See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/325906274 Improving ASTM D445, the Manual Viscosity Test, by Video Recording Article in Journal of Testing and Evaluation · January 2019 DOI: 10.1520/JTE20170341 CITATIONS READS 3 2,318 7 authors, including: Alexander F. Gutsol Actively searching for new opportunity 182 PUBLICATIONS 10,376 CITATIONS SEE PROFILE All content following this page was uploaded by Alexander F. Gutsol on 30 May 2022. The user has requested enhancement of the downloaded file. Journal of Testing and Evaluation doi:10.1520/JTE20170341 / Vol. 47 / No. 1 / 2019 / available online at www.astm.org Hung Khuu,1 Nay Yee,1 Albert Butterfield,1 Mark Meiser,1 Tao Wei,1 Alexander Gutsol,2 and Michael Moir1 Improving ASTM D445, the Manual Viscosity Test, by Video Recording Reference Khuu, H., Yee, N., Butterfield, A., Meiser, M., Wei, T., Gutsol, A., and Moir, M., “Improving ASTM D445, the Manual Viscosity Test, by Video Recording,” Journal of Testing and Evaluation, Vol. 47, No. 1, 2019, pp. 310–323, https://doi.org/10.1520/JTE20170341. ISSN 0090-3973 ABSTRACT Manuscript received June 14, 2017; accepted for publication November 20, 2017; published online June 21, 2018. 1 2 Chevron Energy Technology Company, 100 Chevron Way, Richmond, CA 94801, USA Chevron Energy Technology Company, 100 Chevron Way, Richmond, CA 94801, USA Corresponding author, e-mail: alexander.gutsol@gmail.com, https://orcid.org/0000-00020458-4774 Video recording of the manual version of the ASTM D445, Standard Test Method for Kinematic Viscosity of Transparent and Opaque Liquids (and Calculation of Dynamic Viscosity), effectively automated this method, resolved ergonomic concerns, reduced chemical exposure of an operator, drastically reduced probability of errors of the operator, significantly improved productivity, and simplified the analyst training process and work turnover. The accuracy of results is acceptable from the standpoint of the ASTM D445 method. Keywords video registration, manual viscosity test, ASTM D445, automation Introduction ASTM D445, Standard Test Method for Kinematic Viscosity of Transparent and Opaque Liquids (and Calculation of Dynamic Viscosity) [1], was approved for the first time in 1937 and is still one of the most frequently used methods for liquid petroleum products and crude oil. There are many methods and standards for viscosity measurements (see the list of the methods in an appendix of the book in Ref. [2]). Under the jurisdiction of ASTM Committee D02 on Petroleum Products, Liquid Fuels, and Lubricants and in the direct responsibility of Subcommittee D02.07 on Flow Properties, there are two other methods, ASTM D7042, Standard Test Method for Dynamic Viscosity and Density of Liquids by Stabinger Viscometer (and the Calculation of Kinematic Viscosity) [3], and ASTM D7945, Standard Test Method for Determination of Dynamic Viscosity and Derived Copyright © 2018 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959 310 KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING Kinematic Viscosity of Liquids by Constant Pressure Viscometer [4], that have about the same accuracy and are applicable to the same petroleum products with Newtonian viscosity. Nevertheless, so far, these methods have not reached as wide a penetration in the oil industry as ASTM D445. This might be related to the industry inertia as well as to some range limitations in viscosity (≈30,000 mm2/s for ASTM D7042 [5] and 1,000 mm2/s for ASTM D7945 [4] versus 100,000 mm2/s for ASTM D445) and temperature (135°C for ASTM D7042 [5] and 120°C for ASTM D7945 [4] versus 200°C for ASTM D445). Another important limitation of ASTM D7042 and ASTM D7945 is that the equipment they use measures dynamic viscosity together with density for the calculation of kinematic viscosity. Measurement of density is realized using a very precise method based on the oscillation of a U-shaped tube filled with a specimen. This equipment is rather complex and needs to be serviced by specially trained technicians in contrast with rather simple and robust equipment for the manual version of ASTM D445. Equipment complexity becomes especially important when specimens are very viscous or difficult to wash out using conventional solvents, so-called difficult-to-run specimens. For example, our experience with ASTM D7042 revealed significant disadvantages of this method in comparison with ASTM D445 when working with viscous specimens. Thus, when putting the viscous specimen into a syringe, it is necessary to heat the specimen before transferring it into the preheated syringe without any bubbles, and the injection of the specimen into a cell with such a small opening caused pain in the operators’ hands after the injections. Also, cleaning parts after a test takes much effort and time. Under the jurisdiction of Subcommittee D02.07, there are two other active standard methods, D7483, Standard Test Method for Determination of Dynamic Viscosity and Derived Kinematic Viscosity of Liquids by Oscillating Piston Viscometer [6], and D7279, Standard Test Method for Kinematic Viscosity of Transparent and Opaque Liquids by Automated Houillon Viscometer [7]; however, they have significantly lower accuracy and applicability ranges. Different automatic and automated viscometers were developed for ASTM D445 that provide higher productivity and have better determinability: 0.15 % at 40°C and 0.20 % at 100°C [8] versus 0.37 % at 40°C and 0.36 % at 100°C for the manual version of the method [9]. Per ASTM D445, determinability is “the difference between successive determined values obtained by the same operator in the same laboratory using the same apparatus for a series of operations leading to a single result, would in the long run, in the normal and correct operation of this test method, exceed the values indicated only in one case in twenty” [1]. Despite the availability of automated viscometers, there is still a need for a manual version of this method. Examples that prove a reason for such a necessity can be found in paragraph 10.2.3 of ASTM D445: “Viscometers used for silicone fluids, fluorocarbons, and other liquids which are difficult to remove by the use of a cleaning agent, shall be reserved for the exclusive use of those fluids except during their calibration.” Another reason for limited applicability of the automated method is related to the nature of the automation. Most of the available automated viscometers have sensors (e.g., thermistors) for flow time determination that are embedded in the volume marks of the viscosity tubes. Therefore, the viscosity tubes are permanently fixed in the temperature bath. This normally causes the automated instrument to have a limited viscosity measurement range, and thus it is not appropriate to measure difficult-to-run specimens since the viscosity tube cannot be removed easily from the bath for cleaning. In different laboratories, this necessity can be caused by other reasons; for example, a small number of tests of a particular type don’t justify the purchase of an automated Journal of Testing and Evaluation 311 KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING 312 viscometer. For example, in our kinematic viscosity lab, in addition to four automated viscometers, we use seven baths for manual viscometers. These baths cover a temperature range from −40°C to 190°C. Note that, according to the “Precision and Bias” section of ASTM D445, “Results from Manual and Automated Instruments in Test Method D445 may be considered to be practically equivalent” [1]. The only substantial requirement that is stricter for the manual version of the test is that the flow time for manual viscometers shall not be less than 200 s, while flow times of less than 200 s are permitted for automated viscometers. The absence of the significant accuracy advantage of automatic and automated viscometers in comparison with the manual ones can be explained by the following two factors: (1) The tolerance bands of the certified viscosity reference standards are rather large (from ±0.30 to ±0.73 %, see Table 1 in ASTM D445 [1]). These bands combine both the uncertainties of the certified viscosity reference standards as well as the uncertainties of the laboratories using the certified viscosity reference standards [1]. (2) An extremely important factor in repeatability and reproducibility of the method is temperature control [10], which is based on the same liquid bath technology in both versions of the method. Other factors that have an influence on accuracy that is mostly reflected in the uncertainties just mentioned are specimen handling, viscometer cleanliness, loading procedure, and mounting alignment. Thus, the manual version of ASTM D445 is still widely used and will continue to be used in the foreseeable future. The manual version of ASTM D445 is a well-established method that should not cause any significant problems. However, this is not the case. Results of the D445 Grand Design Inter-Laboratory Study (ILS) that took place in 2005–2009 forced subcommittee D02.07A to form a special Task Group TG7A.27 “Investigation of Manual Viscometer Imprecision in D445.” The precision estimate for manual viscometers that was calculated from the ILS [9] was found to be worse than the previously published precision and much worse than many of the premier ISO 17025–accredited laboratories claim is possible. For example, in the last version of ASTM D445 published in 2014 [11] that used the old precision data, the repeatability (r) and reproducibility (R) were given as r = 0.11 % and R = 0.65 % for base oils at 40°C and 100°C, respectively. Later, starting in 2015 when the Grand Design ILS data were incorporated into a new ASTM D445 standard, for base oils at 40°C, r = 1.01 % and R = 1.36 %; and for base oils at 100°C, r = 0.85 % and R = 1.90 %. The Task Group identified many factors that contributed to the imprecision, and one key factor was formulated as follows: “Loss of experience by technicians due to most labs now using automation version of D445 primarily and manual one only for backup. Also, older technicians who were really skilled running manual viscometers have retired, and the younger ones haven’t been properly trained” (Personal communication, P. Maggi) in addition to precision concerns that are important to customers, a manual version of ASTM D445 raises other issues and TABLE 1 Approximate tolerance bands. Viscosity of Reference Material, mm2/s Tolerance Band <10 ±0.30 % 10 to 100 ±0.32 % 100 to 1,000 ±0.36 % 1,000 to 10,000 ±0.42 % 10,000 to 100,000 ±0.54 % Journal of Testing and Evaluation KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING FIG. 1 Typical position of an operator during the manual ASTM D445 test run. challenges that are important to laboratories and operators. Thus, Fig. 1 demonstrates some of these issues. It is rather difficult to make working spaces free of ergonomic problems. Timing marks on the capillary tube should be at eye level (black line in Fig. 1, consider the lines to be in the plane of the tubes) while the total height of the system (white line) should be convenient for specimen loading. In our case, the loading level is at a height of 147 cm, which is too high for some operators. To have an easily accessible stopwatch (usually a system of stopwatches for multiple simultaneous tests, as in Fig. 1), an additional level is necessary. Eyestrain will be an issue in any case because it is necessary to catch a precise moment in the motion of a small object. It is also very difficult to keep a neutral, unstrained posture during this operation. As the operator’s eyes (and, thus, head) should be in the ultimate vicinity to the test tube, it is almost impossible to eliminate exposure to chemical vapors (level of loading, white line in Fig. 1), or ventilation wind (ventilation snorkel is marked by white arrow in Fig. 1), or both. In addition to ergonomic issues, an operator faces other challenges, for example, filling a precise volume of the specimen into a viscometer tube (e.g., Zeitfuchs Journal of Testing and Evaluation 313 KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING 314 cross-arm viscometer, CANNON Instrument Company, State College, PA [12]) may require several additions and removals of tested liquid. To increase productivity, an operator very often works with several test tubes simultaneously, or even performs other tests in parallel. This multitasking significantly increases the probability of a mistake, and if a mistake is made, the reason is not always obvious. The traditional manual version of ASTM D445 does not create any records for data incident review, and this makes the process of new operator training rather difficult. We use the manual version of the method in our laboratory for specimens that cannot be tested by the available automated viscometers because of their very high viscosity or required odd temperature measurements. Because of the high viscosity of some specimens that we test manually, the test time together with the time for preparation and test tube cleaning can be rather long and can easily reach one hour. In such conditions, the cost of any potential operator mistake becomes very high as the whole hour of work can be lost. Also, there is one potential mistake that is difficult to avoid completely: precise timing. According to ASTM D445, the operator should “measure, in seconds to within 0.1 s, the time required for the meniscus to pass from the first to the second timing mark” [1]. There is also a second measurement requirement in ASTM D445, and the difference between the two measurements should be within the stated determinability for reporting the result. It is necessary to understand that 0.1 s is probably the best repeatability that can be reached by any human who is using a stopwatch for an event that is very well determined in time. The exact position of the meniscus in motion is not an easily determined event, especially if the liquid is transparent (Fig. 2). Two consecutive measurements can have random errors of the same sign, thus the difference between the two measurements will be within the stated determinability, and the result will be reported even if it is rather far from the correct value. To avoid potential systematic mistakes that are caused by invisible capillar contamination, we always use two viscometer tubes for the manual procedure as is allowed in ASTM D445: “Two determinations of the kinematic viscosity of the test material are required. For those viscometers that require a complete cleaning after each flow time measurement, two viscometers may be used” [1]. However, this precaution does not remove the possibility of having two random timing errors of the same sign. Another very costly and common mistake is to hit the wrong button on the stopwatch system array (see Fig. 1). Such a mistake ruins two tests simultaneously. In this article, we present a technical approach that can be used to partially automate the test and thus resolve almost all ergonomic issues, reduce the number of mistakes, and improve the productivity of an analyst. For automation, we used a rather obvious approach that has been applied at sporting events for many years—a “photo finish.” In our version, we recorded the whole test. This approach allows high-precision measurement of the test time (e.g., 1/30 s for 30 frames per second (fps) video), and sometimes also allowed us to determine the reason why two consecutive measurements give significantly different results. For example, sometimes the video recording allows us to find that there are particles or bubbles in a liquid that prevent correct measurements. Technical Solutions in Automation Though the photo finish approach sounds like a rather simple one, there were several technical issues to resolve, which are reported in this section. Journal of Testing and Evaluation KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING FIG. 2 Images showing the meniscus of transparent liquid before and after the first mark. CAMERA At first glance, it would seem that any reliable camera with video-recording capabilities or a camcorder could serve our purpose. This is not completely true. In recent years, there has been very rapid progression in the development of digital video recording devices, so the choice of acceptable and relatively inexpensive cameras is much better than it was several years ago when we performed this automation project. First, a camera or camcorder? We considered several requirements: (a) high sensitivity because lighting for the test tubes cannot be very strong so as to preserve temperature stability and avoid temperature gradients; (b) high-quality optics are required to ensure high resolution, which is necessary to fix precisely the event when the meniscus passes a timing mark while recording several tests simultaneously; (c) per ASTM D445 requirements, a test time should be between 200 and 1,000 s, and several tests with long times can be started at different times, so a recording time limit that is longer than 1,000 s is preferable; (d) limited application, i.e., the same focus during the whole recording and an absence of ergonomic issues for the cameraman because of the stationary position of the recording device, which removes the need for a large zoom; and (e) reasonable cost: good Digital Single-Lens Reflex (DSLR) cameras were (and still are) significantly cheaper than camcorders with comparable quality optics and sensors. Considering all these factors, we selected a Nikon D8000 camera, which was one of the best compact DSLRs at the time. It provided Full HD (High Density, 1,920 × 1,080 pixels) capabilities with a potential Journal of Testing and Evaluation 315 KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING 316 FIG. 3 Camera fixed on clamp support in use for the ASTM D445 test video registration. recording time of up to 20 minutes (4 GB). Per National Television System Committee (NTFC) standard, the recording rate is 29.97 fps, which should provide time registration accuracy of about 0.03 s. A special, easily removed clamp support was designed and manufactured to fix the camera to a test bench in a stationary and safe position (Fig. 3) and to save floor space in comparison with standard tripods. The clamp also allowed for the removal of the camera when it is not in use. LIGHT As it was mentioned earlier, intense lighting can disturb temperature stability and create temperature gradients. As a reminder, per ASTM D445, the temperature should be maintained within ±0.02°C, which is a requirement that is very difficult to obey for bath manufacturers [10]. Fortunately, LED lights were available when we performed this automation. Use of LED is illustrated in Fig. 3. We used a Humanscale Element 790 lamp (Humanscale Ergonomics, Cary, NC) that has the following important features: 360° of adjustability; the same light equivalency as that from a 90 W incandescent bulb all while Journal of Testing and Evaluation KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING using less than 7 W of power; a seven-level dimmer; neutral light with a color temperature of 3,500 K. PC HARDWARE AND SOFTWARE First, video records require significant PC memory for storage (hard drive) and operational memory (RAM) for processing. We used the practice of keeping video records of an ASTM D445 test for 6 months in case customers come back to us with questions. For video handling, we selected an Adobe Premiere software package (Adobe Systems, San Jose, CA) to edit our recordings. However, our experience shows that editing is not necessary. Without editing, we can still use some of the convenient features of the software; for example, zooming in on important frames, looking through the whole video frame by frame, etc. Very important is the ability to replay a part or the whole video as many times as necessary because it prevents unnecessary retesting of a specimen or, vice versa, it shows the necessity for retesting (e.g., when we review the video and see a bubble in the tested liquid). Together with the ASTM D445 method automation, we also automated data input to a system that reports results to customers. Using the same computer where the video is processed (Fig. 4), the data input software (Fig. 5) helps gather all necessary information about the tests. Fig. 5 presents part of an Excel spreadsheet (Microsoft Corporation, Albuquerque, NM) where “green fields” contain information about the required test conditions and “white fields” require operator input before and after the test. When the tests are done, an operator uploads the video to the PC and opens it in the video processing software (Fig. 6). Then, the operator can easily find frames (start and end, Fig. 7) where the events of meniscus passing the first and the second timing marks are recorded. The operator inputs these frame numbers into the table (Fig. 5). The frame number is displayed in orange in Fig. 6. The software calculates time and viscosity (using the tube constant), the report results are delivered to the customer, and all relevant test data (e.g., tube ID) is sent to storage. The total time required for an operator to transfer the video recordings from the camera to the PC, then to find all start and end frames and put the data into a software table is about 10 minutes. Different operators, if they were not trained in the same way, may have different methods for determining the frame when a meniscus passes the timing mark, yet the important advantage of this automation approach becomes obvious: it is very easy to train a new operator using multiple video records. Also, it is possible to discuss any trainee questions after the test because a record of the whole test is available. FIG. 4 Example of video registration of eight tests simultaneously. Journal of Testing and Evaluation 317 KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING 318 FIG. 5 Data input software table. FIG. 6 Analysis of the test record using video processing software. The frame number is displayed in orange in the lower left part of the right window. FIG. 7 The frames selected as the start and end frame of a test. Journal of Testing and Evaluation KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING TIME VERIFICATION Per the NTSC standard, the video recording rate is 29.97 fps that corresponds to a time interval of 0.033374 s per frame. Per ASTM D445, a timing device should be capable of taking readings with a discrimination of 0.1 s or better and has an accuracy within ±0.07 %. To verify that a camera video recording satisfies these requirements, we did five tests using a site with official U.S. time [13]. Fig. 8 demonstrates an example of a record for such a test. Note that we select the frames where the new second digit appeared clearly for the first time. This allowed time verification with high accuracy. Average time per frame calculated using these tests is 0.033365 ± 0.000019 s. Thus the accuracy is within ±0.06 %. VISCOSITY MEASUREMENTS VERIFICATION As a final test of the automated system, we measured viscosities of the available Certified Viscosity Reference Standards during several months using different standards and different batches of the same standards. In parallel to video recording, we used conventional manual (“click”) time registration for a fair comparison. Results are presented in Fig. 9 as differences in percentage between the measured viscosity and that from the sample certificate of analysis. FIG. 8 An example of the time accuracy verification test that shows time interval between frames #2 and #10012. FIG. 9 Differences between the data for viscosity standards and our measurements. Journal of Testing and Evaluation 319 KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING 320 Per ASTM D445 recommendations, we used Table 1 to determine the approximate tolerance bands for the reference materials we used. ASTM D445 requires that when the determined kinematic viscosity did not agree with the acceptable tolerance band of the certified value (e.g., an experimental point for the standard with a certified viscosity of 33.47 mm2/s that has a difference of +0.33 %), it was necessary to recheck each step in the procedure to locate the source of error. However, it was not possible to recheck timing in the case of the manual measurement. Thus, when all other potential sources of errors were eliminated, we simply repeated the test. Note that, for some standards, all experimental points are on one side of the X-axis, e.g., for the standard with a certified viscosity of 310.4 mm2. We did not investigate the reason for such “oriented” differences because this kind of investigation is not required by the ASTM D445 method. As we understand, the reason why it is not required lies in wellbalanced requirements and uncertainties in viscosity measurements and characterization of viscosity standards. Standards are characterized and certified using the same methods all laboratories use. Therefore, for example, in ASTM Cooperative Kinematic Viscosity Measuring Program tests, a corrected value of a sample is accepted as the average value derived from all measurements. As described in the Introduction, the tolerance bands of the certified viscosity reference standards are rather large for multiple reasons, most of which are listed in the Introduction. Only improving the time measurement cannot significantly improve the method accuracy, and the goal of our automation project was not accuracy improvement. It is possible to see that the differences between the video-recorded results and the viscosity standard’s data are within the tolerance bands in all cases. Thus, the accuracy of measurements using video recording is acceptable from the standpoint of the ASTM D445 method. Method Benefits and Limitations Video recording of the manual version of ASTM D445 allowed us to resolve ergonomic concerns for our operators. It is reasonable to estimate an average time per test itself of 10 minutes (600 s) and average simultaneous load per bath of four tubes. Also, it takes about 2 minutes per load before the last of the four specimens reaches the first timing mark. Thus, an operator spends about 3 minutes per test in a rather awkward pose (see Fig. 1). The time for loading and tube cleaning is about 10 minutes per test. This means that the operator who runs a manual ASTM D445 test is spending about one quarter of the total working time in a strained position. Keep in mind, however, that some tests can be as long as 1,000 s, so the operator can be in this position continuously for almost 20 minutes. Video recording resolved this issue completely. From the numbers above, we estimate that exposure to chemical vapors was reduced by about 25 % when video recording is used. With the conventional manual method, we estimated that about 10 % of the tests needed to be rerun because the difference between the two measurements was larger than the specified determinability that was due to timing errors. With video recording, this percentage is virtually zero. Just this factor alone provides a 10 % time savings. More significant time savings arise from the substitution of observing time with video analysis. As we mentioned earlier, the analysis requires about 10 minutes per load. As estimated earlier, Journal of Testing and Evaluation KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING observation time was about 12 minutes plus time for uploading the results. All together it takes at least 15 minutes. Thus, automation provides at least a 30 % time savings. Also, from a time management perspective, it is not necessary to perform a record analysis just after the test. This yields much more flexibility for the operator, who can then run other tests simultaneously or take a break. It also enables work turnover: one operator can start a test while another operator can finish it. We have a rather long history of participation in ASTM Cooperative Kinematic Viscosity Measuring Program, and, starting from 2014, we use video registration for manual tests. Fig. 10 provides information on the difference between our results and the mean value for all participating laboratories. The number of participating laboratories varied with time, and in 2017, this number was 24. Not all laboratories participated in testing of all materials at all requested temperatures. In recent years, we tried to participate in all possible tests. The presented results are not differentiated by substance, only by method (manual or automated) and temperature. It is possible to see that with an automated method, the relative differences between our results and the mean values are always within a rather narrow range. The situation with the manual method is worse, and it is significantly worse than with measurements of viscosity standards (Fig. 9). One of the reasons for this difference can be larger uncertainty for uncertified samples in comparison with certified viscosity standards. In general, the situation agrees with the conclusion of the Task Group TG7A.27 (personal communication, P. Maggi), which is that there are systemic problems with the manual version of ASTM D445. On the other hand, we can compare our results and the mean values before and after we started to used video registration. It is possible to see that during the last four years we had only one rather big difference of −1.00 % that is FIG. 10 Results of our participation in ASTM Cooperative Kinematic Viscosity Measuring Program. Journal of Testing and Evaluation 321 KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING 322 still considered acceptable now when acceptable reproducibility R = 1.36 %. During the previous four years, in 2010–2013, we had three big differences of about 1 % in absolute value, and one difference of −1.75 %, which is outside the acceptable reproducibility. We cannot conclude that we have statistical evidence of higher accuracy using the manual method with video registration because much more data need to be produced by different labs; however, we obviously have encouraging progress. Though we are happy with the automation we did, and we have used it for several years already, this automation has some limitations. The first one being the inability to record tests in several baths simultaneously using a single camera. Secondly, the recording time is limited for the camera we use (20 minutes = 1,200 s), which is comparable to the longest possible time of a test (1,000 s); therefore, we need to be efficient when we are setting several long tests for simultaneous recording. Note that in many other ASTM standard tests developed long ago, it was necessary to “catch” an event (which was not always obvious). Such standard tests can benefit from the addition of affordable video recording, at least from the standpoint of new operator training and odd result investigations. For example, the flash point for some liquids is difficult to distinguish from the appearance of a blue halo or an enlarged flame [14]. Smoke point determination [15] by manual testing requires a significantly experienced operator, and it would be much easier done with the test video record. Conclusion We developed a video recording process for the manual version of ASTM D445 and started to use it in 2013. This approach effectively automated the test, resolved ergonomic concerns, reduced chemical exposure of operators, significantly improved productivity, and simplified analyst training and work turnover. The accuracy of the results is acceptable from the standpoint of the ASTM D445 method. With good cameras that have videorecording capabilities (though not necessarily a DSLR) becoming much more affordable, along with bright LED light sources, this approach can be attractive to many labs that use the manual version of ASTM D445. Since the probability of an uncorrectable operator mistake becomes negligible with this simple automation, the method improvement is especially beneficial for the labs that use two similar viscometer tubes to produce duplicate results. In addition to obtaining the viscosity measurement, the method allows for verification of the condition of both tubes. References [1] ASTM D445-17a, Standard Test Method for Kinematic Viscosity of Transparent and Opaque Liquids (and Calculation of Dynamic Viscosity), ASTM International, West Conshohocken, PA, 2017, www.astm.org [2] Gupta, S. V., Viscometry for Liquids, Springer International Publishing, New York, 2014, 256p. [3] ASTM D7042-16, Standard Test Method for Dynamic Viscosity and Density of Liquids by Stabinger Viscometer (and the Calculation of Kinematic Viscosity) (Superseded), ASTM International, West Conshohocken, PA, 2016, www.astm.org [4] ASTM D7945-16, Standard Test Method for Determination of Dynamic Viscosity and Derived Kinematic Viscosity of Liquids by Constant Pressure Viscometer, ASTM International, West Conshohocken, PA, 2016, www.astm.org Journal of Testing and Evaluation KHUU ET AL. ON IMPROVING D445 BY VIDEO RECORDING [5] “Stabinger Viscometer,” Anton Paar, 2018, https://web.archive.org/web/2017 1027044000/https://www.viscopedia.com/methods/instrument-categories/stabingerviscometertm/ (accessed 27 Oct. 2017). [6] ASTM D7483-13a, Standard Test Method for Determination of Dynamic Viscosity and Derived Kinematic Viscosity of Liquids by Oscillating Piston Viscometer, ASTM International, West Conshohocken, PA, 2013, www.astm.org [7] ASTM D7279-16, Standard Test Method for Kinematic Viscosity of Transparent and Opaque Liquids by Automated Houillon Viscometer, ASTM International, West Conshohocken, PA, 2016, www.astm.org [8] Interlaboratory Study to Determine Suitability of ASTM D445 for Automated and Automatic Instruments Used to Measure Viscosity of Base Oils and Formulated Oils, Research Report D02-1787, ASTM International, West Conshohocken, PA, 2014, 162p. [9] Interlaboratory Study to Determine Precision of ASTM D445-11a Manual Method when Used to Measure Viscosity of Base Oils and Formulated Oils, Research Report D02-1788, ASTM International, West Conshohocken, PA, 2014, 40p. [10] Lane, J. L. and Henderson, K. O., “Viscosity Measurement: So Easy, Yet So Difficult,” ASTM Standardization News, Vol. 32, No. 6, 2004, pp. 42–45. [11] ASTM D445-14a, Standard Test Method for Kinematic Viscosity of Transparent and Opaque Liquids (and Calculation of Dynamic Viscosity) (Superseded), ASTM International, West Conshohocken, PA, 2014, www.astm.org [12] ASTM D446-12, Standard Specifications and Operating Instructions for Glass Capillary Kinematic Viscometers, ASTM International, West Conshohocken, PA, 2017, www.astm.org [13] Official NIST US Time Widget, https://web.archive.org/web/20171031195254/https:// time.gov/widget/widget.html (accessed 31 Oct. 2017). [14] ASTM D92-16b, Standard Test Method for Flash and Fire Points by Cleveland Open Cup Tester, ASTM International, West Conshohocken, PA, 2016, www.astm.org [15] ASTM D1322-15e1, Standard Test Method for Smoke Point of Kerosine and Aviation Turbine Fuel (Superseded), ASTM International, West Conshohocken, PA, 2015, www.astm.org Copyright by ASTM International (all rights reserved), pursuant to License Agreement. No further reproduction authorized. View publication stats 323