Do online oil sensors make OELCHECK laboratory analyses unnecessary?

Year of publication: 2007


We permanently filter our hydraulic oil with 5μ filters in the bypass flow. Our particle counting, which we perform on site with a mobile device, shows correspondingly good cleanliness classes.

But how can it be explained that the wear elements such as iron, copper and chromium are increasing in the OELCHECK laboratory reports, although the purity classes are consistently good or even improve from sample to sample?

Is there a correlation between the values for wear, impurities and particle count?

OELCHECK answers:

More and more often, oil sensors are being touted for continuously measuring oil quality, operating online and being integrated directly into the oil circuit. Based on usually just a single measured value, they are supposed to report irregularities detected in the oil flow and help reduce damage to a minimum or find the right time for an oil change. Such sensors use the oil as a source of information. They are expected to replace much of the laboratory analysis that can only be carried out offline. But evaluating the condition of oil and machinery based on sensor data is a very complex issue. OELCHECK determines about 30 measured values for the diagnosis of a single oil sample. Sensors that record only one or a few measurement parameters hardly allow an accurate assessment. At best, they are suitable for displaying a trend. However, since the sensor output does not correlate with conventional laboratory data, there are no absolute limits for a meaningful interpretation. A conventional laboratory analysis combines a large number of individual values. It not only evaluates the complexity of the oil change based on viscosity change, oxidation and additive degradation, but also shows the type and amount of impurities and of wear metals. Experienced experts interpret the results and make diagnoses.


However, each laboratory analysis is only a snapshot of the past, because a certain amount of time always elapses between sampling and the result, despite 24-hour service. OELCHECK is convinced that an increased use of online oil sensors does not mean less laboratory analyses, but that such sensors will stimulate the sensitivity of maintenance personnel with regard to oil as an important information carrier. We are interested in online sensors helping to avoid oil analyses that show an unproblematic oil condition. This is because experienced tribologists then have more time to use their expert knowledge to deal with problematic samples.

Current state of sensor technology

OELCHECK is involved in several projects funded at national and international level in the development of sensors. Various sensor manufacturers use laboratory analyses to verify and correlate their sensor values or have their sensors tested directly at OELCHECK using a variety of different oil samples.

The sensor systems offered today work online directly in the oil or they are immersed in the oil or wetted with oil as portable onsite devices. These sensors, which usually only work for one type of oil or one measured value, are not really cost-effective with a purchase price of often more than €1,000.

Almost without exception, they measure individual values that unfortunately do not correlate with the physical or chemical analysis methods known from oil analytics. This means that there is no direct reference to the established limit and warning values. Calibration of the sensors is usually quite time-consuming and prone to failure. During operation, deposits on the sensor surface or air in the oil lead to problems. In most cases, the informative value is limited to only one type of oil or a specific unit. Inexpensive oil sensors only have the possibility to provide information about oil parameters by changing electrical parameters such as voltage or frequency. They require a reference value as an initial condition and then display the current oil condition as a change.

The unbeatable advantage of online sensors, however, is that they are constantly in use and report any change immediately.


Automakers have made several attempts to use conductivity sensors to determine the condition of engine oils and the optimum oil change interval. But despite evaluating thousands of measurement results, they have not been able to establish a direct link to oil condition. They are often referred to as QLT (quality level temperature) sensors. They use the change in electrical conductivity, often referred to as permittivity, εr, tan δ or dielectric factor, to measure oil quality, oil level in the pan, and oil temperature. Today, they are often used only to monitor oil consumption.

In practice, it has been shown that the measurement of conductivity is not sufficient to describe the oil condition with sufficient accuracy. It is not possible to represent a single parameter that is important for assessing the oil condition. In addition to the oil temperature, the conductivity of an oil is also influenced by impurities, wear metals, mixing with other oil, change in additives, soot content, water content, fuel content, acid reaction products (tribopolymers), change in viscosity or oxidation, in some cases in opposite directions.



Viscosity is the most important physical parameter of an oil. However, it alone cannot be used to assess either the oil quality or the further usability of the oil. Only if the viscosity change can be determined precisely enough within a relatively narrow tolerance range is it possible to make a statement about the oil condition under specific operating conditions. An increase or decrease in viscosity is not only dependent on oil oxidation due to temperature and running time. In most cases, viscosity increases as a result of oil aging. However, depending on the oil type, viscosity can also decrease due to shearing of viscosity index improvers or mixing with other oil types. Refilling with the wrong type of oil can cause both an increase or decrease in viscosity. Even an initial value assumed on the basis of fresh oil viscosity may already be incorrect if, for example, too much low-viscosity flushing oil was left in a gear when it was delivered. Various newer low-cost sensor types are said to be able to measure the "relative" viscosity change. For this purpose, the "shear transducers" based on oscillating crystals or silicon chips measure with an extremely high shear rate a change in the viscosity-dependent resistance in an oil or additive layer that is only a few nanometers "thin". Therefore, for almost all oil types, the measurement result is significantly affected by deposits of oil aging products, strongly polar acting additives or VI improvers. In addition, the high shear rate due to the high frequency at which the molecules are excited can cause heating of the oil in the contact area between the sensor surface and the oil. This results in an erroneous measurement that can hardly be correlated. Comparative measurements in the laboratory have not yet shown any usable results with such shear transducers. Tests with acoustic wave sensors promise more success. They use sound waves to vibrate a "thicker" oil film at a low frequency. The difference between the sound waves generated and those received on the same sensor chip depends on the oil viscosity. The energy absorbed by the oil is output as dynamic viscosity at the particular operating temperature. With the density, the dynamic viscosity can be converted into the kinematic viscosity related to 40°C or 100°C. The "low-shear" sensors available today now also show a relatively good correlation for higher viscosity oils (up to ISO VG 220), but this is still strongly dependent on the oil type.

Water content

Water in the oil can severely reduce the service life of the lubricated components due to corrosion formation. If water droplets interrupt the lubricating film, localized welding of the roughness peaks or increased material fatigue can occur on the surfaces in addition to corrosion.

Many sensors today advertise that they can detect water in the oil. However, they usually only indicate a change in the conductivity of the oil. It is also affected by water. However, the sensors do not determine the absolute amount of water in the oil, as measured by laboratory methods in % or ppm. They only determine the "relative humidity" in the oil according to the conductivity principle. If it is completely saturated with water, a value of 100% relative humidity in the oil is displayed in the ideal case. This can be as low as 180 ppm water in a hydraulic oil, but over 2,000 ppm in a hydraulic oil type with different additives.


The saturation limit, i.e. the absorption capacity of water in the oil, depends on the air pressure, the relative humidity in the ambient air, the base oil, the oil temperature, the oil viscosity, the oil aging and also the additivation. Attempts to indicate the water content in ppm with sufficient accuracy by measuring relative humidity have therefore failed.

Element analysis

A determination of the wear, impurity and additive metals present in the oil is the backbone of conventional oil analysis. Based on the amount of elements present in the oil, clear conclusions can be drawn about the wear condition of components, contamination, mixing of the lubricant or additive degradation.

Online particle counters can provide information on the number of particles present in the oil. However, these do not distinguish between wear or contamination particles and between hard particles and soft reaction products, whose hazard potential for the aggregate must be assessed differently. To date, no sensors have been tested at OELCHECK that can make statements as a substitute for complex ICP element analysis.

Particle counting

Online particle counters are already installed today in some hydraulic systems operating with low viscosity oils (ISO VG 46 or thinner). However, the values obtained cannot be compared with the standard values according to ISO 4406. They can, however, be well suited for trend monitoring. But even such sensors should be recalibrated for each system and each oil type. Depending on the principle, however, offline devices or online sensors will not work if the oil contains many air bubbles or water droplets, or if hydraulic systems in mobile use are exposed to strong vibrations. Even when online particle counters are used to determine the purity of gear oils, they fail either because the oil viscosity is too high (ISO VG 320) or because the oil is too dark.

Particle analysis

For the determination of magnetizable iron, methods similar to the PQ index are used especially for helicopter gearboxes and aircraft turbines as well as for ship engines. In magnetic sensor systems, the iron chips accumulate on the sensor, causing a change in the output magnetic field. However, these "chip detectors" can only detect relatively large particles in relatively thin oils and thus indicate acute wear processes. Online photography of wear particles and associated software development are already being considered. Similar to OPA (Optical Particle Analysis) in the laboratory, this should allow conclusions to be drawn about the mechanism of particle formation. OELCHECK is involved in an EU-funded project dealing with this topic.

Acid number

When deciding whether to change the oil, the oxidation state of the oil plays a decisive role. Oil becomes "acidic" when it has stored oxygen due to aging, reaction products from additive degradation, or consumption of alkaline-acting additives due to sulfur-containing residues from fuel combustion. If the oil is too acidic, even better oil care to extend the service life is no longer worthwhile, because the acids dissolved in the oil can lead to corrosive wear.

Today, OELCHECK uses low-cost sensors that operate in the range of a spectrum of visible light. Using the oil samples received daily, these sensors are correlated with laboratory values in such a way that AN or NN, and thus oil oxidation and additive degradation, can be readily visualized and compared with conventional test methods.

However, even with these sensors, there is a need to calibrate them to the particular oil type and its change over the period of use using hundreds of practical samples.


Conclusion: In various funded projects, in which OELCHECK contributes with expertise, oil samples and database evaluations, as well as at well-known companies that manufacture sensors for other areas, work is being done at full speed on the development of online oil sensors. In addition to the measuring principles described above, OELCHECK has also tested other sensors – mostly with negative results. However, an increasing trend in the use of online oil sensors can be expected in the near future. Nevertheless, such sensors cannot completely replace oil analysis in the laboratory.

A combination of oil sensors and supporting laboratory analyses is unbeatable.

  • The oil sensor or several sensors with different measuring principles are constantly online.
  • As long as the sensor signal is in the "green zone", no oil analysis is required.
  • Sensors react immediately in the event of irregularities and can thus warn of damage or prompt an oil change in good time.

OELCHECK oil analyses with an expert diagnosis are only necessary if


  • this sensor warning occurs at an unexpectedly early point in time, the reason for a too short-term change of the sensor signal is of interest.


So the laboratory analysis is performed only when it is necessary and not while the oil is still fine. It then comments in detail on the causes of the irregularities reported by the sensor. And: Oil analyses are indispensable to verify full-bodied sales statements when a sensor is first used.


Oil sensors and accompanying oil analyses:

  • improve the maximum availability of engines, machines and plants
  • allow an almost risk-free optimization and extension of oil change intervals
  • minimize the effort required for effective plant and fluid management