 5-5. Steps Taken to Minimize Magnetic Noise

1. Install Mu Metal (high permability iron) shieldingover all connecting rod wiring and connector.

2. Change wiring on connecting rod to give pairing ofsense wires and excitation wires to remove loops.

3. Twist pairs of copper braid shielding on grasshopperwires.

4. Demagnetize crankshaft as much as possible withhand- held degausser.

5. Smooth sharp edges on crankshaft to remove fieldconcentrations near grasshopper/connecting rod.

6. Measure signal at each speed without strain-gaugeexcitation, so that any remaining noise can besubtracted from strain-gauge signal.

81

INTAKE COMP. POWER EXHAUST240

180

m: . 60 Ik20

90 1 8 27K6.40 50 3 2

U.CS 60

LL cc

120

180

24 0-0 __90 ISO 270 360 450 540 630 720

THETA (DEG)

Figure 5-44. Motoring Piston and Ring Friction, 1200 RPM,29.17 PSIA Manifold Pressure, 200 F Water andOil, Friction MEP = 3.03 PSI

82

connecting rod of this engine. Compared to gasoline engineexperience, the jumps at the dead centers are not aspredominate. All in all, the friction appears to berelatively hydrodynamic throughout the 720 degree stroke.Figure 5-45 shows results of firing at 1200 rpm with oneatmosphere pressure boost. Compared to motoring in Figure5-44, large differences may be noted around top dead-centercompression. This large change between motoring and firingsuggests that the technique will be useful for identifyinglubrication problems in a low-heat rejection engine, wherethe friction forces at top dead-center combustion mightbecome very large. This firing curve exhibits someproblems, which are artifacts of the measurements. Inparticular, on the power stroke, the friction force drops toa negative level, which is incorrect.

Accounting for Instantaneous Crank Speed. The inertiaforces of Figures 5-44 and 5-45 were calculated based onmeasured instantaneous crankshaft speed. Figure 5-46 showsa trace of instantaneous speed at 1200 rpm firing with 30inches of mercury boost pressure. The more than 50 rpmvariation in speed over the cycle of this one-cylinderengine affects the inertia force significantly. Thus,consideration of the effect of instantaneous crank-speed isnecessary in this case. The friction trace is distorted, asshown in Figure 5-47, if this is not done. Considering theinstantaneous crank-speed of the engine is essential toobtaining good friction results, if there is significantspeed variation such as might be expected in engines withfew cylinders and heavy reciprocating inertias.

Results With SDL-1 and Uncooled Operation. Figures 5-48 and5-49 show motoring and firing results with syntheticlubricant SDL-1 at 1200 rpm, intake manifold pressure of 30psia, about 40 psi bmep, and without cylinder block cooling.These curves may be compared to Figures 5-44 and 5-45, inwhich SAE 40-oil was used in the conventionally cooledengine. The tests with SDL-1 were run using chrome oxidering and liner coatings, whereas the data with SAE 40 usedproduction chrome-plated iron rings and Lubrited gray ironliner. In comparing these results with those of the SAE 40lubricant, it is noticed that the PR FMEP is about the same(2.97 versus 3.03 psi) motoring, but is lower for the SDL-1(1.55 versus 2.25 psi) firing. Note that near TDCcompression, the SDL-I gave higher friction in the mixed orboundary friction regime, which is characteristic of thelubrication there, even for motoring. It is not known whythe SDL-1 gave significantly lower firing friction withuncooled operation. The result may be real, or it be due tomeasurement inaccuracy. More study on tuis is needed.

83

INTAKE COMP. POWER EXHAUST240

180

120 r ty I

ww

120

240so 180 270 360 450 540 630 720

THETA (DEG)

Figure 5-45. Firing Piston and Ring Friction, 1200 RPM,29.17 PSIA Manifold Pressure, 200 F Water andOil, Friction MEP = 2.25 PSI

84

"1288

1278

S"1268

~ 1258 0 "

S 1248

S"1238

. 1228

1218

190

1188 188 360 540CRANK ANGLE (DEG)

Figure 5-46. Instantaneous Crank Speed at 1200 RPM, Firing,30 In. Hg. Boost

85

-NTKE COMP. POWER EXHAUST240 -4

ISO

t80

120 W

.-U, I./,:,

IJ W

,, > 60_ _ LI

U.L 0 -PA_ _ _

LL i

90 10 270 360 450 540 630 720

THETA (DEG)

Figure 5-47. Results shown in Figure 5-45 Before Correctedfor Instantaneous Speed Variation

86

Intake Compression Expansion Exhaust

240

180

M 120-J

LýL 60

0cm

0c 60'(L

120

180

240 S0 it 27T 37T 47t

Crank Rngle (rad)

Figure 5-i Motoring Piston and Ring Friction, 1200 RPM,30.01 PSIA Manifold Pressure, UncooledOperation, Friction MEP = 2.97 PSI, SyntheticLubricant

87

Intake Compression Expansion Exhaust

240

180

00 120-J

UL0U. 60

L A A

0 600

-4

120

180

2400 T2i 3n 47r

Crank Angle (rad)

Figure 5-49. Firing Piston and Ring Friction, 1200 RPM,29.91 PSIA Manifold Pressure, UncooledOperation, Friction MEP = 1.55 PSI, SyntheticLubricant

88

General Comments on the Instantaneous IMEP Method. TheInstantaneous IMEP method, when properly implemented, hasthe potential of determining piston and ring assemblyfriction with good accuracy over a large portion of the fullcycle. It is estimated that about 90 percent of the cycleis reported with good accuracy. At the very highestpressures and combustion temperatures the uncertaintiesbecome greater. Thus about 10 percent of the crank anglesof the cycle, around top dead-center combustion, haverelatively high uncertainty with respect to friction forces.

The method is projected to be able to assess changes infriction coming about from engine or lubrication failures.The foregoing statement is very significant, since often onewishes to anticipate a problem as it develops. Therefore,even if the absolute values with the Instantaneous IMEPmethod are less than perfect, the indication of a relativechange is useful. A major plus for the Instantaneous IMEPmethod is that essentially no engine structuralmodifications are required. Our overall experience withthis method, in both gasoline and diesel engines, shows thatit has very good repeatability. Even with uncooled engineoperation, there was no indication that the technique wouldnot work.

Unfortunately, this program t