National Academies Press: OpenBook

Measuring Tire-Pavement Noise at the Source (2009)

Chapter: Chapter 4 - Evaluation of OBSI Test Parameters

« Previous: Chapter 3 - Evaluation of Alternative Test Methods
Page 11
Suggested Citation:"Chapter 4 - Evaluation of OBSI Test Parameters." National Academies of Sciences, Engineering, and Medicine. 2009. Measuring Tire-Pavement Noise at the Source. Washington, DC: The National Academies Press. doi: 10.17226/14212.
×
Page 11
Page 12
Suggested Citation:"Chapter 4 - Evaluation of OBSI Test Parameters." National Academies of Sciences, Engineering, and Medicine. 2009. Measuring Tire-Pavement Noise at the Source. Washington, DC: The National Academies Press. doi: 10.17226/14212.
×
Page 12
Page 13
Suggested Citation:"Chapter 4 - Evaluation of OBSI Test Parameters." National Academies of Sciences, Engineering, and Medicine. 2009. Measuring Tire-Pavement Noise at the Source. Washington, DC: The National Academies Press. doi: 10.17226/14212.
×
Page 13
Page 14
Suggested Citation:"Chapter 4 - Evaluation of OBSI Test Parameters." National Academies of Sciences, Engineering, and Medicine. 2009. Measuring Tire-Pavement Noise at the Source. Washington, DC: The National Academies Press. doi: 10.17226/14212.
×
Page 14
Page 15
Suggested Citation:"Chapter 4 - Evaluation of OBSI Test Parameters." National Academies of Sciences, Engineering, and Medicine. 2009. Measuring Tire-Pavement Noise at the Source. Washington, DC: The National Academies Press. doi: 10.17226/14212.
×
Page 15
Page 16
Suggested Citation:"Chapter 4 - Evaluation of OBSI Test Parameters." National Academies of Sciences, Engineering, and Medicine. 2009. Measuring Tire-Pavement Noise at the Source. Washington, DC: The National Academies Press. doi: 10.17226/14212.
×
Page 16

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

11 Introduction As explained in Chapter 3, the OBSI method was selected as a basis for developing an onboard, at-the-source measure- ment procedure for tire-pavement noise. As a portion of the work conducted to develop such a procedure, test variables and measurement uncertainties were examined. Based on input from current OBSI users as well as information contained in the draft ISO CPX procedure, some pertinent variables that could affect the measurement results were identified. The sen- sitivity of OBSI results to variations in pavement temperature, the configuration of the OBSI measurement fixture, tire infla- tion pressure, test vehicle type, test speed, and load were inves- tigated. The intent of this investigation was to provide initial guidance on test variables and the control limits needed to implement the OBSI procedure. This chapter summarizes the evaluation and results of the test parameter investigation and makes recommendations on parameter limits and controls (additional information on the measurement sites and proto- col, along with a more detailed explanation of the results of this analysis, are provided in Appendix C). Description of Field Measurements Parameters Evaluated Measurements were conducted to evaluate vehicle vari- ables and test execution variables on OBSI measurement results. The test matrix is shown in Table 2. Environmental variables, such as air and pavement temper- atures, wind conditions, and moisture conditions, could not be systematically controlled for these tests. However, temper- ature and wind conditions were measured throughout, and testing conducted over the extremes encountered was evalu- ated. All testing was conducted under dry conditions. Vehicle variables, including loading, tire inflation pressure, and vehicle-to-vehicle variation, were evaluated systematically. Because of the time period of the testing, longer term variables of tire wear and hardness were not evaluated and wheel align- ment was not evaluated except as it occurred from test vehicle to test vehicle. Test execution variables including probe loca- tion, run-to-run and day-to-day repeatability, probe configu- ration, small variations in test speed, and reproducibility were also measured. Reproducibility across multiple users was not assessed. Measurement Sites The initial portion of this testing was conducted at Minnesota DOT’s MnROAD Low Volume Road facility in Albertville, MN. This facility is a 2.5-mile closed loop that contains 20 pavement test sections. Two of these sections, a fine textured AC and a random transversely tined PCC, were selected as test surfaces. Due to an extended period of rain, testing was limited to the SRTT tire and only a portion of the test matrix was completed. The remainder of the testing was conducted at the General Motors Desert Proving Ground (DPG) in Mesa, AZ, on relatively smooth AC and exposed aggregate PCC test sections. The site location, photographs of the pavement sections, and the average 1⁄3 octave band spec- trum for each surface under baseline conditions are provided in Appendix C. Measurement Protocol The baseline test condition for each test pavement and test tire followed the measurement protocol presented in Attach- ment 1 using “full-sized” rental vehicles along with a baseline load consisting of two people and the OBSI instrumentation. A photograph of the OBSI equipment installed on a test vehi- cle is shown in Figure 8. Ideally, the same test vehicle would have been used as the baseline for all of the test scenarios. However, due to the relo- cation of the second portion of the testing, two different base- line vehicles were used. The test vehicle used at MnROAD C H A P T E R 4 Evaluation of OBSI Test Parameters

12 was a 2007 Buick Lacrosse CX. At the GM DPG, a 2007 Pon- tiac Grand Prix was used as the primary (baseline) test vehi- cle. The baseline tire was the Michelin/Uniroyal SRTT, with the Dunlop SP Winter Sport M3 tire (Dunlop) used in those conditions where tire-specific results are suspected to occur due to tread pattern differences. Photographs of the two test tires were provided in Chapter 3. Measurements were conducted using the two-probe approach (15) at a baseline test speed of 60 mph and a “cold” tire inflation pressure of 30 psi. For the baseline condition, the probe was positioned 3 in. from the pavement surface and 4 in. from the face of the tire, at locations opposite the lead- ing and trailing contact patch of the tire, and oriented so that the sensitive axis was positioned toward the tire. For evaluat- ing the effects of temperature, testing was not restricted to a specific temperature range. Three vehicle passes were made for each test parameter, which were averaged together during post analysis. A series of repeat baseline configuration measurements was performed at the completion of each set of tests for each parameter. In addition to the repeat baselines, 10 or more consecutive base- line passes were measured for each test tire to examine the run-to-run repeatability under the baseline configuration. These consecutive baseline measurements were assessed indi- vidually to examine the run-to-run repeatability under opti- mal conditions. To evaluate the variations in OBSI levels attributable to the testing parameters, each 3-pass set of parameter measurements was compared to the 3-run sets of baseline measurements performed at the start and comple- tion of each series of tests for each parameter. The microphone signals were acquired with a five channel commercial analog to digital converter, which also powered the microphones and provided signal conditioning. This unit was interfaced to a laptop computer that used commercial soft- ware to produce first Fourier transform (FFT) narrow band and 1⁄3 octave band sound pressure and sound intensity levels using a 5-second averaging time. The microphones were cali- brated using a Class I precision acoustic calibrator set for 94 dB at the beginning and end of the measurement period. OBSI quality metrics of coherence between the two microphones comprising each probe and the difference between sound pres- sure and sound intensity level were monitored during data acquisition. The actual time signals of the four microphones were also monitored in order to identify any data acquisition abnormalities. Meteorological Conditions Noise measurements at the MnROAD facility were con- ducted on August 17, 2007, from 8:00 am until 8:15 pm. Air temperatures ranged from about 60°F at 8:00 am to a high of about 74°F at 2:00 pm and down to 66°F by 8:00 pm. The sky was clear during the early part of the testing period and then became overcast in the late afternoon into the evening. Over the four days of testing at the GM DPG (September 10–13, 2007) clear skies prevailed and air temperature ranged from 86°F to 107°F. Easterly winds of up to about 18 mph were present on September 11th and 12th, parallel to the ori- entation of the test sections, resulting in almost no crosswind. Results of Parameter Investigation Run-to-Run Repeatability of Baseline Condition At the beginning of the testing for each tire, ten or more consecutive passes were measured to examine the run-to-run repeatability under the baseline configuration. The tests were conducted using the SRTT tire at MnROAD and the Dunlop tire at the DPG. Testing for the SRTT runs occurred over a period of about 50 min, with an air temperature varying no more than 2°F. The Dunlop measurements were made over a period of about 25 min, with air temperatures varying no more than 2°F. A summary of the total range (difference between maximum and minimum for all runs) in overall A-weighted sound intensity levels and 1⁄3 octave bands for the consecutive baseline runs is shown in Table 3 along with the standard deviation.Figure 8. OBSI equipment installed on DPG test vehicle. Table 2. Test parameter matrix. Parameter Variable Values Tire Repeatability (run-to-run) 10 consecutive runs SRTT, Dunlop Repeatability (day) Nominal conditions each day SRTT, Dunlop Probe configuration Single probe, dual probe SRTT Probe location, vertical ±¼”, +½” vertical SRTT, Dunlop Probe location, fore/aft ±½”, ±1” fore/aft SRTT, Dunlop Probe location, from tire ±½”, -1” from tire SRTT, Dunlop Test speed ±2, ±4 mph SRTT, Dunlop Inflation pressure ±4, ±8 psi SRTT, Dunlop Load +100, +200 lbs SRTT, Dunlop Test vehicle 4 vehicles SRTT, Dunlop

13 The total range in overall A-weighted OBSI levels for the consecutive baseline runs was 0.8 dB for the SRTT tire on both the AC and PCC pavements. For the Dunlop tire, the range in level was 0.6 and 0.7 dB for the AC and PCC pavements, respectively. The baseline runs for this portion of the analysis were made consecutively and no changes in the fixture config- uration or measurement protocol were made between runs. As a result, the difference measured for the consecutive baselines can be considered to be measurement uncertainty. Where OBSI levels under different parameter values fall within the standard deviation of the consecutive baselines, the changes in noise level cannot be reasonably attributed to changes in the given parameter because of this uncertainty. Test Tire (SRTT versus Dunlop) Data obtained using the SRTT and the Dunlop tires were examined using the baseline measurement results from the DPG, where both test tires were assessed on the same set of pavements. Because baseline measurements were conducted for each test tire over a period of several days, some variation in the baseline levels occurred because of temperature varia- tions (discussed under Environmental Variables). To more readily examine the differences in noise between the two test tires, baseline measurements were averaged for each tire on both the AC and PCC pavements. The Dunlop tire baseline measurements conducted prior to 8:30 am were not included because similar early morning measurements were not con- ducted with the SRTT. The Dunlop tire resulted in overall sound intensity levels that were 2.2 and 1.9 dB higher than the SRTT levels for the AC and PCC pavements, respectively, with an average difference of 2.0 dB. Higher 1⁄3 octave band levels occurred with the Dunlop tire for all frequencies except the 2,000 and 2,500 Hz bands, where levels with both tires were similar. The average 1⁄3 octave band spectrums for each surface at the DPG facility under baseline conditions for both tires are shown in Figure 9. The differences in these measurements were somewhat lower than those measured previously for Test Sites S1, S4, S5, W5, and Waverly which ranged from 2.3 to 3.0 dB, with the Table 3. Variation in OBSI levels for consecutive baseline runs. Figure 9. 1⁄3 Octave band OBSI levels for SRTT and Dunlop tires. SRTT - AC SRTT - PCC Dunlop - AC Dunlop - PCC Oct. A-WtdA-Wtd Oct. A-Wtd Oct. A-Wtd Oct. Range, dB 0.8 0.5 to 2.5 0.8 0.7 to 1.8 0.6 0.5 to 1.2 0.7 0.6 to 2.0 Std Dev, dB 0.3 0.2 to 0.9 0.3 0.2 to 0.6 0.1 0.1 to 0.3 0.2 0.2 to 0.6 65 70 75 80 85 90 95 100 105 40 0 5 00 630 80 0 1 00 0 1 25 0 1 600 2000 2500 3150 4000 500 0 Frequency, Hz So un d In te ns ity L ev el , d BA SRTT AC SRTT PCC Dunlop AC Dunlop PCC

14 Dunlop producing higher levels than the SRTT. They were, however, similar to the differences found for the passby sites (see Chapter 5), which ranged from 0.9 dB to 3.1 dB with an average difference of 2.0 dB. Environmental Variables Three-run series of baseline configuration measurements were performed at the completion of each set of tests for each parameter. Over the day of testing at MnROAD, air tempera- ture ranged from about 66.2°F to 74.3°F for the baseline con- figurations and pavement temperature varied from 79.9°F to 107.2°F for the AC pavement and from 78.1°F to 98.6°F for the PCC pavement. The temperature fluctuation throughout the day in Mesa was greater than that in Minnesota and, unlike the MnROADs testing, the baselines at the DPG were acquired over multiple days. During baseline measurement runs at the DPG site, the air temperature ranged from about 99.0°F to 102.9°F and pavement temperature varied from 99.0°F to 141.8°F for the AC pavement and from 95.4°F to 131.4°F for the PCC pavement over the four-day testing period. The rela- tionship between overall A-weighted OBSI levels and the mea- sured air temperatures are plotted in Figure 10 for the SRTT and Dunlop test tires on both DPG AC and PCC pavements, along with a linear regression for each data set. The data indicated no clear correlation between the SRTT and air/pavement temperature. A slight downward trend with an increase in air and pavement temperature was found, but r2 values were very low (0.0 to 0.4). In addition, the range in over- all levels for the SRTT tire was only slightly higher than the standard deviation of the consecutive baselines: 0.5 and 0.3 dB for the AC pavements, and 0.6 and 0.5 dB for the PCC pave- ments at MnROAD and the DPG, respectively. For the Dun- lop tire at the DPG, the ranges in level between baselines were 1.1 and 1.0 dB for the AC and PCC pavements, respectively, with the levels showing a decreasing trend with an increase in temperature. The results indicate a decrease of 1 dB in the over- all OBSI level measured with the Dunlop tire with an air tem- perature increase of about 18°F. This corresponds to 1 dB decrease in level for a 48.6°F increase in pavement tempera- ture. For the data of Figure 10, the r2 values for noise-to- temperature regressions for the Dunlop tire were 0.76 and 0.82. The spectra for temperature changes increased or decreased with temperature in a uniform manner. Systematic Vehicle and Test Execution Variables Measurement parameters including probe location in the vertical and fore/aft directions, probe distance from tire side- wall, vehicle test speed, vehicle loading, and tire inflation pressure were evaluated incrementally for both the SRTT and Dunlop tires. Three-run sets of baseline repeats, as discussed for the evaluation of environmental variables, were con- ducted prior to and after each series of tests for each param- eter. For some of the parameters, no difference in level within the established measurement uncertainty could be deter- mined. The most sensitive parameters were found to be vari- ation of probe location in the vertical direction, vehicle speed, tire inflation pressure, and vehicle loading, as summarized in Table 4 for the range of parameter values defined in Table 2. The trends noted in Table 4 only apply within the range of the parameter variations measured in this testing. Given the limited number of data points for each parameter, only linear relationships were considered. The range of slopes reported for any one parameter reflects the range found for Figure 10. OBSI level versus air temperature for DPG pavement sections. -0.6 -0.4 -0.2 0. 0 0. 2 0. 4 0. 6 0. 8 1. 0 85 90 95 100 105 Air Temperature, °F D iff er en ce fr om A ve ra ge B as el in e L ev el , d BA SR TT AC SR TT PC C Dunlop AC Dunlop PCC Dunlop AC Dunlop PCC SR TT AC SR TT PCC

15 specific tires and pavement (complete results are provided in Appendix C). There was a consistent downward trend in noise levels as the probe location was moved incrementally from 1⁄4 in. below to 1⁄2 in. above the standard probe location in the vertical direction (about 0.4 dB decrease in noise levels per 1⁄4 in. of movement). For vehicle test speed, OBSI noise levels increased with speed (by about 0.3 dB per 1 mph). Sim- ilarly, noise levels increased with an increase in the vehicle load (0.2 to 0.4 dB increase per 100 lb load increase). For probe location in the vertical direction and vehicle test speed, similar trends were indicated over both the AC and PCC pavements for both the SRTT and Dunlop test tires and the spectral characteristics of each pavement were maintained. Vehicle loading resulted in slightly lower increases on the AC pavement (and SRTT tire) than on the PCC pavement (and Dunlop tire); there was a 0.2 dB increase per 100 lbs load for the AC pavement, as compared to 0.3 and 0.4 dB increases for the PCC pavement. The loading-related increases occurred primarily in the frequencies below 1,000 Hz, although a small increase in the mid to high frequencies occurred on the AC section. As tire inflation pressure increases, 1⁄3 octave band levels below 1,000 Hz decrease and levels above 1,000 Hz increase, resulting in small overall changes to the sound intensity level (0 to 0.5 dB increase per 10 psi increase) as shown in Figure 11. These changes are within the repeat baseline variability. How- ever, the frequency shifts are notable; a 2.4 dB increase per 10 psi increase was indicated in the 1,250 Hz band for both pavements; shifts in the other frequency bands were smaller. The data did not indicate a clear correlation between OBSI levels and probe location in the fore/aft directions. A small downward trend in noise levels occurred as the probe location was moved further from the tire sidewall (about 0.2 dB per 1⁄2 in. of movement). The changes in noise level due to varia- tion of the probe distance from the tire sidewall are generally within the standard deviation for the consecutive baselines and slight variation of these parameters in the testing config- uration is not anticipated to affect the OBSI result (assuming testing is conducted following the standard protocol). The spectral characteristics of each pavement were maintained. Test Vehicle At the GM DPG, a 2007 Pontiac Grand Prix was used as the primary (baseline) test vehicle and results were compared to three other vehicles, including a second 2007 Pontiac Grand Prix, a 2007 Chevrolet Impala, and a 2007 Buick Lacrosse. The same measurement system and tires were used for all vehicles. For the SRTT tire, the overall levels varied by up to 0.6 dB for the AC pavement and by up to 0.8 dB for the PCC pavement. For the Dunlop tire, the overall levels varied by up to 1.2 dB for both the AC and PCC pavements. Although the differ- ences in level between test vehicles exceeded the standard deviation for the consecutive baseline runs, the variability of Table 4. Linear relationships between test parameters and OBSI levels. Figure 11. 1⁄3 Octave band levels at various tire inflation pressures, SRTT test tire. Parameter Linear Regression Slopes Probe Location, Vertical -0.3 to -0.4 dB per ¼” upward movement Vehicle Test Speed +0.2 to +0.3 dB per 1 mph increase Tire Inflation Pressure Frequency shift, see explanation below Vehicle Load +0.1 to +0.4 dB per 100 lb increase 65 70 75 80 85 90 95 100 105 400 500 630 800 1000 1250 1600 2000 2500 3150 4000 5000 Frequency, Hz So un d In te ns ity L ev el , d BA AC - 42 psi PCC - 42 psi AC - 26 psi PCC - 26 psi

16 the environmental conditions was also considerably higher for these runs. The differences between vehicle results can be attrib- uted to differences in temperature (discussed in this chapter under Environmental Variables), which were more apparent with the Dunlop tire. A considerable amount of time was required to change tires and vehicles between measurement sets, resulting in notable air and pavement temperature dif- ferences between vehicle sets. Measurements conducted during the morning period (prior to 9:30 am), when tem- peratures were 10.8°F to 12.6°F lower than during the late morning and afternoon, resulted in the highest levels. The tests conducted during midday (between 11:00 am and 4:00 pm) using three vehicles yielded results within 0.5 dBA for both the SRTT and Dunlop tires and similar spectral characteristics. Although the test vehicle variation did not produce substantial differences in OBSI levels, the same vehicle family, measure- ment system, and tires were used. Differences resulting from a wider range of vehicle types, OBSI measurement equipment, and multiple test tires of the designs, were not evaluated. Fixture Configuration (Single Probe versus Dual Probe) Single versus dual probe configurations were examined using the SRTT tire at the GM DPG site (photographs of the probe configurations are included in Appendix C). The comparison of the probe configurations was made for test speeds of 45 and 60 mph. At 45 mph, the dual probe produced levels that were 0.1 dB to 0.5 dB lower for both pavements, while at 60 mph, the dual probe levels were 0.1 to 1.0 dB lower. These typically small and varied differences in level are consistent with those reported previously (14). The spectral shapes for both probes were very similar throughout the measured frequency range. Data Quality Criteria During the data acquisition, the coherence between the sig- nals from the two microphones comprising each probe and the difference between sound pressure and sound intensity level (PI Index) were monitored and recorded for each 1⁄3 octave band. Coherence is a measure of the linear dependency of two signals with a value of 0 being no dependency, and a value of 1 being perfect linear dependence (16). Mathematically, it is the magnitude of the cross-spectrum between two signals squared divided the product of the auto-spectrum of both signals. For sound intensity measurements made in flow such the OBSI measurements, it is generally found that the data are contami- nated with flow noise when the coherence falls below 0.8 (15). With only a few exceptions, the coherence was greater than 0.8 in all 1⁄3 octave bands from 400 to 4,000 Hz during the parameter measurements. In the 400 and 4,000 Hz bands, slight decreases in coherence occurred at the DPG site in 4 out of 578 runs when high temperatures caused equipment overloads and overheating (these data were discarded). Above 4,000 Hz, coherence is typically lower due to limitations in the finite difference approximation used in the algorithm for determining sound intensity (15). At the 5,000 Hz band, coher- ence was less than 0.8 for 38% of the parameter runs. The PI index is also used as a data quality check. Generally, if the PI index is above 5 dB, the measurement is contaminated by flow noise (14). In the parameter testing, the PI index was less than 5 dB in all 1⁄3 octave bands from 500 to 5,000 Hz. PI index values for the trailing edge position occasionally exceeded 5 dB in the 400 Hz band (about 3% of the runs). Because the levels in the 400 Hz 1⁄3 octave band were sufficiently low so as to have minimal effect on the overall level, the 400 Hz band was not included if the PI index exceeded 5 dB. Recommendations on Parameter Limits Based on the results of this research, parameter limits listed in Table 5 for the run-to-run variation, variation of probe location in the vertical direction, vehicle speed, tire inflation pressure, and vehicle loading are recommended. Reasonable variations in some of the testing parameters including loca- tion of the probe in the fore/aft direction and probe distance from the tire sidewall would not be anticipated to adversely affect the OBSI results. Parameter limits on these less sensi- tive variables and on the data quality criteria are based on the results of this study, as well as general experience in conduct- ing these field measurements. Sufficient data on the effects of environmental variables on OBSI levels are not available to set limits at this time. Measure- ment, monitoring, and documentation of air temperature, pavement temperature, wind speed and direction, and pave- ment dampness, as indicated in the standard protocol, may help researchers to establish these variables in time over a larger data set. Table 5. Recommended parameter limits. Parameter Recommended Criteria (Limit) Run to Run Repeatability, Overall A-Wtd OBSI level Within 1 dB Run to Run Repeatability, Octave Band Levels Within 2 dB Probe Location, Vertical 3 ± ¼” above pavement Vehicle Test speed 60 ± 1 mph Tire Inflation Pressure (Cold) 30 ± 2 psi Vehicle Load ± 100 lbs Probe Location, Fore/Aft Leading/Trailing edge ± ½” Probe Distance from Tire Sidewall 4 ± ½” Coherence > 0.8 for frequencies below 4,000 Hz PI Index < 5 dB for data reported as valid

Next: Chapter 5 - Demonstration Testing of OBSI Procedure »
Measuring Tire-Pavement Noise at the Source Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s National Cooperative Highway Research Program (NCHRP) Report 630: Measuring Tire-Pavement Noise at the Source examines a suggested procedure for measuring tire-pavement noise at the source using the on-board sound intensity (OBSI) method.

The following appendixes to the report are available online.

Appendix A: Review of Literature

Appendix B: Test Evaluation of Candidate Methods and Recommendation for Test Procedure Development

Appendix C: Results of Test Parameter Evaluation

Appendix D: Demonstration Testing of OBSI Procedure

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!