Choosing a Mechanical Pathology Laboratory

I have always likened oil analysis to mechanical pathology. You give the doctor a blood sample, the doctor tests it, and you get a report. The same with oil analysis: you pull a sample, analyze it, and get a report. And that’s about where the comparison ends. In medicine, you either have malaria or you don’t. You either have tuberculosis or you don’t. With oil analysis it sometimes gets a little more complicated – as one of my mentors used to say, “with oil analysis you can find yourself being a little bit pregnant”, because analyzing an oil sample is not always cut-and-dried. Choosing a pathology lab is serious business. Why do it any differently for machines?

In my work, I have come across many organizations which utilize commercial oil analysis laboratories. Out of interest I always ask: "Why are you using this lab in particular?" The answers usually include one or more of the following responses: "I don't really know"; "it's what we've always used"; "it's close by"; or my personal favorites: "they're nice and cheap" or "the service is free". Choosing a commercial oil analysis laboratory should not be taken lightly. With more than 200 laboratories in North America which analyze used oil, fortunately there is a large selection to choose from. This article focuses on factors to consider when selecting a lab, in no particular order except that the price is discussed at the end of the article, because it should be the last thing to consider.

Turnaround Time
Turnaround is one of the most critical factors, in my opinion. Receiving results two weeks after the sample was sent to the laboratory is unacceptable. Turnaround time for a routine analysis should be 24 to 48 hours after the sample is received at the laboratory (note that some of the more exotic tests may take longer, so waiting longer is occasionally acceptable). Every extra day spent waiting for results means the decisions become less and less relevant. The turnaround time also presumes that the oil from the machine gets to the lab in an acceptable amount of time. I recommend that an oil sample should be dispatched through an overnight service from a facility within 24 hours or less of being extracted.

Alert Notifications
If an abnormality is detected, the system should include some type of alert notice beyond the normal report delivery service, which may include one or more of the following:

• E-mail

• Text notification on a mobile device

• Telephone call

• Fax

The few hours saved could mean the difference between a timely intervention or a catastrophic failure.

Knowing the Target Industry
Is your laboratory familiar with your industry? Many labs specialize in a particular industry (or industries). However, this does not mean that a laboratory specializing in automotive applications can't analyze an industrial sample, but can they interpret it correctly?

This can be a tricky situation. In an ideal world, one would rely on the laboratory to perform the analysis and the end-user would make the interpretation - after all, the user should be more familiar with the environment and the machine than a diagnostician 500 miles away. But in practice, few oil analysis end-users have the time to hone their interpretation skills to the point where this is practical, so one must rely on the lab for the interpretation. A happy medium is to rely on the lab for a good, reliable interpretation that points the user in the proper direction. The user should also use his or her own skills and knowledge of the environment to fine-tune the interpretation if further problems are suspected.

Depth of Interpretation
When teaching Oil Analysis seminars, I always explain the process of interpreting a report as follows:

1. Consider the environmental information (machine running hot, dusty/wet surroundings, etc.).

2. Examine the condition of the fluid (fluid properties analysis).

3. Examine the fluid contamination parameters.

4. Examine machine wear parameters.

5. Make a diagnosis. This is a statement of the status quo of the fluid and the machine, taking into account the factors listed in the preceding steps.

6. Make a prognosis, how to handle the problem, if one is detected.

Hopefully the last step in the process is as simple as "no action required".

Concerning the environmental factors, always inform the laboratory about any information that might be relevant. Chances are that your diagnostician has never seen your plant's No. 4 kiln drive primary reducer, and doesn't know that it is running 40 degrees above normal because the oil cooler is broken.

I began my diagnostic career under the mentorship of Gary Brown of Wearcheck Africa. He had accumulated many years of experience in interpreting samples and running a successful commercial laboratory. He always said that it doesn't matter what the analyst says about the condition of the oil, contamination or wear; all that counts is that the end-user is told what to do. Make the decision for the end-user. The average person isn't particularly interested if the particle count has increased, or if the iron and chromium have doubled since the last sample. He or she wants to be told to filter the oil, check the seals, or check the shaft alignment for possible imbalance.

If you are considering choosing a new commercial lab, ask to see some demo reports, if available. The diagnosis might not be perfect, but it should be accurate enough to guide your staff in the initial trouble-shooting process, so you can concentrate your already-stretched resources on other job functions.

Readability of the Reports
At the Lubrication Excellence 2004 Conference in Nashville, I attended a stimulating presentation by Mark Smith of Analysts Inc. He provided 10 steps to improve one's diagnostic skills, with one of the steps being "read the report, today". Stunningly obvious, yet often ignored. There may be a number of reasons this step is overlooked, but one could be the readability of the report.

To put it simply: if the reports are unreadable, they won't get read.

Check the reports for factors such as:

• An obvious statement of condition, like "critical", "actionable" or "normal" at the top of the report.

• Obvious machine identification. Are the machine and component in question easy to find on the report, or do you have to hunt around in the small print to find out what machine they are talking about?

• Are statements of diagnosis and prognosis easily readable, or are they in a cryptic code that needs a key to decipher?

In a nutshell, in a few seconds, one should be able to determine the following:

• Is there a problem?

• If there is a problem, where is it, what is it, and what actions need to be taken to remedy it?

A common error is to ignore reports that indicate "normal" at the top. Just because it says "normal", does not mean the report should be disregarded. It might contain information that should be noted. The readability of the report should contribute to reports not only getting read, but read on time as well.

Depth of the Report
The readability of the report relates to the first few seconds of viewing it. The depth of the report relates to the information density in it.

I have personally been involved in designing the structure of a report. A lot of information needs to be presented on a single page, or possibly two, and it's a task not easy to perfect. Some labs do it better than others. Here are some factors to consider:

• Are new-oil baselines represented in the report? For those not accustomed to studying oil analysis reports on a daily basis, it's difficult to determine if 100 ppm or 500 ppm of phosphorous is OK. If the new-oil baselines are reported by a comparison process, the decision process tends to be easier. Of course it is the customer's responsibility to supply the laboratory with new oil data and, even better, a sample of new oil for analysis.

• Are limits represented? Nonconforming data should be easy to identify. Ideally the limits should be represented next to or above the parameter in question, and nonconforming data should be highlighted or underlined for easy visual identification.

• Is information presented graphically? Graphs emphasize the severity of nonconforming data better than tables of numbers do, and it is even better if the limits are overlaid on the graph. Graphs are particularly useful in comparing related parameters, such as elemental analysis of iron and ferrous density in the case of wear, or silicon and aluminum when referring to dirt.

• If microscopy was performed, is an image presented? Not everyone knows what he or she is looking at when confronted with an image of a severe-sliding wear particle, but even to the untrained eye, such a condition does look bad. Including a photo or illustration of one on a report attracts attention, which helps get things done. As the saying goes, a picture is worth a thousand words.

Range of Tests
A good lab will also be able to provide a wide range of tests. Some will be routine tests, but a substantial number of exception tests should also be available. It's advisable to have a good working relationship with your lab. If you have a concern about a particular result, then you should be able to call the lab to ask "what now?", and there should be a Plan B in place.

Test Methods
When used-oil analysis came onto the scene, there were no tests designed specifically for it. Therefore, some laboratories took existing tests for similar applications (like the manufacturing of lubricants) and adapted them for their own use, while others created their own procedures from scratch. As a result, unless paying for a certified ASTM or ISO test procedure, the obtained results may not have the accuracy required.

My personal opinion is that for most (but not all) applications, oil analysis doesn't need the full accuracy that the ASTM or ISO tests provide. But one should know that a test method used is a closely modified version of an accredited one, or alternatively, the repeatability of the test should be known. This information can be either requested from the laboratory or tested with independent control samples. If the lab is tested with independent control samples, make sure the samples are identical. For example, don't open the drain port and merely fill up three sample bottles. Instead, open the drain port, drain a quart of oil into a container, thoroughly agitate the oil and then dispense immediately into the sample bottles.

Another problem I've encountered with oil analysis users is they don't know which test they should use. For example, there are many ways of testing for moisture content in an oil sample; some are relevant to specific applications and others are not. There is no easy way to get around this - just because the lab says it's OK doesn't mean it actually is.

Flexibility in Test Schedules
Flexibility in test schedules is another important factor. Most commercial laboratories have standard testing packages for different application types, such as hydraulics, gearboxes and turbines. In many cases, these standard test slates will be acceptable for general needs, but occasionally one might want to adapt the test slate to better suit specific needs. It is reasonable to expect a slight cost penalty with this, but ensure this can be done without incurring excessive cost and administrative overhead.

Integrating Commercial Lab and On-site Results
The best oil analysis programs use a mixture of on-site and commercial lab testing. If using on-site oil analysis, the program likely contains either a commercial or in-house-designed software package to manage it. Ensure that you can easily integrate the results of the two. This might require designing software filters to integrate the two information streams.

Geographical Location
Having a lab close by, especially within easy driving distance, is extremely useful. When that urgent sample crosses your desk and you think you need results fast, you probably do. Sometimes this is not feasible, but if there are labs in close vicinity, consider them first.

Price
I purposefully left this factor until last, because it should be the final thing to consider. A "big fish" oil analysis could potentially save more than hundreds of thousands of dollars, and pay for the entire oil analysis program for years to come. Like everything else, oil analysis is simply a case of "you get what you pay for." Unfortunately, choosing quality over price is not always easy, because many business decisions are made based on cost. Try to focus on the quality of service and expected benefits as much as possible.

Oil analysis can be viewed as a tool to direct resources for emergency repairs, a tool to pin the blame on others, or a tool to measure the health of a lubrication management program (the latter is preferred). I've even seen it used as a tool to ensure machines even had oil in them ("Aw, if they got some oil out of it, then it must have oil in it"). However, if you view oil analysis as a proactive maintenance resource, and hopefully you do, then choosing a commercial laboratory should be as important as choosing a health-care provider. Would you choose a pathologist who does only half the job?

Reference

Smith, M.K. "Oil Analysis Reports: How to Improve your Diagnostic Skills." Conference Proceedings, Lubrication Excellence 2004. Noria Corporation. Tulsa, Okla.