the tribological evaluation of compressor contacts ...€¦ · the tribological evaluation of...
TRANSCRIPT
The Tribological Evaluation of Compressor Contacts Lubricated
by Oil-Refrigerant Mixtures
ACRC TR-19
For additional information:
Air Conditioning and Refrigeration Center University of Illinois Mechanical & Industrial Engineering Dept. 1206 West Green Street Urbana, IL 61801
(217) 333-3115
B. Davis, C. Cusano
May 1992
Prepared as part of ACRC Project 04 Compressor--Lubrication, Friction, and Wear
C. Cusano, Principal Investigator
The Air Conditioning and Refrigeration Center was founded in 1988 with a grant from the estate of Richard W. Kritzer, the founder of Peerless of America Inc. A State of Illinois Technology Challenge Grant helped build the laboratory facilities. The ACRC receives continuing support from the Richard W. Kritzer Endowment and the National Science Foundation. Thefollowing organizations have also become sponsors of the Center.
Acustar Division of Chrysler Allied-Signal, Inc. Amana Refrigeration, Inc. Bergstrom Manufacturing Co. Caterpillar, Inc. E. I. du Pont de Nemours & Co. Electric Power Research Institute Ford Motor Company , General Electric Company Harrison Division of GM ICI Americas, Inc. Johnson Controls, Inc. Modine Manufacturing Co. Peerless of America, Inc. Environmental Protection Agency U. S. Anny CERL Whirlpool Corporation
For additional iriformation:
Air Conditioning & Refrigeration Center Mechanical & Industrial Engineering Dept. University of Illinois 1206 West Green Street Urbana IL 61801
2173333115
ABSTRACT
The tribological characteristics of the most common contact
geometries found in compressors of air conditioning and
refrigeration systems have been experimentally investigated by
means of a unique high pressure tribometer (HPT). The HPT has
been used to experimentally simulate the friction and wear
behavior of various metal contact pairs lubricated by oil
refrigerant mixtures in environments found in compressors. The
refrigerants used in this program are CFC-12 to obtain baseline
data and its prime replacement candidate, HFC-134a. The CFC-12
has been tested with mineral oils and synthetic alkylbenzenes
while the HFC-134a has been tested with monoether polyalkylene
glycol (PAG's) and pentaerythritol polyolester oils. Since the
amount of refrigerant dissolved in the oil is a function of both
pressure and temperature, and the friction and wear of a given
contact can be significantly affected by the concentration of
refrigerant in the oil, the friction and wear data obtained from
this test program should be a good indicator of what can be
expected in compressors.
iii
TABLE OF CONTENTS
CHAPTER PAGE 1. INTRODUCTION........................................... 1
1.1 Overview......................................... 1 1.2 Oil-Refrigerant Mixtures ......................... 2 1.3 Tribological Background .......................... 3
1. 3.1 Lubricant Requirements..................... 4 1. 3.2 Specimen Testing................... . . . . . . . . 6 1.3.3 Pressurized Friction and Wear Machines ..... 7
1.4 Scope of Research ................................ 8
2. HIGH 2.1 2.2
2.3
PRESSURE TRIBOMETER (HPT) SYSTEM ................. . Overall Facility ................................ . High Pressure Tribometer ........................ . 2.2.1 Pressure Chamber .......................... . 2.2.2 Temperature Systems ....................... . 2.2.3 Tribometer Motion ......................... . 2.2.4 Tribometer Controls ....................... . External Equipment .............................. . 2.3.1 Purging Facility .......................... . 2.3.2 Charging Facility ......................... . 2.3.3 Sampling Facility ......................... . 2.3.4 Data Acquisition .......................... .
10 10 11 11 19 21 23 25 26 28 29 30
3. EXPERIMENTAL METHOD. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 32 3. 1 Overview......................................... 32 3.2 Compressor Contacts .............................. 32
3.2.1 Counterformal Contact - Rolling Piston Compressor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 33
3.2.2 Area Contact - Swash Plate Compressor...... 36 3.2.3 Conformal Contact - Reciprocating Piston
Compressor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 38 3.3 Specimen Holders ................................. 41 3.4 Lubricants....................................... 45.
4. EXPERIMENTAL PROCEDURES...................... . . . . . . . . .. 51 4.1 Overview......................................... 51 4.2 Z-Axis Motion Control ............................ 51
4.2.1 Travel Mode ................................ 55 4.2.2 F·orce Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 56
4.3 a-Axis Motion Control ............................ 57 4.4 Temperature Control .............................. 59 4.5 Force & Torque Measurement ....................... 61 4.6 Installation of Specimens ........................ 66 4.7 Purging Procedure ................................ 67 4.8 Charging Procedure ............................... 71 4.9 Configuring the HPT Controls and Conducting
the Test ......................................... 75 4.10 Sampling ......................................... 79 4.11 Refrigerant Reclamation .......................... 81
5. RESULTS & DISCUSSION................................... 83 5.1 Measurement of Wear .............................. 83 5.2 Surface Analysis ................................. 86 5.3 Counterformal Contact Results .................... 87
5.4 Area Contact Results .............................. 90 5.5 Conformal Contact Results ........................ 92
6. CONCLUSIONS............................................ 95 6. 1 Research Summary ................................. 95 6.2 Recommendations for Future Research .............. 96
REFERENCES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 98
APPENDIX A - Equivalent Cylinders ......................... 100
APPENDIX B - Raw Data ...................................... 103
vi
1.1 Overview
CHAPTER 1
INTRODUCTION
Refrigeration and air conditioning systems require
lubrication to reduce friction and wear in the compressor, to act
as a coolant to remove heat from the compressor contacts, and to
help to seal the compressed gas between the high pressure and
suction ports. From a thermodynamic standpoint, the refrigerant
and lubricant should be completely separate to maximize
efficiency. Such an approach can be used in larger air
conditioning and refrigeration systems by means of oil separators.
For smaller systems, however, due to costs, the refrigerant and
oil should be miscible in order to aid in returning oil which
inevitably migrates from the compressor.
The decrease in production and use of dichlorofluoromethane
(CFC-12), as required by the Montreal Protocol, has forced the
refrigeration and air conditioning industries to develop
replacement refrigerants. The prime replacement for CFC-12 is
(1,1,1,2) tetrafluoroethane (HFC-134a). Although HFC-134a is
thermodynamically quite similar to CFC-12, it poses new problems
for compressor lubrication. The new refrigerant is not miscible
with the lubricants now used with CFC-12. Currently, lubricant
manufacturers are developing synthetic lubricants that are
compatible with HFC-134a and provide adequate lubrication for
critical contacts in compressors.
1
1.2 Oil-Refrigerant Mixtures
This section describes the basic relationship between
refrigerants and oils in compressor environments. Since a
lubricant in a compressor can be subjected to refrigerant vapor at
relatively high pressures and temperatures, and assuming that the
oil and refrigerant are at least partly miscible, the properties
of the mixture are very dependent on the environmental conditions.
Of prime interest to the lubrication of compressor contacts is the
amount of refrigerant in the oil and the tribological properties
of the resulting mixture. Much of these mixture data are taken
from previous research conducted to determine thermodynamic
properties of oil-refrigerant mixtures.
Data on the miscibility of oil-refrigerant mixtures as a
function of pressure and temperature have been investigated for
many oil-refrigerant combinations. Little [1], Parmelle [2],
Spauschus [3], Spauschus and Speaker [4], Pate et al [5,6,7], and
Grebner [8] have researched the effects of temperature and
pressure on the miscibility of specific oil-refrigerant mixtures.
In general, as long as the oil and refrigerant are at least partly
miscible, the amount of refrigerant in the oil depends on the
temperature and pressure. Figure 1.1 shows the saturation
properties of CFC-12 with a paraffinic mineral oil. Note that by
increasing refrigerant vapor pressure or decreasing the
temperature, the amount of refrigerant saturated into the oil
increases. The tribological properties of these resulting
mixtures are often less well defined.
Viscosities of oil-refrigerant mixtures are of prime concern
for compressor lubrication and the efficiency of the overall
2
system. Data for the viscosities of currently used oil-
refrigerant mixtures are well known in the literature
[1,2,4,5,6,7]. These evaluations have centered on CFC-12 with
various mineral oils and some synthetic oils. The data for
viscosities involving replacement refrigerants with the
appropriate lubricants are less complete. Some limited viscosity
data involving HFC-134a with PAGs has been obtained by Thomas [9]
but most other research with HFC-134a is currently in progress.
2.0,-----------------------------~~--~------~
1.5
-40
Figure 1.1
Oil Concentration
By Mass
-20
a 0%
• 20%
• 40% o 60%
• 80% 1:1 95%
o 20 40 60 80 100 120
Saturation Curves for CFC-12-Paraffinic Mineral Oil Mixtures at Various Concentrations
1.3 Tribological Background
This section provides general lubrication requirements for
compressor contacts, as well as some available tribological data,
in the open literature, for oil-refrigerant mixtures. These data
3
are a result of an extensive literature search involving lubricity
of oil refrigerant mixtures for compressor applications. Most of
what is included is a result of either actual compressor tests, or
specimen modeling tests. The last portion of this section will
outline the type of modeling necessary to more accurately simulate
conditions in compressors.
1.3.1 Lubricant Requirements
Lubricant requirements for air conditioning and refrigeration
compressors are discussed in chapter 8 of the ASHRAE Handbook of
Refrigeration [10]. Properties such as lubricity, chemical
stability and solubility between the refrigerant and lubricant all
contribute to the overall effectiveness of an oil-refrigerant
mixture. These requirements must be met for the lubricant to be
useful in compressor lubrication applications.
The main purpose of the lubricant in a compressor is to
decrease friction and wear and, therefore, prolong the life of
critical components. Ideally, the lubricant should completely
separate the contacting surfaces. These conditions can be
characterized by two lubrication regimes: hydrodynamic
lubrication, which occurs between converging surfaces, and
elastohydrodynamic lubrication, which occurs in counterformal
contacts which entail material deformation. Ideal conditions are
not always easily obtained, therefore, the conditions in the
compressor often are not capable of generating full fluid films
between the critical contact pairs.
Under start-up, shut-down, transient, rough surfaces, or
elevated temperature conditions, boundary or mixed lubrication can
4
be dominant. In boundary lubrication, there is insufficient oil
film thickness to separate the surfaces. This allows some
asperity interaction between the contacts which results in wear of
the surfaces. However, this wear can be minimized through the
formation of surface films. These films tend to reduce adhesion,
abrasion, surface fatigue, and corrosion, which are the four
prominent wear modes under boundary conditions. The lubricants
designed for use in compressor applications often require some
additives to aid in the formation of films to reduce wear under
boundary lubrication conditions.
Other than lubricity, the properties of the oil should also
encompass several additional properties. The stability of the oil
in the presence of refrigerants and compressor materials is an
extremely important property of the lubricant. Among these
stability requirements are: oxidative stability, primarily for
storage considerations, thermal stability, to avoid lubricant
polymerization/depolymerization at elevated temperatures, and
chemical stability, to resist reaction with refrigerant and
materials in the compressor. In addition, the oil should be
miscible with the refrigerant. The refrigerant vapor transfers
some of the lubricant from the compressor to the evaporator and
condenser. This oil must have sufficient low temperature
miscibility to allow the refrigerant to carry it back to the
compressor within a reasonable amount of time. If the refrigerant
fails to return the oil, the compressor can rapidly fail under
starved lubrication conditions. In addition, the oil should
remain miscible with the refrigerant to permit good heat transfer
while it is migrated to the evaporator.
5
1.3.2 Specimen Testing
If the oil and refrigerant are miscible, subjecting the oil
to the refrigerant under pressure tends to saturate the
refrigerant into the oil. The lubricative properties of the
resulting oil-refrigerant mixtur.e can be greatly affected by its
composition. Friction and wear tests must be capable of
accurately modeling these mixtures. The initial approach for
modeling the compressor environment has historically been to use
standard test equipment, Falex tests or pin on disk tests, and
bubble refrigerant through the. lubricant to achieve a mixture.
Huttenbacher [11] used just such a setup to determine .oil-CFC-12
wear characteristics using a Falex machine. Although this method
seems to have gained popularity with equipment manufacturers and
oil suppliers to screen prospective lubricants, the validity of
these tests in modeling actual compressor conditions has not been
established.
The limitations of this type of testing are mainly with its
inability to accurately model the actual environmental conditions
existing in compressors. Although by bubbling refrigerant through
the oil some refrigerant does tend to saturate into the oil, the
environment found in compressors is not simulated. In addition,
since the test occurs at atmospheric pressure, there is the
possibility of other gases saturating into the oil. Among these
is oxygen, which would show adverse effects on wear. The other
major drawback from these types of tests is the inability of the
apparatus to accurately model the actual contact geometries and
operating conditions of tribo-contacts in compressors.
6
The use of actual compressors in accelerated testing programs
is an alternative to specimen testing. Sundaresan [12,13] has
completed many long term tests in which various PAGs were used
with HFC-134a in a reciprocating piston compressor. The results
of these data are extremely useful in lubricant screening for that
one compressor, but results for other types of compressors cannot
be accurately inferred. Another drawback with this type of
testing is that it requires long term tests and is therefore not
particularly cost effective. For these reasons, the trend in the
tribological evaluation for tribo-contacts in compressors has been
towards the use of pressurized friction and wear machines.
1.3.3 Pressurized Friction and Wear Machines
To better simulate the environments found in compressors, the
use of pressurized friction and wear machines is gaining
popularity. Sanvordenker [14] utilized a modified Falex machine
to test CFC-12-mineral oil mixtures under pressure. Although this
work did not model the wide range of conditions found in actual
compressors, it was capable of reasonably modeling the
refrigerant-oil mixture at low pressures. More recent innovations
and advances in pressurized testing equipment for tribological
evaluation have lead to the development of high pressure friction
and wear machines. Such a machine has been used by Komatsuzaki et
al [15], Komatsuzaki & Homma [16], and Komatsuzaki et al [17] for
tribological evaluation of CFC-12-mineral oil, CFC-12-
alkylbenzene, and HFC-134a-PAG combinations, respectively. The
apparatus used in their investigation was a modified four ball
tester that had been equipped with a pressure chamber surrounding
7
the contact. However, the pressure capability of their machine is
not capable of modeling the pressures found in some compressors.
Since the pressure directly affects the amount of refrigerant
saturated into the oil, it is critical that the pressure existing
in compressors be duplicated in any testing program. In order to
accurately model environments in compressors, any testing
apparatus should have the following capabilities: the tests
should occur within a pressure chamber to allow for high pressure
refrigerant vapor testing, the apparatus should be capable of
modeling a wide variety of contact geometries for a wide range of
operating conditions, and the apparatus should have full
temperature control of the contact as well as the chamber itself.
Such a high pressure tribometer (HPT) was developed over the past
two years, as part of the Air Conditioning and Refrigeration
Center (ACRC) at the University of Illinois. This HPT is capable
of simulating environments found in almost all compressors.
1.4 Scope of Research
This project has been part of the ACRC, an industry
university cooperative, to decrease development time of ozone safe
air-conditioning and refrigeration compressors. As a result, the
main thrust of this research has been governed by specific
compressors and lubricants of interest to the sponsors. Other
projects within the ACRC have provided data on oil-refrigerant
miscibility and saturation pressure curves that have been used in
this test program. It is felt that an effective means of
simulating refrigeration and mobile air conditioning compressor
contacts and conditions have been developed.
8
The accuracy of the results of this research are contingent
upon the extent to which the actual conditions of the compressor
are modeled. With the assistance of the manufacturers of
compressors, critical contacts are identified, as are their
appropriate operating conditions. Included among these conditions
are refrigerant vapor pressure, contact temperature, load, speed,
materials, geometry, surface finish, and type of lubricant. Since
the lubricant is one of the main factors affecting wear, a concise
discussion of the lubricants tested is warranted. The
capabilities and design of the HPT is examined as are the test
procedures used in this research program. Next, the friction and
wear results obtained from the HPT for baseline CFC-12 tests as
well as initial tests with HFC-134a, are presented. Finally,
recommendations for future investigations are included.
9
CHAPTER 2
HIGH PRESSURE TRIBOMETER (HPT) SYSTEM
2.1 Overall Facility
The design .of the facility for the tribological evaluation of
critical contacts in compressors, in pressurized refrigerant
environments, centers on the development of a tribometer enclosed
in a pressure chamber. Because of the relative complexity of the
HPT, it was decided not to build it in-house. The remainder of
the system, though, was developed and constructed, at the
University of Illinois, using commercially available materials and
components.
Figure 2.1 Photograph of High Pressure Tribometer System
10
The completed HPT system, shown in Figure 2.1, is located in
the Tribology Laboratory in the Mechanical Engineering Building on
the Urbana-Champaign campus of the University of Illinois. The
design of the HPT system began in the Spring of 1990 and consists
of five sections: the HPT, a purging facility, a charging
facility, a sampling facility, and a data acquisition facility. A
thorough description of the design and operation of the complete
high pressure tribometer system follows.
2.2 High Pressure Tribometer
The tribometer was designed and manufactured by Advanced
Mechanical Technology Inc. (AMTI) of Newton Massachusetts.
Central to its design is a special pressure/vacuum housing, which
surrounds the test, capable of testing any inflammable non
corrosive gas. Multiple thermal control loops are included to
permit testing of any temperature typically found in compressors.
Two separate servo motors provide motion and loading capabilities.
The rotational, 9-axis, motor is capable of unidirectional
rotation and oscillatory motion. The load, Z-axis, servo motor
provides either static or oscillatory loads for the contact. A
complex transducer measures applied load, frictional forces, and
moments during a friction and wear test. The feedback from this
transducer, as well as other sensors, provide the HPT with an
excellent control system.
2.2.1 Pressure Chamber
In order to allow for a pressurized refrigerant environment,
the test must occur within the confines of a pressure chamber.
11
Figure 2.2 Cross-Sectional Assembly Drawing of HPT Chamber
12
The chamber of the HPT was designed using the ASME code and is
rated for a 1.725 MFa operating pressure. The chamber consists of
two separate units that come together to form the pressure
housing. The upper half of the chamber is stationary, while the
lower half can be raised or lowered by the Z-axis servo motor.
When the lower half telescopes into the upper half, it engages a
seal and subsequently can be pressurized with refrigerant.
The upper half of the chamber is an assembly of fastened
pieces sealed together by o-rings. Figure 2.2 shows a cross
section view of the entire pressure chamber. The telescopic seal
is held securely in the upper half by an aluminum ring bolted to
the upper housing. There is also a pair of seals used to seal the
joint where the shaft of the 9-axis motor enters the chamber.
The pressure chamber is sealed at the joint where the two
halves meet by a custom made six-inch telescopic seal. The seal
is kept lubricated by Braycote 803RP, a perfluoronated polyester
grease. The telescopic surfaces must be kept clean in order to
prevent abrasion of the seal and the surfaces. The 9-axis motor
shaft is sealed by a custom high pressure seal, designed for dry
sealing in refrigerant environments for pressures of up to
1.725 MFa and speeds of up to 2000 rpm. The design of this seal
is such that when a positive pressure is generated in the chamber,
the lip of the seal is held securely, by the pressure, against the
wear sleeve on the shaft as shown in Figure 2.3a. The higher the
pressure, the tighter the seal. In a vacuum, however, the lip is
actually pulled away from the shaft. For this reason, there is
also a V-ring seal, shown in Figure 2.3b, used in conjunction with
the rotary seal to maintain vacuum integrity. This seal is a
13
Custom-Made Rotary Seal
Upper Bearing Housing
Figure 2.3
(a)
(b)
(a) Custom-Made Rotary Seal Cross-Section (b) Vacuum Seal Cross-Section
14
standard Forshida V-ring seal, part #400500, which is lubricateq
with a PAG ACME screw grease from Nook Industries. When the
chamber is subjected to a vacuum, the lip of the V-ring seal is
pulled down and forms a tight seal between the shaft and the upper
bearing housing.
10-32 THD Self Locking Inserts 5 Places
Figure 2.4 Specimen Mounting Hole Arrangement on Spindle Face
The spindle serves as the mounting face to accept the upper
specimen holder. Figure 2.4 shows the specimen mounting hole
pattern in the spindle face. It has four 10-32 UNF self-locking
inserts on a 63.5 rom pitch diameter and a fifth center insert for
mounting specimens. Thus, the spindle face can be used to mount a
disk with a center hole or a specimen holder by up to four 10-32
screws. The vertical runout of the face, and th~refore the
specimen, can be adjusted by appropriately tightening or loosening
the three runout screws, shown in Figure 2.2, on the spindle.
15
Through the use of a dial indicator, the vertical runout of this
face can be easily adjusted to less than 0.0013 rnm.
The spindle also has a unique internal passage. This passage
serves to allow a heat transfer fluid to be pumped through the
spindle. This fluid, along with the high thermal conductivity of
the aluminum spindle, basically allow the spindle to remain at a
constant temperature throughout the test. The temperature of both
the upper specimen and the oil, therefore, are also maintained at
a constant temperature throughout a test.
The main bearings of the tribometer are intentionally located
outside the pressure chamber. Since CFCs act as solvents for most
oil-based materials, allowing the refrigerant to come in contact
with the grease in the bearings would degrade the grease and could
cause their premature failure. Recall that the V-ring seal for
vacuum integrity is located above the lower bearing. This
subjects the lower bearing to a vacuum every time the chamber is
purged. To minimize the effects of outgassing of the grease, a
synthetic grease with extremely low vapor pressure was selected.
A PAG, ACME screw grease from Nook Industries, was selected on
this basis as well as its ability to provide adequate bearing
lubrication. The bearings are lightly packed with this grease
prior to installation.
The lower half of the chamber is telescoped into the upper
half by the Z-axis servo motor. This motion is also what provides
loading during the test. The lower half consists of several
critical components shown in Figure 2.2. Also apparent is the
complex suspension system which allows for the accurate
application of load while the chamber is pressurized. These and
16
other features are described in greater detail in the sections to
follow.
10-32 THD Self Locking Inserts 7 Places
Sampling Hole
63.5 mm 0
Figure 2.5 Specimen Mounting Hole Arrangement on Cup
The tribo-contact occurs inside the cup. The cup is a
removeable aluminium piece which serves two important functions.
First, it serves as the mounting surface for the lower specimen
holder. The hole pattern, as shown in Figure 2.5, permits the
holder to be securely mounted by up to seven 10-32 UNF self-
locking inserts. Secondly, the cup serves as the lubricant
reservoir during the test. A Pyrex sleeve, sealed at the bottom
by an O-ring, surrounds the cup. This permits the cup to be
filled with a lubricant, completely submerging the contact to be
tested. Since the Pyrex sleeve is entirely within the pressure
chamber, there is no pressure difference across it, and only the
hydrostatic pressure of the oil requires sealing. Three sight
ports, machined in the wall of the cup, corresponding to sight
17
ports in the chamber walls, allow for viewing of the contact
during testing. Also shown in Figure 2.5 is a small hole labeled
sampling hole. This hole communicates with the sampling port on
the outside of the chamber and allows an oil-refrigerant mixture
sample to be drawn off during a test.
The removeable cup is bolted to a complex force transducer
module. The transducer is outfitted with an intricate array of
strain gages, which are used to measure the forces during a test.
Frictional forces (Fx , Fy), load force (F z ), as well as moment (Mz )
are of interest and are relayed to the control box outside the
pressure chamber. There is also a thermal sensor installed in the
transducer that is used to monitor the temperature of ~he lower
specimen. In general, it is preferred to maintain the transducer
at ambient temperatures, but the thermal design considerations
prevented this option for the tribometer. Accordingly, the strain
gages and adhesives were selected for their ability to function at
elevated temperatures and high refrigerant pressures.
The transducer module is firmly mounted to an internal
suspension system. This consists of a pair of diaphragm springs
which provide compliance in the Z-direction while maintaining high
stiffness in the x, y, and e directions. These diaphragm springs,
in conjunction with a Belleville spring, are used to permit
accurate loading while the chamber is pressurized. When the
chamber is pressurized to 1.725 MPa, it takes approximately
31,000 N to hold the two halves closed. Most of this force is
taken up by the suspension system, so that with proper strain gage
amplifier configuration, test loads as low as 4 N can be
accurately applied and monitored.
18
The lower half of the chamber also houses several safety
devices. There is an adjustable pressure relief valve, preset to
1.89 MPa, to prevent accidental over pressurization of the
chamber. A second pressure safety device is a rupture disk rated
for 2.41 MPa at 1200 C and 2.59 MFa at 210 C. This unit contains
a thin metal disk that will break at the rated pressure and
temperature to relieve the chamber. The other safety device,
shown in Figure 2.2, is an overload stop pin. This pin ensures
that, in the event of an overload, the transducer will bottom out
against the pin and will not be damaged due to buckling.
2.2.2 Temperature Systems
Thermally, the test chamber has several important features
that are described in this section. Virtually all internal
surfaces of the chamber can be heated. This is required to
prevent condensation of refrigerant on these surfaces at the high
test pressures. At the upper pressure limit of the pressure
chamber, 1.723 MPa, the temperature required to prevent
condensation is approximately 71 0 C. Both halves of the pressure
chamber are outfitted with cartridge heaters that are used to heat
the chamber walls above the condensation temperature. The upper
half is outfitted with a 400 W cartridge and the lower half
contains two 500 W cartridges. The temperature of each half can
be controlled from the main control panel.
The temperature of the rotary spindle, and therefore the
upper specimen, is controlled from -300 C to 1500 C by an external
recirculating unit A NESLAB RTE-110 Refrigerated Bath pumps a
heat transfer fluid, through a rotary union, down into the
19
spindle. Due to the high value of the heat transfer coefficient,
and the unique design of the passages machined in the spindle, the
upper specimen can be maintained within a couple of degrees of the
fluid temperature. A mixture of Union Carbide Heat Transfer Fluid
(UCAR) and distilled water serves as the recirculating fluid. By
using the appropriate concentrations of this fluid, different
temperature ranges can be tested. For higher temperature tests, a
ratio of 90% UCAR and 10% distilled water is recommended by the
manufacturer in order to maintain a high boiling point of the
mixture. Testing at temperatures below ambient temperatures is
accomplished solely through the use of the recirculator set to
deliver chilled fluid. A mixture of 50% UCAR and 50% distilled
water should be used to keep the viscosity of the fluid low and
permit easy pumping through the spindle. The controls for setting
temperature are located on the NESLAB unit and allow for setting
the desired temperature as well as providing a safety shutoff for
over-heating. The user should maintain the level of the fluid in
the RTE 110 above the fill line in the reservoir.
The last thermal system to be discussed is the chiller.
Similar to the recirculator, the chiller is an independent unit
with its own controls. A NESLAB CFT-33 Recirculating Chiller
pumps a 50/50 mixture of laboratory grade ethylene
glycol/distilled water through passages machined into portions of
the HPT. This unit is set at ambient temperature and is used to
cool critical parts of the tribometer. The chiller should be
turned on whenever the tribometer is operated.
20
2.2.3 Tribometer Motion
Motion in the tribometer is generated by two independent dc
servo motors. A large 9-axis servo motor provides rotational
motion for the upper specimen; while a second, somewhat smaller,
Z-axis servo motor provides axial motion and loading during the
test. These motors are thoroughly described in this section.
The 9-axis dc servo motor (3 kW) is controlled through a
pulse width modulated (PWM) amplifier. The low inertia motor
coupled with the high performance amplifier provides excellent
response and permits complex motion. The shaft of the motor is
attached through a flexible helical coupling to the shaft entering
the chamber. The position of the 9-axis is monitored by a
differential optical encoder that is used in controlling spindle
motion.
The Z-axis dc servo motor is controlled separately through
its own PWM amplifier. This fast response motor amplifier
combination supplies both Z-axis motion, up to 1.69 mm/sec, and
test loads, up to 4500 N. This motion is transmitted by a lead
screw which is driven through a backlash-free 100:1 harmonic
drive. An encoder feedback loop supplies a means to monitor the
location of the lower half of the chamber; while the transducer,
described in Section 2.2.1, acts as the force feedback loop that
controls the applied axial load. A pair of limit switches ensure
that the Z-axis cannot travel beyond preset positions. The lower
switch protects recirculator lines, while the upper switch is
adjustable and prevents the lower half of the chamber from
telescoping too far into the upper half.
21
N N
~r----. ~L-
Temperature Feedback
Al + RS- 2321 rOisPlayl Temp Control ~ 4 a Board , , r
-Z-Axis . Motor
MotherBoard ----(Microprocessor) -11. .,
-- Motor Amps .. a-Axis - Tribometer - -.. Panel Controls ~ -;7 .. Motor (Mechanical) • Z-Axis i---
• Transducer • a-Axis -. • Temp Sensors ... • Temp
Strain Gage Amplifiers
• Fx ~ ~ ~ • Fy "7 ~ • Fz
• Mz
Z Encoder Feedback , a Encoder Feedback 'r Force Feedback 'r
Figure 2.6 - Tribometer Control System Schematic
2.2.4 TribometerControls
Figure 2.6 shows a schematic of the units that serve as the
tribometer control system. There are four strain gage amplifiers:
Fx , Fy , Fz , and Mz . Each consist of an amplifier board that plugs
into the control box and interfaces with panel controls and
switches as shown in Figure 2.7. Also shown in Figure 2.7 are the
front panels of the two motor control circuit boards, and the
front panel of the temperature control board. The motherboard,
with a microprocessor, interfaces with the control circuits and
provides a wide variety of functions. It interfaces with a four
line LCD display and eight panel switches. In addition to
interfacing with the front panel, the microprocessor can perform
control, limit, and alarm features. It can also communicate with
a personal computer, via an RS-232c port, to allow for external
configuration and data acquisition.
Feedback for the loads is provided by the transducer in the
form of Fx , Fy , Fz , and Mz . Each of the force (torque) directions
has its own independent amplifier that excites the strain gages,
and is used to condition the load to equivalent engineering units.
This allows for each direction to be set to the appropriate
sensitivity so that the accuracy of the force (torque) reading can
be improved. This conditioning of the strain gages is permitted
by proper adjustment of gain and excitation voltage on each
amplifier control panel. A detailed explanation on how to
determine the proper settings is given in Section 4.5.
The Z-axis motor control board permits axial loads to be
applied during the test as well as providing motion to open and
close the chamber. The force can be static, up to 4500 N, or
23
oscillatory from O-~500 N at frequencies up to 5 II~. In addition,
this oscillation can be synchronized with the 8-axis oscillation
so that the load and 8-motion can be in phase. These controls can
also be used to set the speed at which the lower half of the
chamber can be raised or lowered.
Figure 2.7 Photograph of HPT Main Control Panel
The 8-axis motor control loop, with optical encoder feedback,
permits the 8-axis servo motor to be precisely controlled. The
motor is capable of simple unidirectional rotation (0-2000 rpm)
and oscillatory motion with amplitudes of up to 180 0 and
frequencies of up to 5 Hz. Presently, the tribometer controls
permit oscillatory motion with either sinusoidal or triangular
waveforms. The controls also permit a const~nt torque to be
applied to the test. In this mode, Mz is used as the feedback
24
signal and the speed will automatically adjust to whatever is
necessary to produce the required torque.
The on-board temperature controllers are used to control the
temperatures of the cartridge heaters. The upper and lower
heaters can be independently set from ambient to 95° C. Thermal
sensor feedbacks from the chamber allow the temperatures to be
accurately controlled to ±1° C. The recirculator and chiller can
be turned on from the front panel, but their controls are housed
within their independent units and their functions will be
described in Section 4.4
The last tribometer control is the microprocessor. The Intel
80C188EB 16-bit microprocessor, located on the motherboard, is
capable of independently controlling nearly all tribometer
functions. It can be used to control the two cartridge heaters as
well as 9-axis and Z-axis motion. The interface to the
microprocessor consists of a front panel keypad with eight
switches and a 4 line ~ 40 character LCD display. This display is
used to set test parameters and monitor data values. The
microprocessor can also interface with a personal computer to
permit for more complicated functions.
2.3 External Equipment
The HPT has also been outfitted with apparatuses for purging,
charging, and sampling, as shown in Figure 2.1, as well as a data
acquisition system. Two vacuum pumps work in tandem to purge the
system. The larger of the two pumps is used to purge the chamber
and external lines; while the smaller pump removes any vapor
outgassed from the grease in the main bearings. An external
25
pressure vessel is used to charge the chamber with refrigerant. A
silicone heating blanket around the vessel is used to generate the
required refrigerant pressure to charge the chamber. A 13.6 kg
refrigerant tank, attached to the pressure vessel, is used to
supply refrigerant to the vessel. The chamber is also outfitted
with a 50cc sampling cylinder which can be used to siphon off an
oil-refrigerant mixture sample during a test. There is also a
separate 6.8 kg drain tank which collects used refrigerant so that
it can be recycled. Data acquisition is possible through the use
of a PC linked to the motherboard by an RS-232c connector.
2.3.1 Purging Facility
In order to purge the system prior to initiating a test, a
facility to evacuate the chamber was developed. This facility is
shown in Figure 2.8. A Welch Scientific Duo-Seal Vacuum pump,
model 1402, is connected to the pressure chamber at a Whitey Model
60 ball valve by a 25.4 mm ID vacuum hose. This large pump is
able to rapidly evacuate the chamber and external lines, of the
charging facility and drain tank, to a very high vacuum. Recall
that the V-ring Forshida seal, placed above the lower bearing, was
installed for vacuum integrity. This places the lower bearing,
and its grease, in a high vacuum environment each time the chamber
is purged. To remove any vapor that is outgassed from the grease,
a second vacuum pump is used. The Thermal Engineering Co, model
1825Z, vacuum pump is connected by a (9.525 mm 0 by 0.9 m) Yellow
Jacket charging hose to a port located between the V-ring and
bearing. This serves to effectively remove any outg~ssed vapor.
26
Note: Yellow Jacket hose connects to the Lower Bearing Housing
Whitey Model Ball Valve (B-635S12-B)
25.4 rom 0
25.4 Hose
Welch Scientific "Duo-Seal" Vacuum Pump Model 1402
HPT Pressure Chamber
1/4" Whitey Needle Valve (B-14F4)
Thermal Vak-Check Model 4501
Yellow Jacket 9.525 rom 0 Charging Hose (0.9 m)
Thermal Eng. Vacuum Pump Model 1825Z
Ashcroft Vacuum Gage (SS-1279SS-0-30Hg)
Figure 2.8 HPT Purging Facility
The level of the vacuum is monitored by two separate vacuum
gages. The first is an Ashcroft model SS-1279SS-0-30v gage,
accurate to 0.5%, which is used to give a rough vacuum level in
inches of mercury. The second gage, a Thermal Vak-Check model
4501, is capable of extremely accurate readings, in microns, down
to an absolute vacuum. Both gages are isolated from the chamber
by a 1/4" Whitey needle valve, model B-NF4-B. This valve is used
to protect the gages from damaging positive pressures.
27
2.3.2 Charging Facility
In order to supply the chamber with refrigerant vapor at
pressures up to 1.725 MFa, a charging facility was developed.
Figure 2.9 shows the relevant components of this facility. The
3.63 kg pressure vessel, manufactured by E.F. Britten & Co, is
used to hold the refrigerant being evaluated. The vessel is
supplied with refrigerant from a standard DOT-39NRC 260/325, M1073
E. I. du Pont de Nemour & Co container with 13.6 kg capacity. A
720 W silicone heating blanket, manufactured by Conrad Corp, is
wrapped around the pressure vessel and supplies heat to generate
pressure within the vessel. A standard 5 amp 0-120 volt variac
supplies power to the heating blanket and is used to control the
temperature of the pressure vessel. The pressure is monitored by
an Ashcroft SS-1279SS02c30-300 dial pressure gage with an accuracy
of 0.5%. There is also an adjustable pressure relief valve,
Swagelok model SS-4CPA2-150, which is set at 2 MFa to relieve the
pressure vessel for safety purposes. Once sufficient pressure is
generated, a series of valves connecting the vessel to the chamber
can be opened, allowing pressurized refrigerant vapor to be
transferred to the chamber. This valve assembly consists of a
Swage 10k SS-43F2 valve at the pressure vessel, which connects to
Nupro Union Tee, SS-400-3, at a quick-connect (not shown in Figure
2.9). From the tee, a 1/4" Nupro plug valve connects to a 1/4" 0
Nupro hose, which connects to the pressure chamber at a Whitey
1/4" needle valve. The quick connect allows the pressure vessel
to be disconnected and weighed to determine the approximate amount
of refrigerant used in a test.
28
1/4" 0 Nupro Hose SS-7R4TA4TA4-48
1/4" Nupro Plug Valve (SS-4P4T-TB)
Nupro Union (SS-400-3) .
1/4" Nupro Plug Valve (SS-4P4T-TB)
1/4" 0 Nupro Hose SS-7R4TA4TA4-36
DuPont 13.6 kg Supply
Tank
Whitey 1/4" Needle Valve (B-1RF4)
HPT Pressure Chamber
Ashcroft Pressure Gage (SS-1279SS02C30-300)
Nupro Cross (SS-200-4)
Conrad Corp 720 W Silicone Heating Blanket
Nupro Relief Valve (SS-4CPA4-150)
Variac 5 Amp (0-120 V)
EF Britten & Co 3.63 kg DOT-4B Pressure Vessel
Figure 2.9 HPT Charging Facility
2.3.3 Sampling Facility
The design of the HPT permits easy sampling of the oil-
refrigerant mixture during a test. The sample is taken while the
chamber is pressurized, so the amount of refrigerant saturated
into the oil can be determined by weight analysis. Figure 2.10
shows the sampling facility of the HPT. A sampling port on the
chamber communicates with the cup by a 1.275 mm 0 capillary tube.
A 1/4" Nupro plug valve, model SS-2P4T-TB, is connected to the
sampling port on the chamber. This valve permits the chamber to
be sealed off from the outside. The valve is connected to a
29
Swagelok quick-connect, model SS-QF4, so that the sample cylinder
can be removed from the chamber while the chamber is still under
pressure. The sample cylinder, a 50 cc Swagelok SS-4CS-TW-50, has
a 1/4" Nupro plug valve, model SS-6P4T-TB, which is used to seal
the sample cylinder. This setup allows the oil-refrigerant to be
weighed on a Sartorius model LC1200S balance, with an accuracy of
0.001 g. The refrigerant is then evaporated and the weight percent
of refrigerant saturated into the oil may be determined.
Quick Connect (Swagelok: SS-QF4)
1/4" Plug Valve (Nupro: SS-6P4T-TB)
50 cc Sample Cylinder (Whitey: SS-4CS-TW-50)
/1/4" Plug Valve ~ (Nupro: SS-2P4T-TB)
-'Connects to Sample Port On HPT
Figure 2.10 HPT Sampling Facility
2.3.4 Data Acquisition
Data acquisition is accomplished by an IBM compatible
personal computer with an 80386SX processor running at 20 MHz.
The PC is connected by ribbon cable directly to the microprocessor
on the motherboard of the HPT by an RS-232c connector. The PC and
microprocessor use SECS-I communication protocol to communicate
with each other.
The interface program is written in Magic/L, a powerful yet
simple language which only requires 64 k to run. This low memory
requirement makes it an ideal program for use with the
30
microprocessor. The interface is command-line based and actions
are implemented by entering key words or commands on the PC
keyboard. The PC can be used to configure the strain gage
amplifiers, set the cartridge heater controls, and monitor the
output from the test. The maximum sampling rate is determined by
the number of parameters being sampled. In general, by sampling
all forces and all temperatures, samples can be taken every
40 msec. These data are read directly into a Ramdisk and then
transferred to the hard drive. They may later be imported as a
numeric file into Lotus 1-2-3 for further analysis and data
smoothing.
31
3.1 Overview
CHAPTER 3
EXPERIMENTAL METHOD
The goal of this research project is to use specimen testing
to accurately model the critical tribo-contacts of compressors in
a high pressure tribometer. The specimens should be able to
accurately simulate the geometries and materials found in the
actual critical tribo-contacts. In addition, the operating and
environmental conditions such as load, speed, pressure,
temperature, and lubricant need to be accurately simulated with
the HPT.
This chapter will describe the tribo-contacts under
evaluation. This includes examination of specific compressor
types, determination of their critical contact geometries, and
determination of actual material pairs used in the contacts. The
environmental operating conditions under which these contacts
operate will be described. Finally, lubricants will be identified
and described on a general bulk material property basis.
3.2 Compressor Contacts
Three critical contacts from mobile air conditioning and
refrigeration compressors were chosen to be modeled in the HPT
test program. After reviewing the requirements of the sponsors
with the function and use of many compressors, it was decided that
the rolling piston, swash plate, and reciprocating piston
compressors should be the focus of this research. The swash plate
compressor is primarily used in automotive air conditioning
32
systems, while the other two are normally used in refrigeration.
systems. The contacts chosen were the most critical based on
input from the sponsors. The actual compressor contacts and their
equivalent specimen models are described in the sections to
follow.
Piston: Gray Cast Iron Tool Steel
Suction Port Discharge Port
Cylinder
Figure 3.1 Schematic of Rolling Piston Compressor Section
3.2.1 Counterformal Contact - Rolling Piston Compressor
The critical contact for the rolling piston compressor is
shown in Figure 3.1. The piston is driven by an eccentrically
mounted shaft which produces oscillatory relative motion between
the vane and piston [18]. This contact serves as the primary seal
between the low pressure refrigerant on the suction side and the
33
high pressure refrigerant on the discharge side of the compressor.
Therefore, since the oil is subjected to relatively high pressure
refrigerant, more refrigerant is saturated into the lubricant, and
the need to model the high pressures is crucial.
Figure 3.2
Equivalent Radius (R=3.175 mm)
Section A-A
Upper Specimen: 76.2 mm 0 Flat Disk (Piston)
19.05 mm
Lower Specimen: 6.35 mm 0 Pin (Vane)
tLoad
A
.. . • I • ~
A
Equivalent Geometry of Counterformal ContactRolling Piston Compressor. Note that the lower specimen is secured in place by a specimen holder.
The vane is made from a hardened tool steel, while the piston
is a hardened gray cast iron. The geometry of this contact pair
is counterformal, and is modeled by using the concept of
34
equivalent cylinders that is developed in Appendix A. The
equivalent geometry, shown in Figure 3.2, is modeled as a
cylindrical pin rubbing against a flat plate. The relevant
geometrical and material data for the model are shown in
Table 3.1. The geometry, materials, and surface treatments for
this contact are all accurately modeled in the HPT.
a e . :ipec~men T bl 3 1 S P ropert~es
Counter-Area Description formal Conformal
Contact Contact Contact
Geometry 76.2 mm0 76.2 mm0 25.4 mm • Upper Flat Disk Flat Disk Square Pad
• Lower 6.35 mm0 5.08 mm0 6.35 mm 0 Pin Pin Flat Shoe with 1220 mm 0
L=9.52 mm L=9.52 mm Materials
• Upper Gray C.l. Ductile C.l Die Cast Al
• Lower Tool Steel Bronze Mild Steel Hardness (±1 Rc)
• Upper 50 42 ---• Lower 65 --- 63
Surface Topography Ground Ground Ground • Upper
• Lower Ground Lapped Ground Surface Finish (Ra)
• Upper (±O. 025 J.Lm) 0.127 0.127 0.254
• Lower 0.127 0.203 0.101
The operating conditions used in the specimen testing program
are also representative of those found in,typical rolling piston
compressors. These conditions are shown in Table 3.2. The type
of motion, speed, pressure, and temperature are all consistent
with the actual compressor operation; while the load was increased
to obtain measurable wear in a reasonable test duration of one
hour. The contact load, shown in Table 3.2, for the counterformal
35
contact is based on the Hertzian contact pressure between the
specimens [19].
Table 3.2
Operating Conditions
Conditions
Area Contact
3.2.2 Area Contact - Swash Plate Compressor
The critical contact examined for the swash plate compressor,
the shoe-plate, is shown in Figure 3.3. The swash-plate drive
mechanism features an i~clined plate which is rigidly attached to
the rotating shaft. The unidirectional rotation of the shaft
causes a nutation of the plate, which is used to transform the
shaft rotation to simple reciprocal motion at the piston. This
contact occurs within the crankcase and, therefore, the oil is not
normally subjected to refrigerant at as high a pressure as is
found in the rolling piston compressor.
The shoe is made from a bronze alloy that has a lapped
surface, and the plate is a hardened ductile cast iron with a
ground surface. The shoe-plate pair form an area contact. The
flat shoe rubs, under simple sliding conditions, against the flat
. plate. The load is transferred from the piston through a steel
ball onto the bronze shoe. This assures that the shoe can self
align with the plate during operation. The equivalent geometry,
36
shown in Figure 3.4, is essentially the same as that found in the
compressor. The area contact is made up of a bronze shoe, loaded
by a steel ball to permit the same degree of freedom found in the
actual compressor, sliding against a flat ductile cast iron plate.
Table 3.1 shows the geometry, material, and treatments used in
modeling the area contact. These data correspond identically to
those found in the actual compressor.
Shoe: Bronze Ductile Cast Iron
Figure 3.3 Schematic of Swash Plate Compressor Section
The modeled operating conditions for the area contact, shown
in Table 3.2, are essentially the same conditions of the swash
plate compressor with the exception of load and speed. When the
area contact was run at loads and speeds equivalent to the
compressor, hydrodynamic liftoff was observed. In order to avoid
liftoff and generate measurable wear, it was necessary to run the
area contact at higher loads and slower speeds than what is
37
typically encountered in the swash plate compressor. The contact
load, shown in Table 3.2, is given as a contact pressure. This
pressure is the test load (P), divided by the contact area (A) of
the bronze shoe. The temperature and refrigerant pressure modeled
in this program are typical of those found in the actual
compressor.
Upper Specimen: 76.2 rom 0 Flat Disk (Plate)
•
Lower speCimen:~t-5.08 rom 0 Pin (Shoe) Load
A
A
Section A-A
Ball to avoid edge loading
Figure 3.4 Equivalent Geometry of Area Contact-Swash Plate Compressor. Note that the lower specimen is secured in place by a specimen holder.
3.2.3 Conformal Contact - Reciprocating Piston Compressor
The critical contact chosen for the reciprocating piston
compressor is that of the wrist pin-bearing as shown in
Figure 3.5. The compression of refrigerant vapor is accomplished
by a piston which is driven through a connecting rod by a
crankshaft. The critical contact of this compressor occurs
between the wrist pin and connecting rod. This contact pair is a
conformal contact which essentially is a journal bearing. The
motion observed at this contact is sinusoidal. As with the area
38
contact, the conformal contact occurs within the crankcase and,
therefore, is not normally exposed to the high refrigerant
pressures found in the counterformal contact.
Wrist Pin: Mild Steel
~~~ischarge
~~~ort ~~~~~~~~~
Connecting Rod: Die Cast Aluminum
Figure 3.5 Reciprocating Piston Compressor Section
The wrist pin is made from case hardened mild steel, which
experiences sliding motion with a die cast aluminum connecting
rod. The geometry of this contact is also modeled by using the
concept of equivalent cylinders that is developed in Appendix A.
The equivalent geometry, shown in Figure 3.6, is modeled as an
approximate equivalent radius, R=1220 mm, rubbing against a flat
plate. The radius is approximate because of machining
limitations. The remaining contact conditions, shown in Table
3.1, are identical to those found in the actual compressor.
39
Upper Specimen Holder
Upper Specimen 25.4 rom Square Pad (Con. Rod)
19.05 rom
Connecting Rod
A
Lower Specimen: 6.35 rom Square Pin ·(Wrist Pin)
tLoad
A
Approximate Equivalent Radius (R=1219 rom)
Section A-A
Figure 3.6 Conformal Contact-Equivalent Geometry of Reciprocating Piston Compressor. Note that both the upper and lower specimens are mounted in specimen holders in the HPT.
Table 3.2 shows the operating conditions for the conformal
contact. All of these conditions accurately represent the
conditions found in the actual compressor. The motion is
approximately sinusoidal and the resulting maximum speed is
representative of the actual compressor. The only exception is
the load. The load is increased to obtain measurable wear within
40
a reasonable test duration. The load, shown in Table 3.2, is
given in terms of a P/LD pressure. This pressure, commonly used
in journal bearing theory, is found by dividing the applied load
by the projected area of the pin. The remaining conditions,
refrigerant pressure and temperature, are identical to those
typically found in the actual compressor.
3.3 Specimen Holders
In order to effectively be able to test the equivalent
geometries of the critical compressor contacts, specimen holders
were designed and made. The upper specimens for both the
counterformal contact and the area contact are 76.2 mm 0 flat
disks that are mounted directly to the spindle. They, therefore,
do not require any specimen holders. The remainder of the
specimens are held in place with various specimen holders.
The upper specimen for the conformal contact is made of a die
cast aluminum alloy. Due to difficulties in obtaining this
material cast in 76.2 mm 0 flat disks, and the fact that the
conformal contact is subjected to oscillatory motion with only
± 20° amplitude, a smaller specimen was used. The square aluminum
pad, shown in Figure 3.6, is a 25.4 mm square pad with a thickness
of 6.35 mm. This pad is sufficiently large to simulate the motion
of interest, but it is too small to mount directly to the spindle.
The required specimen holder is shown in Figure 3.7. The specimen
holder is made from a mild steel disk with a 25.4 mm slot to
accept the aluminum pad. The pad is held securely in the holder
by a pair of 4-40 UNC set screws. The specimen holder, with
mounted aluminum pad, can then be secured directly to the spindle
41
by four 10-32 machine screws. This configuration provides secure
mounting of the aluminum pad, while allowing the contact to
experience the appropriate motion to model the wrist pin-bearing
contact in the compressor.
Drilled and C'Bored for 10~32 socket head machine screws 4 @ 900 on 63.5 mm 0 circle (Ref)
/////// //////// //////// //////// //////// //////// ////////
- - - - i - .. ! - ::: !: !-.- - - - -- - - - ~~.:r = ......:;.....:;...:~~....:...~- i:.::-.r - - --
• • • • • •
6.35 mm
76.2 mm 0
Aluminum Pad (Lower Specimen)
Drilled and Tapped for 4-40 set screws 4 Places
Figure 3.7 Upper Specimen Holder for Conformal Contact
42
Specimen
6-32 UNC Set Screws
A
1
I
..-l A
Section A-A
Figure 3.8 Lower Specimen Holder for Counterformal and Conformal Contacts
All the lower specimens require specimen holders. The lower
specimens for the counterformal and conformal contacts utilize the
same specimen holder, while a different holder secures the
specimen used for testing the area contact. The lower specimen of
the counterformal contact, the tool steel pin, has a 2.38 mm 0
hole drilled radially through it. Similarly, the steel pin of the
conformal contact also has a 2.38 mm 0 hole drilled through it.
43
The aluminum specimen holder, shown in Figure 3.8, has a series of
2.38 mm holes drilled through it and a 6.35 mm slot milled into
the face. The specimens are held in this slot by a 2.38 mm 0
steel dowel that is held in place by two 6-32 UNC set screws.
This setup allows the specimen to self align in the direction
along the milled slot while retaining rigidity transverse to the
slot. The lower specimen holder is then bolted directly to the
cup with two 10-32 UNF machine screws.
Bronze Shoe (Specimen) 5.556 mm 0 Ball
Section A-A
o
A
I ~
A
00 00 ~
Figure 3.9 Lower Specimen Holder for Area Contact
44
The only other specimen holder required is for the bronze
shoe of the area contact. In order to accurately represent the
degree of freedom that the contact in the actual compressor
experiences, the specimen holder, shown in Figure 3.9, has a
5.556 mm 0 spherical joint, with a 52100 steel ball placed between
the specimen holder and the bronze shoe. The degree of pivoting
is restrained by the clea~ance between the shoe and the specimen
holder. This ensures that the shoe cannot be dislodged from the
specimen holder during a test.
3.4 Lubricants
The lubricants that have been evaluated are classified into
four types: mineral oils, alkylbenzenes, polyalkylene glycols
(PAGs), and polyolesters. The first two are used with CFC-12 for
obtaining baseline friction and wear data; while the latter two
are the more promising oil types for use with HFC-134a. The
mineral oils are currently used in the swash plate and
reciprocating piston compressors, and the alkylbenzene is used in
the ro~ling piston compressor. The PAGs tested are possible
replacements for swash plate compressors, while the rolling piston
and reciprocating piston compressors will possibly be lubricated
by polyolesters. Regardless of the intended application, both the
PAGs and esters were cross tested with different contact pairs.
The relevant lubricant properties for all the oils tested are
shown in Table 3.3.
The mineral oils used with CFC-12 are classified according to
their molecular structure as either paraffinic, naphthenic, or
aromatic. The paraffins are characterized by containing only
45
linear or branched carbon chains and are saturated hydrocarbons.
Naphthenic mineral oils are also saturated hydrocarbons, but they
are made up of ring structures. Aromatics are unsaturated
hydrocarbons that have a ring structure with double bonds
alternating between rings. Mineral oils that are classified as
paraffinic or naphthenic, also contain aromatic hydrocarbons but
the concentration can vary from 10% to 30%. The classification of
the oil always refers to the saturated hydrocarbon structure,
either naphthenic or paraffinic. The synthetic alkylbenzene oils
that were tested are classified as 100% aromatic hydrocarbons.
Due to problems of insolubility of HFC-134a with mineral or
alkylbenzene oils, other synthetic replacements have been
developed for use with HFC-134a. The two most promising types
that are used in this program are PAGs and esters. PAGs are
classified into families by their free hydroxyl groups. All of
the PAGs tested are known as monoethers, having one hydroxyl
group. The ester lubricants are classified depending on the
functionality of the alcohol or acids used in their manufacture.
The esters used are all of the pentaerythritol (PE) variety, a
polyfunctional alcohol.
Where possible, both base (B) and formulated (F) versions of
each lubricant were tested. Although proprietary in nature, the
formulated oils are versions of the respective base oil with an
additive package to improve lubricative properties. These
additives can include extreme pressure (EP) agents, oxidation
inhibitors, foam reducers, and other types of agents to improve
lubrication characteristics. The exact nature of the additive
46
packages for the lubricants tested were not supplied by the
manufacturer because of proprietary reasons.
Oil-refrigerant mixtures are classified as completely
miscible, partially miscible, or immiscible according to their
mutual solubility with each other. Completely, or fully, miscible
oil-refrigerant mixtures are mutually soluble at any temperature.
Therefore, with the appropriate refrigerant pressure, the oil
refrigerant mixture can exist anywhere from zero percent
refrigerant to zero percent oil as a single phase at any
temperature. Partially miscible oil-refrigerant systems are
mutually soluble to a limited extent. Above a critical solution
temperature, the oil-refrigerant mixture is completely miscible.
Below this critical temperature, however, the liquid may separate
into two phases. This phase separation does not imply that the
oil and refrigerant are insoluble in each other, rather that each
phase is a separate solution. One may be oil-rich and the other
refrigerant rich and these two mixtures are immiscible with each
other. Table 3.4 shows the relevant miscibility data as well as
the weight percent of refrigerant in the oil as tested. All of
the lubricants tested, with the exception of the PAGs, are
completely miscible with their respective refrigerants. The PAGs
are only partly miscible and exhibit inverse miscibility with HFC-
134a. This type of immiscibility is characterized by a lower
critical solution temperature (LCST). At temperatures below the
LCST, there is complete miscibility; while at temperatures higher
than the LCST, the oil and refrigerant are immiscible. For the
conditions tested in this program, however, all of the PAGs are in
the fully miscible regime.
47
.J::> co
Table 3.3 Lubricant Properties
Oil Oil 1 Number Type Family Additives
Min1 Mineral Oil Para No Min2 Mineral Oil Naph No Alkbenz-B Alkylbenzene Arom No Alkbenz-F Alkylbenzene Arom Yes PAG1-B Polyalkylene glycol Mono No PAG1-F Polyalkylene glycol Mono Yes PAG2-B POlyalkylene glycol Mono No PAG2-F Polyalkylene glycol Mono Yes Est1-B Polyolester PE No Est1-F Polyolester PE Yes Est2-B POlyolester PE No Est2-F Polyolester PE Yes Est3-B Polyolester PE No Est3-F Polyolester PE Yes
1 Arom- Aromatic Para- Paraffinic Mono- Monoether Naph- Naphthenic PE- Pentaerythritol
Viscosity Il (cS)
@40o C @100o C
10? 11 12 12 2.6 57 5.8 57 5.8
135 25 135 25 100 20 100 20
23.94 4.88 23.9 4.87 91.37 10.19 91. 4 10.18 11.5 2.8 11.5 2.8
.t:. ~
Oil Number
Min1 Min2 Alkbenz-B Alkbenz-F PAG1-B PAG1-F PAG2-B PAG2-F PAG2-B PAG2-F Est1-B Est1-F Est2-B Est2-F Est2-B Est2-F Est3-B Est3-F
Ref Type
R12 R12 R12 R12
R134a R134a R134a R134a R134a R134a R134a R134a R134a R134a R134a R134a R134a R134a
Table 3.4 Lubricant Test Conditions
Temp Press. Weight % Miscibility Contact Type (oC) (MPa) Ref in Oil
Full Area 73.9 0.172 4.9 Full Conformal 100 0.172 1.7 Full Counterformal 80.6 1. 550 42.5 Full Counterformal 80.6 1.550 41.1
Partial Area 73.9 0.172 3.1 Partial Area 73.9 0.172 2.5 Partial Conformal 100 0.172 1.8 Partial Conformal 100 0.172 ---Partial Counterformal 80.6 1.550 17.1 Partial Counterformal 80.6 1. 550 21.2
Full Counterformal 80.6 1. 550 32.9 Full Counterformal 80.6 1. 550 28.5 Full Area 73.9 0.172 0.3 Full Area 73.9 0.172 0.5 Full Conformal 100 0.172 1.4 Full Conformal 100 0.172 1.9 Full Conformal 100 0.172 1.3 Full Conformal 100 0.172 0.9
Table 3.4 also shows the weight percent of refrigerant
saturated into the oil. These data were obtained by using the
sampling facility described in Section 2.3.3. The sampling
procedure is based on ASHRAE standards [20]. In this procedure, a
sample of the oil-refrigerant mixture is taken during the test and
then weighed. The refrigerant is then carefully removed from the
oil, and the final weight of the oil is measured. The amount of
refrigerant saturated into the oil can then be determined.
50
4.1 Overview
CHAPTER 4
EXPERIMENTAL PROCEDURES
In order to assure repeatability of results, the same testing
procedure was used for all tests. The specimens are cleaned to
remove surface residues prior to installing them into the chamber.
The appropriate lubricant is then added to the cup and allowed to
reach thermal equilibrium before continuing. The chamber is
purged to remove atmosphere and then charged with the required
refrigerant. Then, the system is allowed to reach pressure and
thermal equilibrium before initiating the test. After the test is
completed, an oil-refrigerant sample is taken, and the specimens
are analyzed for wear and surface films. The specific details of
these procedures are examined in this chapter.
4.2 z-Axis Motion Control
Before the HPT can have specimens installed or be charged
with ~ refrigerant, a basic understanding of how to operate the
HPT Z-axis motion is required. Figure 4.1a shows the motion
control panel of the main control box. This panel controls the Z
axis servo motor, which provides motion for the lower half of the
chamber. The panel is divided into four separate sections. These
sections, along with their functions will be described in brief.
The upper section consists of a digital display linked to a
six-position rotary switch. This display is used to monitor any
of the six values for the switch. The display reads out in volts
from -10.0 v to 10.0 v. The equivalent engineering units to the
51
voltage reading can be determined by knowing the appropriate
conversion factors. Speed is displayed in volts, which converts
to SI units of 0.169 mm/sec/volt. The force reading depends on
the strain gage amplifier settings for the Z-axis, which will be
described in detail in Section 4.5. The Position setting
displays the axial position relative to some initial platen
position. The Position setting is not needed for basic HPT
functioning. The Setpoint reading displays the voltage that has
been set on the potentiometer in the Setpoint section of the
panel. Likewise, Frequency and Amp1itude readings display the
volt equivalents of the values set on the potentiometers in the
Osci11ation section of the control panel. Frequency reads out in
volts which has 0.5 Hz/volt equivalent, and amplitude depends on
the strain gage configuration.
The Contro1 Mode, located at the bottom of the panel, has
three toggle switches. The Amp1ifier switch is used to power up
the amplifier for the Z-axis servo motor. The Motor switch is
used to Enab1e/Disab1e the output stage of the motor amplifier
and thereby causes motion to start/stop. The Amp1ifier On/Off
switch is analogous to the ignition of an automobile and the
Enab1e/Disab1e switch is analogous to the clutch. The
CPU/Pane1/MCC switch allows the user to select among
microprocessor, panel, or personal computer control by a motor
control card. The initial positions for these three switches when
the main power is turned on for the tribometer is: the
Enab1e/Disab1e should be in the Disab1e position, the
Amp1ifier should be Off, and the CPU/Pane1/MCC switch should be
in the Pane1 position.
52
Vertical Axls-Z Rotation Axls- 9
-19.99 -19.99
Poaltion ~ Setpoint Force Frequency
Speed 0 Amplitude
PosItion ~ Setpoint Force Fntquency
Speed 0 Amplitude
Display· Display
Setpolnt Setpolnt
Travel Rele
o _Force_ @ PosItion Offset
Speed Rele
o _Force- @ Position Offset
LP ON
o o [XWII OON
r Oscillation r Oscillation """ + @ @ -0 Off
Spindle Sync. Fl1Iquency Fl1Iquency
Sine @ Sine @ 0 0
Triangle Triangle
A""I.ude A""I.ude
Waveform "- Waveform ~ ~
r """
r '" Enable 01 Enable 01
0 0 0 0 Disable Off Disable Off
Molor Ampllf. Molor AmpHf.
CPU CPU o Panel o Panel
MX MX
"- Control Mode Control Mode
AMn AMn
(a) (b)
Figure 4.1 (a) Z-Axis Motion Control Panel
(b) 9-Axis Motion Control Panel
The next section of the panel, the Oscillation section, is
only of use if oscillatory loads are applied. Although this
function is not used for any tests in this program, its functions
53
are included here for reference. Two potentiometers can be used
to set the frequency and the amplitude at which the load is
applied. It should be noted that only positive values of
amplitude are possible since a contact cannot support tension.
The Amplitude setting is used in conjunction with the Setpoint
section. The load will oscillate about the setpoint value with an
amplitude set with the Amplitude potentiometer. The waveform may
be either triangular or sinusoidal and is set with a separate
toggle switch. The Spindle Sync. switch is incorrectly marked
with the (-) and (Off) positions reversed. The Spindle Sync.
function synchronizes the motion of the vertical axis to the
oscillation of the rotational axis. In the activated. (+/-)
position, the frequency control of oscillation is transferred to
the value set with the 8-axis control. This synchronizes the
maximum and minimum load application with the maximum and minimum
8-position during rotational oscillation.
The Setpoint section of the Z-axis panel is the main control
point for the Z-axis motor. This section has a three position
toggle switch which is linked to a potentiometer. Either the
speed of travel, force applied, or relative position is set by
putting the toggle switch in the appropriate position and then
adjusting the potentiometer. Only one of these modes, Travel,
Force, or Position can be set and controlled at anyone time. It
is not possible, therefore, to simultaneously set both the speed
at which the platen travels and the force applied during the test.
The resulting value will be shown as a voltage on the display so
long as the six-position rotary switch is in the Setpoint
position. The speed can be set from 0-10 volts
54
(0.169 mm/sec per volt), the force is set from 0 to Fmax
(determined by strain gage amplifier configuration), and position
is relative from some initial position. The direction, either up
or down, can be set by proper position of the Up/Down toggle
switch.
Now that a basic understanding of the layout of the Z-axis
motor control section has been presented, the motor can be powered
up. The amplifier should not be turned on under load, so before
doing so, it is wise practice to zero all the potentiometers.
This is done by turning them fully counterclockwise until they hit
a stop. Having done this, the Amplifier switch can now be turned
On. The Motor switch should, however, remain in the Disable
position until motor motion is required.
4.2.1 Travel Mode
With the switch in the Setpoint section in the Travel
position, the potentiometer can be used to set the speed at which
the chamber opens or closes. The speed will be a function of the
setpoint of the Rate potentiometer. The speed can be read by
turning the rotary switch to the Speed position. The rate shown
is given as a voltage·from 0 to 10 volts. This corresponds to an
axial speed of 0 to 1.69 mm/sec. Once the rate and direction is
set, the motion can be started by turning the Motor switch to the
Enable position. The Rate potentiometer can be turned clockwise
to speed up the motion, or counterclockwise, to slow it down.
Even though limit switches provide good protection against over
travel, the platen should always be monitored while in motion.
This machine is capable of applying very large forces, which could
55
potentially damage the machine or injure the operator. Once the
desired position of the lower half of the chamber is reached, the
Motor switch should be turned to the Disab1e position, and the
potentiometers should be turned fully counterclockwise.
4.2.2 Force Mode
This mode is used to apply a constant load during a test. To
place the HPT in this mode, the toggle switch in the Setpoint
section should be in the Force position. The potentiometer, also
in this section, can now be used to set the applied force from 0
to 10 volts. This voltage is read on the display when the six
position rotary switch is in the Setpoint position. The actual
force equivalent to the voltage reading depends on the settings of
the Z-axis strain gage amplifier panel. The instructions for
configuring the strain gages are given in Section 4.5. From this
calibration setting, the 0 v reading corresponds to 0 N while a
reading of 10 volts corresponds to 4445 N. The force signal from
the transducer is read by the strain gage amplifier and compared
to the Setpoint value. The motor control then adjusts the
position of the lower half of the chamber until the desired force
is attained. This loop permits the force to be accurately applied
and controlled during a test.
Before the tribometer is used in force mode, several
conditions should be met. The amplifier should be turned on for
at least one hour prior to use. This allows the circuit boards to
warm up so that accuracy is improved. The tribometer should also
be at thermal and pressure equilibrium at the desired test
conditions. Pressure can slightly compress the lead screw of the
56
Z-axis motor and therefore cause slight changes in applied load.
If pressure equilibrium is not attained, the force readings can be
in error. Temperature has a very significant affect on the strain
gages of the transducer. Therefore, thermal equilibrium is
required before starting force mode control to ensure accurate
readings from the strain gages.
4.3 9-Axis Motion Control
The form of the 9-axis control panel, shown in Figure4.1b,
is similar to the Z-axis panel. It is divided into four sections:
Display, Setpoint, Oscillation, and Control Mode. By proper
manipulation of these settings, the 9-axis dc servo motor can be
controlled in a variety of ways. The description of this panel is
included in this section.
The Control Mode section of this panel contains three toggle
switches: Motor Enable/Disable, Amplifier On/Off, and
CPU/Panel/MCC. The functions of these switches are identical to
those in the Z-axis control panel so the reader is referred to
that section. As with the Z-axis, the CPU/Panel/MCC switch
should remain in the Panel position to allow the motor to be
controlled by panel settings. Also, all the potentiometers should
be zeroed before the amplifier is turned on. The Oscillation
section is also similar to the Z-axis panel. Waveform can be set
to either sinusoidal (Sine) or triangular (Triangle), and the
Frequency and Amplitude potentiometers function identically to
those in the Z-axis panel.
The Setpoint section, although identical in form, functions
slightly differently than the Z-axis panel. The most common mode
57
for this axis is the Speed position. With the toggle switch in
this position, the Rate potentiometer can be used to adjust the
speed from 0 to 10 volts. This corresponds to a rotational speed
of 0 to 2000 rpm, so the equivalence factor is 200 rpm/volt. The
voltage value for the speed can be read in the Disp1ay section
when the six-position rotary switch is in the Speed position. The
direction of rotation in set by using the ON/CON switch. The
convention used for clockwise/counterclockwise rotation is taken
from a position facing the shaft coming out of the motor.
The Force position of the Setpoint section has not been
used in this research, but its function may be useful in some
cases. The required torque is set by the potentiometer in the
Setpoint section. Like the Z-axis force control, the value of
torque depends on the settings of the Mz strain gage amplifier.
The torque feedback from the Mz strain gages is compared to the
Setpoint value. The speed of rotation is then adjusted to
whatever is needed to produce the desired torque.
The other mode of the Setpoint group is the Position mode.
This mode is used to set oscillatory motion of the 9-axis. The
Offset potentiometer is used to offset the 0° position. In this
case 10 v corresponds to 180° of offset, so a reading of 5 v will
offset the oscillation by 90°. The amplitude and frequency of
oscillation are set in the Oscillation section. Frequency can
be set from 0 to 10 volts (0 to 5 Hz) while the Amplitude
potentiometer can be set from 0 to 10 v (0° to 180°) amplitude.
58
4.4 Temperature Control
The temperature control panel has control of the cartridge
heaters in the chamber walls and can be used to turn on the
recirculator. This panel also can display any of the tribometer
temperatures. The panel, shown in Figure 4.2, has simple
controls and operation. The CPO/Pane1 switch should remain in
Pane1 position so that control of the cartridge heaters is done by
the panel switches. A separate switch allows the Recircu1ator to
be turned On/Off but the controls for setting the recirculator
temperature are housed in the separate NESLAB unit. The display
portion of the, temperature panel reads out in of and is used to
monitor the actual temperatures as well as the setpoint
temperatures. The six-position rotary switch is used to select
which temperature is displayed. The Lower Setpoint is the
desired temperature of the lower half of the chamber, while Lower
Temp. is the actual temperature. The two cartridge heaters in the
lower half are controlled by this panel to achieve the desired
temperature. The Opper Setpoint and Opper Temp. are controlled
in the same way and use the upper cartridge heater to obtain the
desired temperature. The Spind1e Temp. position reads the
temperature of the spindle at a point just before the shaft enters
the chamber. The P1aten Temp. position displays the temperature
of the cup.
The upper and lower cartridge heaters can be controlled by
this temperature panel. Both the upper and lower heaters have the
same type of controls. Either can be turned On or Off with a
toggle switch, and the desired temperature is set by adjusting the
59
appropriate potentiometer until the temperature is read in the
display.
r Temperature
-19.99 UpperS8lpOint~ ~rTamp. Lower Temp. Spindle Temp.
Lower SIIIpOint O. Plalen Temp.
r Upper Healer
@ 01
o Off
CPU 01 o 0 Panel 011
Reclrcullltor
AMTI ...
Figure 4.2 Temperature Control Panel
The recirculator temperature is controlled from the NESLAB
RTE-II0 temperature bath itself. A digital controller, shown in
Figure 4.3, displays the bath temperature in °c and permits
control of the temperature to ±O.l°C. The desired temperature is
set by depressing a Display button and then turning the Coarse
control dial until the desired temperature is displayed. The Fine
control dial can be used to aid in accurate setting of the desired
temperature. Once the operating temperature is set, the Display
button may be released and the LED display will read the actual
bath temperature.
60
1100.O°C 1
o Heat NESLAB oAtr,ess
RTE-110 D aT OJ
OPERATING TEMPERATURE
D Display
0 0 Fine Coarse
Figure 4.3 NESLAB RTE-II0 Recirculator Control Unit
4.5 Force & Torque Measurement
The three forces of interest are Fx , Fy , and F z , and the
torque of interest, in the Z-direction, is Mz . These values are
measured by a unique strain gage transducer that has low inter-
channel cross talk. In order to configure the transducer signals,
four identical strain gage amplifiers are used. They may either
be controlled by the panels, shown in Figure 4.4, or by the
microprocessor. The Panel mode, since it was easier to
understand, was was used in all tests for this research project.
As with the motor control panels, there is a CPO/Panel
switch in the Force panel. In CPO mode, the configuration of the
strain gage amplifiers is accomplished by the microprocessor
connected to a PC, while in Panel mode the switches on the panel
set the configuration. The display reads out the actual signal
from the strain gage in volts. In order to determine the
equivalent engineering units, the strain gages must be properly
configured with the appropriate gain and excitation voltage.
61
Friction Force-Fx '" Friction Force-Fy """
-19.99 -19.99
100 lk 5 10 10@lOk ~@20
.... (Hz) Exclbdlon (V)
1110 lk 5 10 10@10k 2.5@20
FI .... (Hz) Excitation IV)
2k 4k 8k xl-x2
lk@18k @ GUI GUlAdjuel
2k 4k 8k xl-x2
lk@l8k @ Gain Gain Adjuat
Enable Enable
0 0 Auto-ZMo Disable
0 0 Auto-Zaro Disable
CPU 60ft 0 Panel -
CPU 60ft 0 Panel -
ShuntCaJ Shunt Cal
\.. AMTI ... .) AMTI", ~
(a) (b)
r Axial Force-Fz '" I' Moment-Mz ~
-19.99 -19.99
100 lk 5 10
10@lOk 2.5@20
Filtar (Hz) Exclbdlon (V)
100 lk 5 10 10@10k 2.5@20 .0 F ..... (Hz) Excitation (V)
2k 4k 8k xl··x2
lk@16k @ Gain GUlAdjuet
2k 4k 8k xl-x2
lk@l8k @ Gain Gain Adjuat
Enable Enable
0 0 Auto-Zero Disable
0 0 Auto-Zero Disable
CPU 0011 0 Panel -
CPU o Oft 0 Panel -
Shunt Cal Shunt Cal
AMTI ... AMTI ...
( C) (d)
Figure 4.4 Strain Gage Amplifier Control Panels (a) Friction Force Fx (b) Friction Force Fy (c) Axial Load Fz (d) Moment Mz
62
The excitation voltage is set from 2.5 v to 20 v by a four
position rotary switch. It is recommended that the strain gages
be operated at 10 volts excitation voltage. This will maximize
the signal from the transducer and permit the use of the lowest
gain. Although 20 v will not damage the strain gages, the thermal
stability may be compromised. The gain section of the force panel
consists of three elements: a five-position Gain switch, a Gain
Adjust potentiometer, and Enable/Disable toggle switch. The
Gain is set from 1 k to 16 k by a five-position rotary switch.
The gain can then be further adjusted from 1 to 2 times the
reading of the five position switch by enabling the Gain Adjust
potentiometer. When the toggle switch is in the Enable position,
the Gain Adjust potentiometer allows for continuous variation of
gain between adjacent readings on the Gain switch. This allows
the gain to be set to any value from 1 k to 16 k. For instance,
if a gain of 6000 were required, the Gain switch should be set to
4 k, the Gain Adjust should be enabled and set to the point 1.Sx.
This setting will give a gain of 6000 (1.5 x 4 k = 6000). In this
way, the gain can be adjusted for calibrating the voltage output
into equivalent engineering units.
Appropriate engineering units can be set by proper gain
adjustment.· The equation for calculating the appropriate gain to
obtain an equivalent force/volt is given by:
where:
and
Fv Greq Vexc
Equivalent Force/Volt Required Gain Excitation voltage (normally 10 v)
63
(4 . 1)
Cl = Calibration constant for the appropriate axis = 8.78x10-7 l/N for Fx
8.90x10-7 l/N for Fy = 2.25x10-7 l/N for Fz = 1.63x10-S l/(N-m) for Mz ·
So, for example, to obtain a force/volt equivalent of 44.5 N/v for
Fx, the gain should be set to 2557. Now the difficulty is how to
set the Gain Adjust potentiometer so that a gain of 2557 is
obtained. Since a gain of 2557 is between 2 k and 4 k, the Gain
Adjust potentiometer must be used to obtain the required gain.
In order to set a known gain with the Gain Adjust
potentiometer, it is necessary to use the Shunt Cal switch. This
switch places a precision calibration resistor across one arm of
the appropriate Wheatstone bridge in the force transducer. The
+/- switch selects either of two adjacent arms which provides a
positive or negative output. The value of these outputs should
only vary by the sign proceeding the voltage, therefore, it will
not matter which is used to set the gain adjust. For convenience,
the + setting is used in this description. Before switching the
Shunt Cal switch to the + position, the Auto-Zero button should
be pressed. This will automatically balance, or zero, the output
of the transducer signal. To set the Gain Adjust potentiometer,
the output voltage is found from:
Vout Greq x Vexe x 1000
4 x Gswiteh
64
1 J Real Rbridge (4.2)
where: Vout
Greq Vexe Gswiteh
Real Rbridge
=
= = =
= = = =
amplifier output voltage upon switching Shunt Cal. to + position required gain from Eq. (4.1) excitation voltage (normally 10 v) required Gain Switch setting ego if Greq=6000 then Gswiteh = 4 k Greq=2100 then Gswiteh = 2 k 499,000 0 (calibration resistor) resistance of shunted arm of bridge 350 0 for Fx , Fy , and Mz 700 0 for F z •
The procedure for setting the gain adjust is somewhat
confusing and is best described by an example. Assume that a gain
of 2557 is required for Fx , found in the previous section to yield
44.5 N/v equivalent. This would require the Gain switch to be set
to 2 k so Gswiteh=2000. For Fx , Rbridge=35 00, and assuming an
excitation voltage of 10 volts, Eq. (4.2) reduces to:
Vout = (2557) x (10v) x (1000) [ 1 ] (4) x (2000) 1 + 499,0000
3500
Vout = 2.24 volts
To set this output voltage, the Gain switch should be in the 1 k
position, the Gain Adjust should be enabled, excitation voltage
should be set to 10 volts, and the Shunt Cal. switch should be in
the + position. Now by turning the Gain Adjust potentiometer
clockwise until 2.24 v is read in the display, the appropriate
gain adjust factor is set. The Shunt Cal. switch may now be
turned off, and the Gain switch turned to the 2 k position. The
Gain Adjust is always set while the Gain switch is in the 1 k
position because the display can only show Vout values less than
10.0 volts. After the Gain Adjust is set, the Gain switch
should returned to the Gswiteh position. The other strain gage
65
amplifiers are configured by the same method as Fx except that the
appropriate values for Rbridge and Cl should be used.
4.6 Installation of Specimens
Now that the method for configuring the HPT controls for a
test has been presented, the specimens are nearly ready to be
installed. Before the specimens can be installed in the chamber,
they must be thoroughly cleaned. The specimens, specimen holders,
screws, and cup must all be ultrasonically cleaned for 10 minutes
in a solution of mineral spirits. Then they are rinsed clean with
2-propanol to remove any residues. Care should be taken to avoid
touching the specimens since contaminants transferred ,from the
skin can affect the test, especially for tests conducted without
lubricants. The specimens should only be handled with clean
forceps during the installation process. The specimens are
mounted into the specimen holders, as was shown in Section 3.3,
and then mounted into the cup and onto the spindle. The upper
specimen/specimen holder is held directly to the spindle by four
10-32 machine screws, while the lower specimen holder is secured
to the cup by two 10-32 machine screws. The hole pattern in the
cup permits the specimen holder to be mounted in a variety of
orientations, but the most convenient orientation is that in which
the specimen is near one of the sight ports.
The cup can now be assembled with the glass sleeve, and the
sampling hole should be plugged with the threaded dowel. Now the
required oil can be added to the cup. The oil level in the cup
should reach the middle of the sight port to maintain equivalent
amounts of lubricant for all tests. Now the cup is ready to be
66
installed in the chamber. Care must be taken to align the
sampling hole in the cup with the sampling line in the top of the
transducer. The six steel locator pins on the top of the
transducer assure alignment as well as providing resistance to
torsional buckling. The cup is secured to the transducer by
tightening three 10-32 machine screws. Now the threaded dowel can
be removed and the chamber can be closed. Following the procedure
outlined in Section 4.2.1, the chamber should be closed to the
point just before the upper and lower specimens contact each
other. The gap between the specimens should be approximately 3 rom
to 6 rom.
4.7 Purging Procedure
Now that the specimens are in close proximity and the desired
lubricant, if any, has been installed, the system can be taken to
the test temperature. This is accomplished by setting the various
heaters and external units of the HPT. The recirculator should be
set to the desired test temperature, as should the cartridge
heaters. Although the cartridge heaters are primarily used to
prevent condensation of refrigerant, they should also be used when
testing without refrigerant. They help to heat the oil sample and
promote a uniform chamber temperature, thus minimizing thermal
gradients. The cartridge heaters are set by the temperature
control panel described in Section 4.4. The HPT should then be
allowed to come to thermal equilibrium. This should take
approximately 60 minutes, but can vary with the actual temperature
required. After the cup temperature, labeled Platen Temp on the
67
temperature control panel, and the Spindle Temp are within 5° F
of the test temperature, the chamber can be purged.
Before purging, both the charging facility and the drain
facility should be connected to the HPT so that their lines may be
evacuated at· the same time. The sample cylinder does not need to
be attached at this time, but will be evacuated during the
sampling procedure. The Welch and Thermal vacuum pumps are
attached to the HPT as described in Section 2.3.1. The pressure
transducer for the Thermal Vak-Check vacuum gage should also be
connected to the HPT, as shown in Figure 4.5, but valve #6 should
remain closed until after the vacuum pumps are running.
The system is now ready to be purged. First, valves #1, #7,
#8, and #10 should be opened and all other valves should remain
closed. Now the Thermal vacuum pump can be turned on, after which
the Welch pump can be turned on. Valve #6 can now be opened to
permit the vacuum gages to monitor the pressure in the HPT. Now
the valve to the Welch vacuum pump, valve #5, is opened until the
desired vacuum is read on the Thermal gage. With the two pumps
working in tandem, the entire chamber can be evacuated in less
than a couple of minutes.
For tests involving lubricants, the chamber should be
evacuated down to 500 microns; while for tests without lubricants,
the chamber can be purged to below 100 microns. The difference
between these purge levels arises from the fact that the lubricant
begins to boil violently at vacuums better than 500 microns. The
boiling limit, though, does depend upon the type of lubricant
being tested and its temperature.
68
V Iv 1 2 3 4 5 6 7 8 9
10 11
Legend
Drain Valve Drain Tank Valve Sample Port Valve Sample Cylinder Valve Welch Pump Valve Vacuum Gage Valve Throttle Valve Supply Valve Pressure Vessel Valve Filling Valve Supply Tank Valve
-..- To Welch Vacuum
Servo Motor
1-----1f--+TO The rmal .... Vacuum Pump
6
50cc Sample Cylinder
Pressure Chamber
13.6 kg Supply Tank
tnrl ~ 0)
Ul '-DUl .0)
(Y»
Silicone Heating Blanket
6.8 kg Drain Tank
Z-Axis Servo Motor (Load Cell)
Figure 4.5 HPT Mechanical System Schematic
69
The most efficient method of purging and charging the HPT is
to have the charging facility ready to charge the chamber before
initiating the purging procedure. In this way, as soon as the
chamber is evacuated, and the valves are closed, it may be charged
with refrigerant. This alleviates any problems with possible
vacuum leakages.
Once the desired vacuum is reached, all valves should be
closed. After valve #6 has been closed, promptly remove the
Thermal vacuum gage pressure transducer to avoid damaging it. The
Welch vacuum pump can now be vented to atmosphere with the small
venting screw on the valve #5, as shown in Figure 4.6. This must
be done to avoid damaging the pump seals the next time it is
restarted. The Thermal vacuum pump should continue pumping the
chamber until the chamber is charged with refrigerant and the test
begins. The Thermal pump must also be vented before it is turned
off to avoid seal damage. After the chamber is charged, the
Thermal pump should first be disconnected from the HPT and then
turned off.
Venting Screw
Figure 4.6 Venting Screw on Valve to Welch Vacuum Pump
70
4.8 Charging Procedure
Before the HPT can receive a charge of vapor refrigerant, the
3.64 kg pressure vessel must contain a sufficient amount of the
desired refrigerant. Filling the pxessure vessel will need to be
accomplished -every 8 to 10 high pressure tests, as well as any
time a new refrigerant is to be tested. This filling procedure,
therefore, is provided prior to the actual HPT charging procedure.
The refrigerant in the pressure vessel can then be transferred to
the HPT chamber by proper temperature control of the vessel to
generate sufficient pressure.
The pressure vessel should be filled while the entire system
is being purged. After the vacuum pumps have evacuated the system
to at least 500 microns, valves #7, #8, #9, #10, and #11 are
closed. The pressure vessel should now be placed in an ice bath
to aid in transferring liquid refrigerant from the 13.6 kg supply
tank to the pressure vessel. The supply tank should also be
placed, inverted, above the pressure vessel. This will supply
liquid refrigerant to the vessel. The supply tank should be
placed on a scale to monitor the weight of refrigerant transferred
to the vessel. Now valves #9, #10, and #11 may be opened until
the desired amount of refrigerant has been transferred. In
general, it takes approximately 5 to 10 minutes to fill the
pressure vessel, but this depends on ambient temperature of the
supply tank. After the pressure vessel is full, valve #9 should
be closed, the pressure vessel should be removed from the ice
bath, and the supply tank should be placed upright on the floor.
This will allow the liquid refrigerant trapped in the lines to
drain back into the supply tank. This draining should take less
71
than one minute. After this draining, valves #10 and #11 may be
closed. Now the pressure vessel can be used to charge the HPT.
Assuming that the chamber has just been purged and valve #5
has just been closed, the chamber is now ready to be charged with
refrigerant vapor. Depending on what test pressure is required,
the pressure vessel may need to be heated. For pressures below
0.414 MPa, the pressure vessel does not need to be heated, so the
refrigerant can be directly transferred to the HPT. By opening
valves #8 and #9, and then using valve #7 to throttle the flow of
refrigerant, the desired pressure can easily be attained.
For pressures greater than 0.414 MPa, however, the pressure
vessel must be heated. The heat is provided by the heating
blanket wrapped around the pressure vessel. A thermometer is
placed between the blanket and vessel to monitor the temperature.
The temperature that will generate the required pressure can be
determined from the appropriate refrigerant vapor pressure curve
shown in Figure 4.7. For example, Figure 4.7a shows that in order
to generate a vapor pressure of 1.5 MPa for HFC-134a, a
temper~ture of 58 °c would be required. The temperature of the
refrigerant, therefore would be raised to this temperature. It
should be noted that the temperature measured by the thermometer
is not the temperature of the actual refrigerant but that of the
outside of the pressure vessel. The actual internal temperature
of the refrigerant is generally 50 C lower due to convection,
radiation, and other heat losses from the pressure vessel. It is
therefore necessary to set the variac to generate a temperature of
approximately 630 C to obtain HFC-134a vapor at 1.5 MPa, for this
example.
72
2.0,-------------------------------------------n
cO 1.5
~ (l) 1-1 ::s [/) 1.0 [/) (l) 1-1 ~
(l) 0'1 cO t!J
0.5
O.O;,"TT~~~~TT~~MM~TT~~~~TTTr~~~TTi
-30 -20 -10 o 10 20 30 40 50 60 70
Temperature (oC)
Figure 4.7a HFC-134a Vapor Pressure vs Temperature
2.0,-------------------------------------------~
1.5 ttl 0.. ~
<Ii ~ ::l fIl 1.0 fIl <Ii ~ 0..
<Ii 01 ttl ~
0.5
-30 -20 -10 o 10 20 30 40 50 60 70
Temperature (oC)
Figure 4.7b CFC-12 Vapor Pressure vs Temperature
73
The variac should initially be set to approximately 50% of
its maximum voltage output and then adjusted to obtain the desired
pressure vessel temperature. At 50%, the variac will cause the
pressure vessel to reach a temperature of 77° C and a
corresponding pressure of over 2.0 MPa for either CFC-12 or HFC-
134a, so the pressure and variac setting should be closely
monitored to avoid over pressure conditions. Under no
circumstances should the temperature of the pressure vessel be
allowed to exceed 85° C. If a temperature of 85° C is not capa-ble
of generating the required test pressure, the charging process
should be discontinued. Either there is not enough refrigerant in
the pressure vessel to permit charging, in which case the vessel
will need to be refilled, or the refrigerant is condensing on the
chamber walls, which could be caused by a malfunction or incorrect
setting of the cartridge heaters. After the pressure vessel
reaches the desired pressure, or is slightly above it, the chamber
is ready to be charged.
Before continuing with the charging process, valves #5 and #6
must be closed. Charging is accomplished by opening valves #7,
#8, and #9 and allowing the chamber to reach the desired test
pressure. Once this pressure is reached, valve #7 should be
closed and the chamber should be kept at the test pressure for one
hour to allow the oil-refrigerant mixture to reach an equilibrium.
One hour has proven to be a sufficient amount of time to allow the
refrigerant to saturate into the oil [16]. Valve #7 may have to
periodically opened to maintain a constant refrigerant pressure in
the chamber as the refrigerant saturates into the oil. After one
hour, the variac should be turned off and the pressure vessel
74
should be allowed to return to ambient temperature before valves
.8 and .9 are closed. The HPT is now ready to be configured for
controlling and conducting the tests.
4.9 Configuring the HPT Controls and Conducting the Test
Now that the HPT chamber is·at thermal and pressure
equilibrium, the controls must be configured to run the test. The
strain gage amplifiers must be individually configured for each
test. Since the axial force is the applied test load, the F z
strain gage amplifier configuration is determined by the magnitude
of the applied test load. In order to achieve high accuracy, the
equivalent test load, in volts, should be between 7.0 v and
10.0 v; but for convenience in monitoring loads, it is desirable
to keep a simple force/volt equivalent. Suppose an axial applied
load of 561 N is required for a certain test, a convenient
force/volt equivalent might be 60 N/v. This would mean that the
applied voltage would be 9.35 v, which will satisfy the accuracy
condition as well as providing a convenient conversion from volts
to Newtons. From Eq. (4.1), the required gain to have 60 N/v is
Greq=7416. This gain can be obtained by using Eq. (4.2) and the
procedure of Section 4.5 to set the Gain Adjust potentiometer to
yield Vout=6.49 v.
The two friction forces, Fx and Fy , are set by a similar
method. As a general rule, the coefficient of friction will be
less than 0.20, so the magnitude of the friction forces will be
20% of the applied load. This requires that Fx and Fy be set to a
lower force/volt equivalent than the applied load Fz . For example,
if F z is set to 60 N/v, both Fx and Fy should be set to 12 N/v.
75
Then, from Eq. (4.1), the required gains are 9484 for Fx and 9364
for Fy . These gains correspond to output voltages of Vout=2.08 v
and Vout=2.05 v for Fx and Fy , respectively. These settings will
provide accuracy of the actual reading and convenience in
calculating the force equivalent of the Fx and Fy voltages
displayed on the control panels.
Similarly, the moment Mz is configured in the same manner.
Since, for this example, the maximum frictional force resultant
will be when both Fx and Fy are 10 v (120 N) the tangential force
will be Ft=(1.41) (120)=170 N. Assuming that the average moment arm
of the contact is 19 rom, an approximate maximum moment of
3.238 N-m is possible. By setting Mz to 0.4 N-m/volt, the maximum
measurable moment will be 4.0 N-m (10 v) which is consistent with
the other force measurements. To obtain an Mz setting of
0.4 N-m/v, Eq. (4.1) requires a gain of 15337. This required gain
corresponds to an output voltage of Vout=3.36 v for setting the
Gain Adjust. The final step in configuring the amplifiers is to
use the Auto-Zero button to balance the strain gages before
initiating the test. Now that the strain gage amplifiers have
been configured, both the axial load and rotational motion need to
be set.
The axial force is set according to the procedures outlined
in Section 4.2. To apply a constant axial force, the Z-axis
control panel should be set to the following positions. The
amplifier should be turned on and warmed up at least one hour. In
the Setpoint section, the toggle switch should be in the Force
position and the direction switch should be in the Up position.
Before the entire load is applied, a smaller load of 0.5 v or less
76
should be applied with the Force potentiometer. This smaller load
is to ensure the contact specimens align properly. The Z-axis
Motor switch may now be enabled, allowing the small force to be
applied.
Next, the 8-axis needs to be properly configured as described
in Section 4.3. For unidirectional rotation, the direction is set
with the CW/CCW switch, the Setpoint toggle switch should be in
the Speed position, and the Rate potentiometer is adjusted to
yield the appropriate speed. If oscillatory motion is required,
the procedure for setting the frequency and amplitude in
Section 4.3 should be followed.
Once the 8-axis is configured, the data acquisition system
must be initialized and properly configured to sample the
appropriate data. The interface between the PC and the
microprocessor is initialized by choosing the TMBMI icon from the
PC Windows environment. The interface program then initializes a
connection with the microprocessor and awaits the command to begin
sampling data. Since each of the contact types require slightly
different data to be sampled, each configuration has been given a
separate command. The commands are RLPIST, SPLATE, and RCPIST
for the counterformal, area, and conformal contacts respectively.
By entering the appropriate command at the cursor, the data
acquisition system is automatically configured to sample data for
the particular contact. To begin sampling data and saving them to
disk, the following commands, for the appropriate geometry, should
be typed but not yet entered.
For counterformal contact, type: ISAVE 15 285 "filename"
77
For area contact, type: SAVED "filename" 3600
For the conformal contact, type: ISAVE 10 285 "filename"
where filename is the name of the file to be saved to disk.
Now that the HPT controls have been properly configured, the
test may be started. First, the rotational axis motor should be
enabled. This will cause the 8-axis dc motor to begin its motion,
whether unidirectional or oscillatory. Now the full test load
should be applied. Since the Z-axis motor is still enabled, but
only a small load is applied, the Force potentiometer should be
increased until the desired load is applied. In order to ensure
repeatability of start-up conditions, the load should be increased
at a constant rate over one minute up to the test load. In order
to begin data acquisition, the appropriate save command should now
be entered from the keyboard.
The tests are run for one hour unless seizure or other
problems occur. The status of the contact should be observed
periodically. This is best accomplished by using a flashlight,
and lqoking in one of the viewports in the chamber. The presence
of wear particles, oil cloudiness, or other relevant physical
phenomena should be noted along with the time into the test. The
HPT should not be left unmonitored while a test is underway.
The procedure to stop a test is basically the reverse of
starting the test. The load should be removed and the contacts
should be separated by about 3 mm to 6 mm. The 8-axis and z-axis
motors should then be disabled, but the chamber should remain
closed to allow the oil-refrigerant mixture to be sampled, and the
remaining refrigerant to be reclaimed.
78
4.10 Sampling
Before an oil-refrigerant sample can be drawn off from the
chamber, the sample cylinder should be cleaned and purged. The
cylinder and fittings should be cleaned with mineral spirits and
then rinsed with 2-proponal to remove any residues. The Thermal
vacuum pump should then be connected to the sample cylinder, and
the cylinder purged for 5 minutes. This amount of purging has
been found to be sufficient to remove nearly all atmosphere from
the cylinder and leave a vacuum of less than 100 microns. The
valve of the sample cylinder should then be closed, it should be
disconnected from the vacuum pump, and then the sample cylinder
should be weighed on the Sartorius balance. This weight, Wl, of
the empty sample cylinder will be used to determine the weight
percent of refrigerant in the oil.
The sample cylinder should now be connected to the chamber
using the quick connect on valve #3 of Figure 4.5. Both valves #3
and #4 should now be opened so that the oil-refrigerant mixture is
allowed to flow into the sample cylinder. The time required to
fill the sample cylinder will depend on the viscosity of the
mixture and the pressure in the chamber, but in general, 5 minutes
should be sufficient. Now both valves #3 and #4 should be closed
and the sample cylinder should be removed from the chamber at the
quick-connect. If any oil-refrigerant mixture is spilled on the
outside of the cylinder, the surface should immediately be cleaned
with mineral spirits, rinsed with 2-propanol, and then dried. Now
the full sample cylinder, with the oil-refrigerant mixture, should
be weighed on the balance. This weight, ws, is used in determining
the total sample weight.
79
3.05 m x 0.63 rom 0 Capillary Tube
Sample Cylinder Assembly
1/4" Plug Valve (Nupro: B-4P4T)
Figure 4.8 Sample Cylinder Refrigerant Evacuation Facility
The sample cylinder should now be connected, as shown in
Figure 4.8, to a 3.05 m by 0.63 rom 0 capillary tube assembly.
Initially, valve #12 should remain open to the atmosphere. Valve
#4 is then opened until outward flow has stopped. This may be
monitored by observing the gas flowing out of the capillary tube.
This process may take approximately two hours, but time will
depend on the amount of refrigerant saturated into the oil. To
remove any remaining refrigerant trapped in the oil, valve #12
should be connected to the Thermal vacuum pump, and the cylinder
purged for at least one hour. After one hour, valve #4 should be
closed and the capillary tube assembly removed from the sample
cylinder at the quick connect. The final weight, W3, of the sample
cylinder is now measured on the balance. This weight is the
80
weight of the cylinder and remaining oil. The resulting weight
percent of refrigerant in the oil is found from:
where Wt%= WI W3 W5
Wt% (4.3)
Weight percent of refrigerant in oil Empty weight of sample cylinder Final weight of sample cylinder and oil Full weight of sample cylinder with oilrefrigerant mixture
4.11 Refrigerant Reclamation
Refrigerant reclamation is necessary to avoid contributing to
the release of ozone depleting compounds, as well as to reclaim
the expensive refrigerants so that they may be recycled. The
procedure for reclaiming the refrigerant from the chamber is
efficient yet simple and inexpensive. An empty 6.8 kg DOT-39 NRC
260/325 MI073 E.!. du Pont de Nemour & Co container is used as the
drain tank. Figure 4.9 shows this container connected to the HPT
chamber at the drain valve.
To begin refrigerant reclamation, the 6.8 kg drain tank is
placed in an ice bath and valves #1 and #2 are opened. The cold
refrigerant in the drain tank is at a lower pressure than the
chamber so there is a pressure driving force to transfer
refrigerant to the drain tank. The weight of the 6.8 kg container
should be periodically monitored to ensure that it is not being
overfilled which could cause rupture. Equilibrium will be reached
after about one hour so that valves #1 and #2 can be closed. The
chamber can now be opened so that the oil and specimens can be
removed. There will still be a slight amount of refrigerant in
the chamber. Valve #5 can be opened while the chamber is opening
81
to permit any residual refrigerant to escape. The cartridge
heaters should now be turned off and the recirculator should be
set back to ambient temperature. Once the chamber has cooled
sufficiently, the specimens can be removed and wear analysis can
begin.
HPT Pressure Chamber
<D 1/4" Nupro Plug Valve (SS-4P4T-TB)
1/4" 0 Nupro Hose SS-7R4TA4TA4-24
DuPont 6.8 kg Drain Tank
Figure 4.9 Refrigerant Reclamation Facility
82
CHAPTER 5
RESULTS & DISCUSSION
5.1 Measurement of Wear
Wear results are based on measurements taken immediately
after completion of each test except for the area contact. While
all specimens are ultrasonically cleaned in mineral spirits after
testing and then immediately measured for wear, the bronze shoe of
the area contact is cleaned and then allowed to dry in a desiccant
chamber to remove any moisture. This is only done for the shoe
because its wear is measured as a weight loss which would be
affected by moisture. Wear on the other contacts is measured by
methods not sensitive to moisture so they do not need to be dried.
Each test was repeated at least once to compute the average
friction and wear.
The reported wear for the counterformal contact is only that
measured on the tool steel pin. The surface of the mating piece,
the gray cast iron plate, showed only minor polishing wear. This
wear probably resulted in slight changes in the surface finish,
but no measurements were taken. A typical wear scar on the tool
steel pin is shown in Figure 5.1. The wear scar width is measured
with a Nikon SMZ-2T stereoscopic microscope. This method permits
the width of the scar on the pin to be measured to an accuracy of
0.005 mm. The average wear scar width was determined by taking
three measurements of the wear scar along the length of the pin.
The volumetric wear may be calculated by:
v ~ s r - w( r - 0 )]L 2
83
(5.1 )
where V volumetric wear (mm3 )
wear scar width (mm) r = radius of pin (mm) w
L length of pin (mm) and
s=2r</> where </>= asin(2:) and
Figure 5.1 Typical Wear Scar on Pin of Counterformal Contact. This particular contact, 31RL, was run in HFC-134a, by itself, at 1.55 MFa. (25x)
The reported area contact wear is computed as a weight loss
on the bronze shoe. Figure 5.2 shows a typical surface of the
bronze shoe after testing. The weight of the shoe, Wi, is recorded
prior to testing. After the test, the shoe is cleaned and dried
before re-weighing it. This final Height, W"f, is the Height after
testing. The difference betHeen these tHO Heights is the total
84
weight loss due to wear. All of these weights are measured on a
Mettler model AE 163 precision balance, with a readability of
0.01 mg. Because the mating piece, the ductile cast iron plate,
is very much harder than the bronze, no wear was observed on the
plate. Only small amounts of bronze are transferred to the
surface of the plate during testing.
Figure 5.2 Typical Worn Surface of Bronze Shoe of Area Contact. This contact, 30SP, was run with PAGI-F plus HFC-~34a. (25x)
The wear reported for the conformal contact is that measured
only on the aluminum pad. Figure 5.3 shows a magnified view of a
typical wear scar on the pad. The entire wear scar on the surface
of the aluminum pad is similar to that in Figure 5.4. Using a
Talysurf 10 surface profiler, the depth of the wear scar was
accurately measured to within 0.0025 mm. The scar depth was
measured at three separate locations on the pad and then an
85
average wear scar depth was computed. The mating piece, the casc-
hardened steel pin, showed no wear.
Figure 5.3 Typical Wear Scar on Aluminum Pad of Conformal Contact. This contact, 25RC, was run with Est3-F plus HFC-134a. (25x)
Talysurf Traces
Wear Scar
Figure 5.4 Wear Scar on Aluminum Pad of Conformal Contact
5.2 Surface Analysis
In order to determine critical surface "properties, both
before and after testing, two surface analysis tools were used.
The Talysurf 10 was used to determine surface profiles as well as
86
roughness average (Ra) values for all samples tested. As
previously mentioned, the Talysurf 10 was also used to measure
wear scar depths on the aluminum pad for the conformal contact.
The second surface analysis tool was X-ray Photoelectron
Spectroscopy (XPS); which was used to determine the existence, if
any, of surface films formed during testing.
A Perkin-Elmer/PHI 5400 X-ray Photoelectron Spectrometer was
used in the small spot XPS studies. A regular MG Ka x-ray source
is used to bombard the surface of the specimen with x-rays which
dislodge photoelectrons from the surface. These photoelectrons
are then analyzed by a spherical capacitor analyzer with a
computer data acquisition system. This analysis can be used to
show the formation of surface films, if any, such as metallic
chlorides or metallic oxides, that formed on the specimen surfaces
during testing.
5.3 Counterformal Contact Results
Wear data for the counterformal contact are shown in
'Figure 5.5. The formulated versions of the three lubricants
(Alkbenz-F, Est1-F, and PAG2-F), by themselves, provide better
wear resistance than their base counterparts. This can be
attributed to the additive packages that are included in the
formulated oils to improve wear. Once refrigerant is added to the
lubricants, however, the wear results are less predictable.
When CFC-12 is added to the alkylbenzenes, the wear results
depend on the formulation of the oil. When CFC-12 is added to the
base alkylbenzene, wear decreased. XPS analysis of the tool steel
pin showed formation of iron chloride films (FeC12) on the surface.
87
The hypothesis is that these FeC12 films essentially act as EP
agents in the oil thus reducing wear. A comparable increase in
wear resistance for the formulated alkylbenzene was not observed
when it was tested with CFC-12. This is most probably due to the
additives which already exist in the oil.
0.004
o.ooo~~-
Ref Base Form Base + Ref
Form + Ref
Figure 5.5 Wear Data For Counterformal Co.ntact
While CFC-12, by itself, provides very good wear resistance,
testing in a HFC-134a environment shows extremely high wear rates,
equivalent to testing in air. XPS analysis of the tool steel pins
tested in HFC-134a showed that no surface films were produced.
The addition of HFC-134a to the Estl and PAG2 lubricants also show
no improvements to wear resistance. In fact, testing of both the
esters and PAGs with HFC-134a t~nds to increase wear over the
respective lubricant alone. This is due to the fact that the
88
addition of HFC-134a to the lubricant decreases the effective
viscosity. This lower viscosity decreases the probability of
generating protective oil films between the surfaces, thus
producing higher wear. The raw data, given in Appendix B, shows a
maximum scatter between repeated tests of 8.7%. This suggests
that the repeatability of experimental data is quite good.
s:: 0
• .,f .jJ u
• .,f 1-1 r.. Ii-! 0
.jJ s:: CI)
• .,f u
• .,f Ii-! Ii-! CI) 0 u CI) 01 cu 1-1 CI)
~
0.15
0.10
0.05
Ref Base Form Base + Ref
Form + Ref
Figure 5.6 Friction Data For Counterformal Contact
While wear data are quite dependent on the type of lubricant
and type of refrigerant tested, the friction results, shown in
Figure 5.6, do not generally follow these tendencies. Overall,
only slight variations in coefficient of friction are observed for
all tests conducted. The coefficient of friction for HFC-134a, by
itself, does show the highest value of any of the tests. This
correlates well with the wear results but other general trends are
89
difficult to discern. The only other tendency of the coefficient
of friction results which parallels the wear results, with the
exception of Est1-B, is that the tests involving HFC-134a with a
lubricant, show slightly higher coefficients of friction than the
corresponding tests without HFC-134a. The repeatability of the
coefficient of friction for all tests is within 16%.
5.4 Area Contact Results
The wear results from area contact tests are shown
graphically in Figure 5.7. When the base oils are used alone, the
mineral oil seems to provide the best wear resistance. Although
the two formulated oils by themselves give better wear
characteristics than the base mineral oil, once refrigerants are
added to the lubricants, the mineral oil-CFC-12 mixture provides
the best wear resistance. The presence of CFC-12 promotes
chemical reaction on the bronze shoe, producing copper chlorides
(CUCI2) and traces of zinc fluoride (ZnF2) surface films. As with
the counterformal contact, the surface films help to protect the
surface and therefore lower wear. Although not verified, the R12
most likely formed FeCl2 surface films on the mating ductile cast
iron disk as well. The wear resistance of base PAG and ester oils
in a HFC-134a environment is much lower than that for mineral oil
in CFC-12. From the limited number of lubricants tested, it is
seen that the formulated ester with refrigerant (Est2-F+R134a)
provides wear characteristics similar to the mineral oil-CFC-12
mixture.
As with the counterformal wear results, HFC-134a alone lacks
lubricative properties. Though not shown, very high wear rates
90
and temperature rises, equivalent to testing in air, are observed
for tests run with HFC-134a by itself. When HFC-12 was tested
alone, it also showed higher wear rates than those of the base
oils, but not nearly as high as the HFC-134a. Similar to the
counterformal contact results, the addition of HFC-134a to both
the PAGs and esters tends to increase wear. The raw data, given
in Appendix B, show a maximum scatter between repeated tests of
13.5%.
Base Form Base + Ref
Form + Ref
Figure 5.7 Wear Data for Area Contact
The friction data for the area contact are shown in
Figure 5.8. In general, these data correlate reasonably well with
the wear data. As with the wear, the friction obtained with the
formulated ester and HFC-134a compares favorably with the
presently used mineral oil-CFC-12 mixture. Also the formulated
91
versions of the PAG and ester provide lower coefficients of
friction than their base counterparts. The maximum scatter for
the coefficient of friction between repeated tests was 22%.
s:: 0
• .-1 .j.J 0
• .-1 1-1 r.. "-I 0
.j.J s:: Q)
• .-1 0
• .-1 "-I "-I Q) 0 u Q) 0'1 lIS 1-1 Q)
~
0.08
0.06
0.04
0.02
Base Form Base + Ref
Form + Ref
Figure 5.8 Friction Data for Area Contact
5.5 Conformal Contact Results
Wear data for the conformal contact are shown in Figure 5.9.
As with the area contact results, the mineral oil provides the
best wear resistance for this contact. The addition of CFC-12 to
the base mineral oil improves wear resistance. XPS analysis shows
the formation of FeCl2 films on the steel pin. This film acts like
those formed in the counterformal contact by protecting the
surface against wear. The addition of HFC-134a to the PAG2-B
tends to increase the amount of wear. Although not graphically
shown, when HFC-134a was tested by itself, seizure occurred after
92
15 minutes of testing. The difference in the wear results for the
two esters is probably due to the fact that the viscosity of Est3
is much less than that of Est2. Thinner oil films are generated
for low viscosity oils than for high viscosity oils. This allows
closer proximity of surfaces and therefore more asperity
interaction and wear. The scatter between repeated tests, from
raw wear data given in Appendix B, show a maximum of 11.5%.
i'O ~1\1
p., .c +J S o..::s Q) t: Q .g 1-1 ::s 1\1~ u ..:e U)
t: 1-1 0 1\1 Q)'O :l: Q)
1-1 Q) ::s OIUl 1\1 1\1 1-1 Q) Q) ~
~
0.25
0.20
0.15
0.10
0.05
Base Form Base + Ref
Form + Ref
Figure 5.9 Wear Data for Conformal Contact
Figure 5.10 shows the coefficient of friction data for the
conformal contact. Even though the mineral oil-CFC-12 mixture
shows very good wear characteristics compared to the esters and
PAG with HFC-134a, its friction characteristics tend to be worse
than those of the latter mixtures. Overall, the PAG2-F gives the
93
() I-' 0 0 CD ~ HI CD HI CJ) .... rt () .... ()
Average Coefficient of Friction CD 0 ~ (1)
hj rt HI .... 0 0 0 a a HI \Q a a ...... ...... '" 0 .... s:: 0 lJ1 a lJ1 a HI () Ii .... CD HI CD
Ii ~ U1 .... rt
tIl () I-' III Est2-B rt 0 0 CII .... HI
Ct> Est3-B 0 hj
PAG2-B ~ HI Ii Ii .... 0' .... () CD () rt rt rt .... ~ .... 0 CD 0 ~ t'rJ CD ~
0 ~ t:l
~ Est3-Pl Ii '" rt CD 1-3 "'" Pl PAG2- "0 ::T
CD CD HI Pl 0 rt a Ii CD Pl
tIl 0- X (") Est2-B+R134a .... 0 :x:J III rt e ~ Ct> CII
Est3-B+R134a CD HI
HlCt> CJ) a
0 + PAG2-B+R134a rt Ii CJ) CJ)
a () Pl .... Pl I-' CJ) rt
rt (")
t'rJ I-' CD 0 :x:J 0 Est2-F+R134a w Ii ~ Ct> 11 d(J
rt HlS Est3-F+R134a HI Pl 0 () + PAG2-F+R134a Ii rt
rt ::T CD
6.1 Research Summary
CHAPTER 6
CONCLUSIONS
An experimental facility has been developed to test the
,friction and wear characteristics of oil-refrigerant mixtures. A
high pressure tribometer has been designed to permit specimen
testing in pressurized refrigerant environment at temperatures
that can vary from -300 C to 1500 C and pressures of up to 1.724
MFa. The HPT has been outfitted with various apparatuses for
charging, purging, and sampling as well as permitting data
acquisition by a personal computer. In addition, procedures and
standards were developed to ensure repeatability of the test
results.
The critical contacts were found to be the counterformal,
area, and conformal contacts in the rolling piston, swash plate,
and reciprocating piston compressors, respectively. These
contacts along with their approximate operating conditions were
tested in the HPT. Two refrigerants and various mineral and
synthetic lubricants were evaluated in the HPT to determine
lubricative properties of the mixtures.
Baseline data for mineral oils and alkylbenzenes with CFC-12
have been completed. The results of these tests have shown that,
for all three contact types, CFC-12 has excellent lubricative
properties. This includes testing CFC-12 by itself as well as
with various lubricants. XPS analysis shows that metallic
chlorides are formed on the surfaces of the contact specimens.
95
These chloride films aid the lubrication process by protecting the
surfaces and, therefore, tend to decrease wear.
Comparative tests for HFC-134a and various esters and PAGs
have also been completed for the same contacts and conditions as
the baseline tests. In general, these tests show worse wear
results than the baseline tests,- especially when the refrigerant
is tested alone. XPS analysis shows no surface films produced on
the surfaces of the contact specimens, so there is no equivalent
reduction of wear. When HFC-134a is saturated into the
lubricants, the effective viscosity of the mixture is decreased
and, therefore, thinner oil films are generated and wear
increases. Although the addition of CFC-12 to the lubricant also
lowers viscosity, the formation of surface films counteracts this
effect.
The data for the coefficient of friction do not generally
correspond to their wear results. The coefficient of friction for
the counterformal contact shows only small variations from test to
test. The area contact results show somewhat greater variations
and do seem to correlate with some wear results. The friction and
wear results for the conformal contact are generally not similar.
Although the mineral oil-CFC-12 shows the best wear resistance,
its friction is among the worst. These observations suggest that
there is no clear relationship between friction and wear, and each
property must be evaluated separately.
6.2 Recommendations for Future Research
Although this research has provided excellent baseline data
for friction and wear in oil-refrigerant environments, some
96
improvements can be made. The installation of a high speed data
acquisition board would greatly enhance the quality of sampled
data during oscillatory testing. This board, along with
appropriate control software, could be used to control the dc
servo motors so that considerably more complicated loads and
motions could be tested. In addition, an in-line viscometer
should be added to the system to permit the viscosity of the oil
refrigerant mixture to be measured. This would allow for a more
quantitative analysis of the thinning effect of refrigerant on oil
lubricative properties.
Future testing should focus on three possible avenues:
material development, surface roughness effects, and lubricant
screening. Material development includes both determining new
material pairs to be tested as well as investigating surface
coatings. The surface roughness of the contacts should be varied
to determine possible effects on friction and wear. Finally,
additional lubricants should be used to determine if they provide
better lubrication. In order for the lubricant screening to be
effective, however, lubricant manufacturers will have to divulge
additives used as well as provide more information about the
lubricant structure.
97
REFERENCES
[1] Little, J.L., "Viscosity of Lubricating Oil-Freon-22 Mixture," Refrigerating Engineering, Nov. 1952. p. 1191.
[2] Parmelle, H.M., "Viscosity of Refrigerant-Oil Mixture at Evaporator Conditions," ASHRAE Trans., Vol. 70, 1964, p. 173.
[3] Spauschus, H.O., "Vapor Pressure, Volume, and Miscibility Limits of R-22-0il Solution," ASHRAE Trans., Vol 70, 1964, p. 306.
[4] Spauschus, H.O., and L.M. Speaker, "A Review of Viscosity Data for Oil-Refrigerant Solutions," ASHRAE Trans., Vol. 93, No.2, 1987, p. 667.
[5] Pate, M.B., Van Gaalen, N.A. and Zoz, S.C., "The Measurement of Solubility and Viscosity of Oil/Refrigerant Mixtures at High Pressures and Temperatures: Test Facility and Initial Results for R-22/Naphthenic Oil Mixtures," ASHRAE Trans., Vol. 96, No.2, 1990, P 183.
[6] Pate, M.B., Van Gaalen, N.A. and Zoz, S.C., "The Solubility and Viscosity of HCFC-22 in Naphthenic Oil and in Alkylbenzene at High Pressures and Temperatures," ASHRAE Trans., Vol. 97, No.1, 1991, P 100.
[7] Pate, M.B., Van Gaalen, N.A. and Zoz, S.C., "The Solubility and Viscosity of Solutions of R-502 in Naphthenic Oil and in Alkylbenzene at High Pressures and Temperatures," ASHRAE Trans., Vol. 97, No.2, 1991, p 179.
[8] Grebner, Jeffery J., The Effects of Oil On the Thermodynamic Properties of Dichlorodifluoromethane (R-12) and Tetrafluoroethane (R-134a), M.S. Thesis University of Illinois, Jan. 1992.
[9] Thomas, R.H.P., Wu W-T, and Pham, H.T., "Solubility & Viscosity of R-134a Refrigerant/Lubricant Mixtures," ASHRAE Journal, Feb. 1991, p. 37.
[10] ASHRAE. 1990. ASHRAE Handbook -- Refrigeration, ASHRAE, 1990, p. 8.1-8.21.
[11] Huttenlocher, D.F., "Bench Scale Test Procedure for Hermetic Compressor Lubricants," ASHRAE Journal, June 1969, p. 85.
[12] Sundaresan, Sonny G., "Status Report on Polyalkylene Glycol Lubricants For Use With HFC-134a in Refrigeration Compressors," Procedings of the 1990 USNC/IIR-Purdue Refrigeration Conference, July 17-20 1990, p. 138.
98
[13] Sundaresan, S.G., and W.R. Finkenstadt, "Polyalkylene Glycol and Polyolester Lubricant Candidates for Use With HFC-134a in Refrigeration Compressors," ASHRAE Trans., 1992, Vol. 98. No 1, In Print.
[14] Sanvordenkerr, K.S., "Lubrication by Oil-Refrigerant Mixtures: Behavior in the Falex Tester," ASHRAE Trans., Vol. 90, No. 2B, 1984.
[15] Komatsuzaki, S., T. Tomobe, and Y. Homma, "Additive Effects on Lubricity and Thermal Stability of Refrigerant Oils," Lubrication Engineering, Journal of ASLE, Vol 43, No.1, Jan. 1987, p. 31.
[16] Komatsuzaki, S. and Y. Homma, "Antiseizure and Antiwear Properties of Lubricating Oils under Refrigerant Gas Environments," Lubrication Engineering, Journal of STLE, Vol. 47, No.3, 1991, p. 193.
[17] Komatsuzaki, S, Y. Homma, K. Kawashima, and Y. Itoh, "Polyalkylene Glycol as Lubricant for HFC-134a Compressors," Lubrication Engineering, Journal of STLE, Vol. 47, No. 12, 1991, p. 1018.
[18] Kosberg, Robert C., Jeffrey M. Stokes, and Margaret A. Schneller, Mathematical Modeling and Dynamic Analysis of The Rolling Piston Rotary Vane Compressor, Engineering Design Program, Dept. of General Engineering, University of Illinois, 1988.
[19] Young, Warren C., Roark's Formulas for Stress and Strain, New York: McGraw-Hill Book Company, 1989.
[20] ASHRAE, 1984. ANSI/ASHRAE 41.4-1984, Standard Procedure for experimentally determining the weight concentration of oil in single-phase solutions of oil in refrigerant. Atlanta: American Society of Heating, Refrigerating, and Air~onditioning Engineers Inc.
99
APPENDIX A - Equivalent Cylinders
The distance between two rigid cylinders can be shown to be
equivalent to the distance between an equivalent rigid cylinder
and a flat plate. Figure A.l shows two cylinders, separated by a
gap height of ho at their closest point. An equivalent cylinder
separated from the flat plate is shown in Figure A.2. The
expression for the gap height, h, between two cylinders may be
written as:
(A. 1)
Using the Maclauren series expansion, cos ~ can be expressed by:
cos (A. 2)
If ~ is small, then substituting with x/R and eliminating all
higher order terms, the following can be obtained.
cos ~ = 1 - ~ 2R2
Substituting Eq. (A.3) into Eq. (A.l) yields:
(A.3)
If the final two terms are combined with a common denominator, the
expression for h reduces to:
h (A. 4)
100
The radius of the equivalent cylinder on a flat plate, for
counterformal contacts, is found to be:
R =
and for conformal contacts, the radius of the equivalent cylinder
on a flat plate is:
R =
-A
1 -I C
I 1 .x -- D
B
~ ____ ~ ____ ~~ x
(a) (b)
Figure A.l (a) Counterformal Contact Geometry (b) Conformal Contact Geometry
101
.-x
Figure A.2 Equivalent Cylinder on Flat Plate
102
APPENDIX B - Raw Data
Table B.1 Counterformal Contact Wear Data
Wear Scar Coefficient lest i Qils Befl:igel:aDt Ile~th (mm} Qf El:i!::tiQD
09RL Alkylbenz-F 0.165 0.090 10RL Alkylbenz-B 0.220 0.094 11RL Alkylbenz-B R12 0.140 0.094 12RL Alkylbenz-F R12 0.195 0.100 14RL Alkylbenz-F R12 0.lS5 0.101 15RL Alkylbenz-B R12 0.140 0.104 lSRL R12 0.195 0.105 19RL Alkylbenz-F 0.170 0.102 20RL Alkylbenz-B 0.230 0.104 22RL R12 0.200 0.107 23RL Est1-B 0.265 0.OS5 24RL Est1-F 0.165 0.102 27RL Est1-B 0.295 0.101 2SRL Est1-F 0.165 0.10S 30RL R134a 0.470 0.132 31RL R134a 0.505 0.12S 32RL Est1-B R134a 0.260 0.105 33RL Est1-B R134a 0.270 0.104 34RL Est1-F R134a 0.255 0.120 35RL Est1-F R134a 0.270 0.120 3SRL PAG2-B 0.235 0.076 39RL PAG2-B 0.235 O.OSO 40RL PAG2-B R134a 0.255 0.OS7 41RL PAG2-B R134a 0.250 0.OS7 42RL PAG2-F 0.140 0.102 43RL PAG2-F 0.150 0.101 44RL PAG2-F R134a 0.220 0.097 45RL PAG2-F R134a 0.240 0.105
103
Table B.2 Area Contact Wear Data
Wear Coefficient lest i Qils Bef;r:ige;r:aDt (mg} Qf E;r:i!:tiQD
06SP R12 132.21 0.182 07SP R12 149.44 0.203 16SP Min1 8.06 0.080 18SP Min1 8.99 0.081 19SP Min1 R12 1. 95 0.055 20SP Min1 R12 1.83 0.044 21SP PAG1-B 14.31 0.063 22SP PAG1-B 12.49 0.072 23SP PAG1-B 1.88 0.018 24SP PAG1-B 2.11 0.022 26SP R134a Seized 27SP PAG1-B R134a 37.65 0.042 28SP PAG1-B R134a 38.11 0.048 29SP R134a Seized 30SP PAG1-F R134a 22.08 0.065 31SP PAG1-F R134a 19.04 0.061 32SP Est2-B 1.36 0.060 33SP Est2-B 1. 52 0.067 34SP Est2-F 0.22 0.037 35SP Est2-F 0.17 0.035 36SP Est2-B R134a Seized 37SP Est2-B R134a Seized 38SP Est2-F R134a 0.33 0.057 39SP Est2-F R134a 0.26 0.065
104
Table B.3 Counterformal Contact Wear Data
Wear Scar Coefficient lest i Qils Befz:igez::ant OeJ;2tb. 0.1. in} Qf Ez::ic:tiQn
02RC Min2 1250 0.130 03RC Min2 1000 0.150 04RC Min2 R12 560 0.150 05RC Min2 R12 516 0.143 06RC R12 550A
07RC R12 460A 10RC Est2-B 4530 0.114 11RC Est2-B 4360 0.128 12RC Est2-F 5240 0.147 13RC Est2-F 5130 0.173 15RC Est2-B R134a 4410 0.081 16RC Est2-B R134a 4830 0.081 17RC Est2-F R134a 5000 0.110 18RC Est2-F R134a 5240 0.126 19RC Est3-B 8510 0.117 20RC Est3-B 8330 0.154 21RC Est3-F 8210 0.130 22RC Est3-F 8750 0.136 23RC Est3-B R134a 9000 0.106 24RC Est3-B R134a 9150 0.126 25RC Est3-F R134a 8540 0.146 26RC Est3-F R134a 8930 0.157 27RC PAG2-B 2760 0.124 28RC PAG2-B 2660 0.098 29RC PAG2-B R134a 3160 0.092 30RC PAG2-B R134a 3350 0.099 31RC PAG2-F 1400 0.110 32RC PAG2-F 1250 0.091 33RC PAG2-F R134a 2660 0.111 34RC PAG2-F R134a 2860 0.114
A These two tests show small wear scar depths, but the surface finish of the aluminum pad was very rough. The other tests had wear scars that were smooth.
105