Regulations Specs & Testing

Taming an Unruly Test


Taming an Unruly Test

For decades, the lubricants industry has struggled to tame a critical but unruly test, ASTM D3527, designed to evaluate automotive greases. This high-temperature wheel-bearing grease test could be the poster child for bad behavior. Myriad attempts to better its precision and accuracy have had little success, and the method remains so stubbornly defective that in 2015 a proposal was floated in ASTM to abandon it altogether.

Instead, that motion was voted down and the test method renewed through 2020. Because flawed as it is, a lot is riding on D3527, explained David Turner of Citgo, who chairs ASTMs Subcommittee G on Greases. It is a core test for ASTM D4950, the 30-year-old standard and specification that defines automotive wheel-bearing grease performance. Passing the D3527 test is mandatory for service-fill greases seeking GC-LB certification from the National Lubricating Grease Institute. The test is cited by the U.S. military, auto companies and component manufacturers, plus its readily available. So the industry wants it to survive-if only it can bring it to heel.

D3527 was created under the auspices of General Motors and first accepted as an ASTM method in 1976, said Turner. In the test, an automotive front-wheel hub/spindle/bearing assembly is filled with a grease sample and driven under specified loads by an electric motor. The spindle temperature is kept at 160 degrees Celsius as the hub rotates at 1,000 rpm through repeated cycles of 20 hours on/4 hours off. This stress causes the grease to degrade, harden and thicken to the point where the motor must draw more amperage to continue turning the spindle. When that current shift signals an increase in turning resistance (torque), the test is cut off and the end-of-test time recorded. The number of on-cycle hours that the grease endured helps to estimate its useful service life.

The test is meant to mimic long-term driving where the bearings get hot, then not driving so they cool down, and then heating up under long-term driving conditions again. But its a horrible test, Turner stated bluntly. The tests repeatability is so poor that the outcome of two runs can vary by plus-or-minus 80 percent and still be statistically valid. For example, a grease can claim to average 100 hours in the test if it runs once at 180 hours and then a second time at just 20 hours. Reproducibility across laboratories is even worse, as demonstrated by ASTM round-robin tests.

Since the 1980s, ideas for fixing D3527 have centered on the sampling time, bearing torque loading, bearing chamber temperature and insulation, motor variability, grease packing methods and motor speed control, among other factors. But the test is still relatively unchanged to date, remarked Autumn Chadwick, an automation engineer with ExxonMobil Research & Engineering in Paulsboro, New Jersey.

Chadwick and her colleagues at Paulsboros lubricant laboratories have been taking a serious run at unwinding the tests kinks, she reported last June to the NLGI annual meeting in Coeur dAlene, Idaho. For three decades, ASTM has recognized this as a very challenging test and has tried to improve it, she said in a presentation written with ExxonMobil R&E distinguished research associate Barb Carfolite, and Mark Hague, the technician who maintains and runs Paulsboros high-temperature wheel bearing test rigs.

Dissecting the procedure in depth a few years ago, the team initiated a series of software upgrades, insulation improvements and electric motor studies, which boosted the accuracy on their own labs test rigs. They also began to ask if the rigs motor itself might be a culprit in the variability. Now theyre hoping that their techniques and proposals can gain support in the wider grease industry, resulting in a healthier D3527 all around.

Ironically, the tests original mission has long expired, said Chadwick, who holds twin degrees in electrical engineering and physics. Cars no longer use this configuration of front-wheel bearings, as theyre now all sealed-for-life, she told the NLGI meeting. Yet, because both ASTMs grease standard and NLGI certification rely on D3527, as do numerous customers, her company was determined to get meaningful results with it.

ASTM inter-laboratory proficiency tests confirm that D3527 has plenty of room for improvement. In eight round robins from 2013 to 2017, participating laboratories were given identical samples of a wheel bearing grease to evaluate in their D3527 test stands. Their results showed little conformity, Chadwick pointed out: One recent round, in late 2017, got back results ranging from 80 to 260 hours, from five labs testing the same grease.

With three D3527 rigs of its own in Paulsboro, ExxonMobil decided to tackle this puzzle head-on. One area of inquiry focused on the rigs controllers, automation and motors and whether they were functioning correctly. We were seeing a lot of variability around the current and noticed that there had been changes in rig motors over time, Chadwick related. In fact, the power, speed and voltage of the motors used across the industry in the D3527 test can vary widely, as can their rated torque, her queries found.

As Chadwick explained, all D3527 rigs originally were driven by an ASTM-specified, one-quarter-horsepower direct-current motor. This recommended motor (now no longer available) produced major current fluctuations and was recognized as a source of errors. Eventually an alternative motor, called Option 1, was allowed in 2013–even though it, too, had issues, said Chadwick. In 2017, motor Option 2 was written into the method, and when Paulsboro adopted that model it finally saw repeatability improve. But why?

Armed with her electrical engineering skills, Chadwick began to dig more deeply into how motor selection may influence D3527 results. All three motor options, she observed, can produce the same power and rotational speed, but their voltage differs: The original motor was 120 volts, Option 1 was 90 volts and Option 3 is 130 volts. The motors torque constant also differs, at 91 Kt, 70 Kt and 60 Kt, respectively. Individual units suffer from motor winding differences, even when supplied by the same vendor. These and other variances could be affecting the measurements, Chadwick theorized.

Also, as lab motors age, they begin to require more current to produce the same amount of torque. That rise in amperage may be misread as a failure of the bearing grease-when in fact its the motor thats failing, Chadwick added.

A key takeaway, she said, is that laboratories performing D3527 need to know your motor and how it influences your test performance. She would like to see ASTM poll grease labs to find out what motor models they have and how old they are, to see if that data sheds any light on the round-robins poor results.

Eventually ASTM and NLGI will need to consider replacing D3527 with an alternate test, something thats actually representative of wheel bearings today, proposed Chadwick. That wont be easy, since other industry standard methods do not include the cycling feature that D3527 uses, and they employ bearings that are not representative of those found in automotive wheel hubs. Bearing tests such as the SKF ROF+ rig or the FE-9 from Germanys FAG may merit a look, but they are unlikely to find favor because of their steep cost, she acknowledged.

I call D3527 outdated, Barb Carfolite, Chadwicks co-author, said in an interview with LubesnGreases. Thats because the need for service-fill greases has all but disappeared, along with grease points on vehicles. Wheel bearings are now filled-for-life, and a faulty unit usually is yanked out and replaced rather than relubricated. The challenge is that this test is readily available, and by and large its inexpensive to operate. So people trust it–maybe more than they should, Carfolite cautioned.

The next step, she suggested, would be for the industry to agree on best practices for running D3527 using a torque meter to measure the shift in power demand directly, rather than monitoring the motors current draw; this would more accurately mark the end-of-test time. The change could be written into the method. She estimated that adding a torque meter to an existing rig could cost a few thousand dollars for the meter, plus data acquisition software and some rig set-up costs. Anyone could retrofit for that.

Like so many other companies, ExxonMobil is invested in meeting a quality standard, and this test has intrigued and frustrated us for some time, she continued. Maybe theres hope if we can get everyone to focus on the motor selections effect on the results. To start, electric current is not an accurate measurement for the end-of-test life, and there may be another measurement that improves the repeatability and reproducibility of this test.

David Turner, who also co-chairs NLGIs Technical Committee, noted that the grease institute has a strong interest in keeping D3527 alive-and by extension D4950: Certification of greases is a significant income stream for NLGI. Manufacturers pay an original application fee plus an annual renewal fee for each product they have certified, and for the right to show NLGIs trademarked GC and LB logos on their products.

NLGI charges an initial application fee of $625 per product (discounted to $420 for members), plus $235 per product per year for using its trademarked certification logo. Around 300 products currently are licensed and listed on the groups website. If D4950 were to disappear, so would that income for NLGI, stated Turner. Also gone would be NLGIs assurance of basic grease performance, which has helped buyers for decades.

Back in 1989, when it was first issued as an ASTM standard, Turner recalled, D4950 grease was head and shoulders above what was being sold as wheel-bearing grease at that time.

Lower-quality greases are no longer sold for the automotive market now; its almost 100 percent factory-fill and sealed-for-life, he went on. But the GC-LB certification turned out to represent pretty good service level for lots of other applications, such as multi-purpose greases for industrial plants. D4950 certification indicates the grease meets a minimum performance level for rust, extreme pressure, fretting wear, water resistance and other needs across the performance spectrum. The general grease user needs it, and it has value.

NLGIs president, Joe Kaperick of Afton Chemical, said the institute recognizes that D3527 is not the best test, but I dont see it going away soon. In 2015 it was put up for withdrawal as an ASTM method in order to gauge if there was any real interest in maintaining it-and there was. Many people wanted to keep it, so we should go back to them now and say, OK, help us fix it.

One problem, he continued, is that only a few companies with D3527 rigs have time to participate in round robins and ASTM meetings. There are a lot of things we think we could fix about the test, but theres also the risk that any changes might affect the tests severity as well as its repeatability. You need data to support any shift, and not many companies are looking at it.

Kaperick says NLGI is discussing how to move beyond GC-LB, both to preserve its income from licenses and because the market needs a way to identify top-tier greases. NLGIs GC-LB designation has become the tag that companies can put on their products to say, We have a premium grease.

We will maintain the GC-LB system, but were also in the process of developing a new system that would classify multi-purpose greases by specific performance properties, like water tolerance, high-temperature, extreme pressure, corrosion inhibition and others. The idea, explained Kaperick, is that a marketer could target one or more of these properties and then have its grease certified by NLGI. Thats where I see NLGI going. We would keep GC-LB as long as anyone wants to use it, but if we do a good job with the new system, eventually D4950 can fade away.

An NLGI working group is drafting an outline of what the new classification system would be and what ASTM methods would support it, Kaperick said. And its likely to require NLGIs financial assistance, especially for test development. Id like to see us spend money to get this thing moving, he added. Id hate to see another organization get ahead on this before NLGI does.

At least theres a glimmer of hope now, Turner added. Until recently, the discussion among NLGI and ASTM was that D3527 may not be able to be improved, he said. We were resigned to the fact that its not a great test. But a young, fresh mind like Autumn Chadwick may be what we need to keep it alive until a replacement test can be developed someday.