How To Collect An Oil Sample

By Doug Salvesen

Editor’s Note: We received the following letter from an avid and well-informed reader and felt his information was valuable in article format, so it is available here, as a Web Exclusive article.

I read with interest the article on oil analysis (see “Oil Analysis,” PMM Dec.’ 07). As a lifelong boater, this subject is close to my heart. During a 32-year career in the oil industry, I have run all the physical tests in a company laboratory; I have driven trucks that carry lubricants; I have operated blending and packaging equipment to make and safely contain lubricants; and now I sell lubricants and interpret oil analysis for a living to some of the largest consumers of lubricants in industry. I found your information wonderful and spot-on.

I would like to respectfully submit some additional thoughts to your readers, not about testing oils, but about collecting that all-important sample, the source of all this information.

As the article pointed out, the elements and contaminants we are looking for are very small, therefore any level of distortion of the sample will influence the interpretation and potentially lead to an incorrect conclusion. So in addition to finding the right company to run the analysis, the responsible maintenance technician or owner needs to be very careful and purposeful about the method of collecting an oil sample. I have seen some statistics that as many as one in five samples are unusable due to sample contamination, and even higher waste if you consider labels that aren’t filled out effectively.

The sample bottle should be kept with the top on, until the last practical moment before the sample is introduced to the bottle. A marine environment is rich with opportunities for contamination that is not part of the lubricant stream, especially engine rooms and bilges. Remember, we are endeavoring to measure very small, trace amounts of elements.

In order to get a profile of the lubricant, the sample should be obtained before any top-off of lubricant and be truly representative of the system. There are basically three methods of obtaining a sample. I’ll list them in order of preference for obtaining a representative sample for interpretation.

1. Capture the sample during normal operation, at normal operating temperature. This can be achieved safely via the installation of a sampling valve between the oil circulation pump and before the filter. There are several vendors that make durable sampling valves built for this purpose. These vendors also provide guidance on where to install these devices along the oil gallery to safely retrieve a representative sample.

2. Vacuum extraction. This method can be used only on a system at rest (shut off). Get the system up to normal operating temperature for several minutes, preferably after a voyage of some time. Shut down the system. Then, using a vacuum pump, extract the sample from the fill tube or breather (if equipped). Ensure that you don’t draw sludge or contaminants off the sump bottom or walls—the lubricant sample needs to be a mid-level sample for effective analysis. The bottom sludge may be interesting to understand but will likely not contain representative amounts of the materials circulating in an operating system. (Sludge in a sump or anywhere in the engine will act as an insulator, reducing heat transfer, but only a finite amount will dissolve in the lubricant, and that is what we need to better understand and identify.) The tubing used to snake down into the sump should be kept off the bottom, either by plugging the bottom of the tube and making a hole a few inches up the tube to ensure a mid-range sample, or if room permits, affix the tube to a stick or rod leaving the tubing a few inches short of the rod. When the rod touches the bottom of the sump, your draw tube will be off the sump bottom.

3. Finally, the least desirable and most variable method is the “grab sample” during system drain. It is difficult to determine the middle of the sample, it is messy to grab the sample, and if the oil system is at operating temperature, it will be very hot. This method has the highest degree of variability, the greatest safety risk to the person collecting the sample, and a risk of contamination from both inside and outside influences.

Whatever method you use to collect the sample, try to keep the container contaminant free and use the same method each time you sample. If the system lubricant hasn’t been changed, you may need to resample if the analysis comes back with unusual results (alerts or cautions). If the system has been changed, sometimes a resample can either verify the concern or help ease your mind. Some high readings have reasonable explanations. For example, engine builders are using more exotic materials in coolers and coatings on air pathways that can show up in oil analysis very early in the engine life, until these metals pacify or become oxidized. In the case of some copper based coolers, a tiny layer of copper often dissolves into the lubricant, and the resultant first few samples may contain high copper levels that would require further research or explanation from a well-informed source.

Many engine builders put oil analysis interpretation guidelines on their websites, using new oil values as a baseline. I suggest you consider sending in a new oil sample to your lab of choice so you understand new oil values. For example, if you track Total Base Number (TBN—a test that reveals the ability of the oil to neutralize acid formation in the engine), many original equipment manufacturers would recommend that a 50 percent reduction from the new oil value is cause for changing the fluid, etc.
Finally, your oil manufacturer and engine builder often have free advice available for the asking. Some engine builders and oil companies have their own oil testing program available through select dealers and they have technicians available live via phone or the web to answer questions, at least in the United States during local business hours.

Oil analysis is very much like a medical test. A good sample taken under optimal conditions by an informed technician can be trended effectively to help diagnose a problem. For example, if a patient has high blood pressure, a doctor will probably monitor blood pressure again to verify the condition, and this usually kicks off an investigation into reasons why the patient’s blood pressure is high and may result in other tests. The oil analysis is confirmation of a symptom and other tests are needed to find the true root cause. So in summary, if the sample is not truly representative of the lubrication system, we may spend significant time and money looking in the wrong direction.

I hope this information adds to the overall understanding of how oil analysis can be a predictive maintenance tool if obtaining a sample is done properly.

I love reading PMM, and appreciate the broad base of knowledge the staff brings to the readers with each edition. Keep up the great work!

Doug Salvesen
Sterling, Virginia

About The Author

I have worked in the oil industry for 32 years. I started by performing physical and analytical testing of oils. I have held positions of service and some positions of authority in a large oil company, and have made my life’s work helping customers with oil analysis of machinery great and small. I am a lifelong boater, but until retirement, mostly dayboating. Any passagemaking has been as a crew member rather than my dream of being skipper someday soon. Married for 28 years to a sailboat racer, with three children, we all love the water and reading PMM.