Particle Counters - The Good, the Bad and the Ugly

By Martin Williamson

Anyone who has attended any class with a proactive maintenance focus knows that a particle counter should be at the top of the shopping list when setting up a "mini" onsite analysis program. In fact, if you're already operating a vibration-focused monitoring program, or have established a good relationship with your filter supplier, you've probably been offered a particle counter.

However, after having managed oil analysis programs and been consulted on the establishment of these programs in a variety of industries, it is clear that choosing a particle counter is a tough decision. In fact, after supporting them technically, conducting wide-ranging research and serving as a product development manager on these units, it is apparent that the market is a minefield for the unwary. Thus, many factors should be considered when purchasing a particle counter.

History

Automated light blockage particle counting technology was first introduced in the late 1960s. It existed in markets where process material sizing was required, and some of that technology carries over to today's oil analysis technology.

For example, the original automatic particle counter (APC) units were calibrated using latex spheres, which lead to software that calculated shadow areas as circular dimensions. However, latex spheres cannot be used in the calibration of APCs because the latex swells in the test oil and creates inaccuracies. And so the test dust used is certainly not obligingly spherical creating difference in measured dimensions against the actual random shapes.

Indeed, an oil sample may contain a multitude of problems, which may obscure the goal of counting the solid particles. By treating contaminants such as water and air as shadow- or light-refracting sources, they are included in the overall count.

Figure 1 - the principle of light obscuration Ð shadows are cast onto the light sensor as particles pass through the light beam.

Furthermore, light blockage does not work well with fluid that is dark, multi-phased (for example, water-based glycols) or heavily contaminated with silt or soot. This may cause coincidence error or actually prevent laser beam transmission. CSI have been instrumental in developing techniques to eliminate water related errors, and research is ongoing to allow use in water based glycols.

Figure 2 - the principle of mesh obscuration showing change in differential pressure at constant flow.

As an alternative approach to historical light blockage technology, Dr. Trevor Hunt at Bath University and James Fitch (now President of Noria Corp) independently patented variations of mesh or pore blockage measurement. The focus of the measurement remained purely on the solid particulate, irrespective of the conditions or type of fluid in which the measurement took place. Because of the technology, this actually resulted in two very different measurement instruments with a similar level of performance.

Commercially, the two methods have created some confusion in the marketplace. Customers want to know which is the better technology - optical or pore blockage? While optical particle counters remain unhindered by patent issues, not all units offer the same performance or specifications. Moreover, because patents restrict the availability of the two pore blockage units, conflict exists at the marketing and standard's level as to whether pore blockage constitutes particle counting or solid contaminant trending.

What Fits the Bill?

If you are considering purchasing one of these units, you need to make the following determination: Is it a particle counter you need, or a trend monitor?

If you are conducting research, or are required to certify a fluid's cleanliness by a particle count, then you definitely need a particle counter. As defined in the ISO standards, a particle count requires a measure of several different sizes of particles. This is particularly true in the case of the original NAS 1638 (now SAE 4059), as it defines a number of size brackets to be measured, and requires a particle count via microscope or optical APC.

However, if you are conducting a proactive onsite oil analysis program, then a particle count, per se, is superfluous. A trend monitor measures the overall level of contamination sufficiently to identify a potential harmful situation. Purists may argue that a full particle count is required in order to identify the failure path - an increase in larger-sized particles (>50um) is indicative of the onset of failure as opposed to an overall increase measured at >6um and >14um, or simply at >10um. Any general increase in particles noted by a trend monitor may warrant immediate action, with more advanced sampling tests conducted later by your off-site laboratory.

You also need to consider the applications you intend to monitor. Optical particle counting is influenced by a number of factors - aeration, water, dark fluids, heavily contaminated fluids and sample preparation. In fact, agitating a sample prior to measuring will induce aeration, unless it is gently rolled. Furthermore, water can be eliminated by preparing the sample with a solvent or by drying it, and air can be removed by pressuring the sample, or by placing the bottle in a vacuum or an ultrasonic bath. However, these additional measures increase preparation time and equipment costs.

In online terms, optical APCs are best applied to hydraulic fluid power systems and turbines. In this case, either the pressure and/or temperature generally preclude the issues of water and aeration, and are typically clean enough not to cause coincidence error on the unit's light source. Furthermore, optical APC manufacturers argue that counting the water and aeration is important when identifying the contamination level, but this is a tenuous point given the definition of a particle count and the attendant international standards on particle counting. Work has begun to ascertain the level of water contamination from the data using algorithms that recognize the impact of the water in a sample. However, given the low cost "water-in-oil" test devices available, this approach may be unnecessary.

More recently, laser technology has taken the place of incandescent light units, and is thought to be superior. However, when using an APC online, it is necessary to allow sufficient flushing time to ensure the unit's sampling hoses are thoroughly clear before accepting a reading. In fact, on very clean systems, it is possible to see results starting at ISO -/19/16, dropping rapidly to ISO-/11/8 to as low as ISO -/6/3 after several hours of online monitoring at the same sample port. While it would be too much to expect to wait two hours for a reading from the average on-site oil analysis program, if a permanent mount inline or online unit is used on hydraulics, very clean readings would be the norm compared to those typically expected from bottle samples. Furthermore, there are very few laser sensor OEMs, so despite the variety of units available, so most units from varying manufacturers usually share a common laser cell.

Pore blockage particle monitoring is best applied to a proactive monitoring program where numerous applications are involved, from gearboxes and bearings to diesel engines and paper machine oils. Since pore blockage is not affected by aeration or water (unless it includes large pockets of free air and water), or the oil's colour and contamination level, it is very consistent under a wide variety of applications. These units work by measuring the rate of change of the differential pressure across a mesh at a constant flow rate, or the decay in the flow rate at a constant pressure. Then they interpret the number of trapped particles. This obviously includes particles smaller than the measuring screen size trapped in the interstices of the larger particulate, but by interpolation of the pressure or flow curve against a calibration value, it can realistically estimate the particle distribution.

The constant flow unit made by Pall Corp uses two screens to establish a measurement greater than 6um and 14um as per the ISO 4406:1999 code, although earlier versions of the Pall unit used 5um and 15um screens. However, the constant pressure unit sold by Rockwell Automation Entek uses one measuring screen of an applicable size (5um, 10um or 15um) for the viscosity of the sample tested and extrapolates the data against its calibrated curve to provide an estimated particle count. From a user perspective, the latter device is ideally suited to condition monitoring. It has the flexibility to be used directly online on an oil test port with a minimum 2 bar pressure, or with a small bottle sample (less than 100mL) extracted from the system.

Because it needs larger volumes of fluid for analysis, Pall's constant flow unit with onboard pump is better used online, either at the tank or on a pressure line. Through automation it can be operated successfully by computer or PLC control, which makes it ideal for research on test beds. While both units are generally restricted to an ISO -/11/8 as their lowest measurement (with the exception of a modified constant flow unit), they also work with very clean fluids.

Contamination Control

If particle counting is being considered as an extension to a vibration analysis (VA) program, the VA hardware and software supplier may already provide additional oil analysis tools to include a particle counter or trending unit, with the necessary interfaces into their condition monitoring (CM) software. Interfacing the cleanliness code within CM software can be a precursor to more alarming problems, which can be solved by advanced oil analysis or vibration analysis. If your supplier's unit does not offer this type of interface, be sure the supplier is willing to provide the output data from the instrument so you can import it into your CM software. (Exporting in a comma delimited format such as *.csv is very useful.)

You should avoid using a unit that has its own software for storing measured data, as it can be cumbersome when you try to interpret data signal patterns on a single screen against trends and results from other sources such as vibration analysis. Of course, a PC interface requires the unit to have a communication port, preferably one easily connected between the computer and APC to upload results. While these ports are typically an RS232 style, they limit cable length and require a free serial port. In the past the serial port was often already in use by another device, but USB connections have eliminated this problem. That said, many PCs no longer offer an RS232 port so the unit needs to have USB connectivity, preferably.

In addition to data upload, it is important that you have the ability to enter a fixed identity such as a bar code or radio signal from the oil analysis sample port into the unit's memory while conducting a test. This minimizes errors and results in the accurate uploading of data without accidentally creating new identities.

In terms of onsite analysis, the choice is obviously between bottle samples and direct online measurement - permanent or otherwise. It is useful to have both capabilities, of course, as other routine oil tests (i.e. water and viscosity) can be conducted on the same bottle sample. If bottle sampling, make sure there is a suitable attachment for preparing the sample (agitating and de-aerating), and that the work area is suitably clean. Bottles may need to be super clean for critical or very clean systems.

On the other hand, a permanent mount unit is justifiable for a critical plant. For walk-around measurements, primary and secondary connections to a system allow you to immediately repeat tests of abnormal samples, and possibly troubleshoot the problem (such as inspecting filters and breathers). This will require quick connect fittings, and some units may even require a custom manifold. Most units, however, work with the "Minimess" style fittings, but be aware of the correct places to mount these - avoid long, stagnant pipe lengths and laminar flow zones, or areas where significant pockets of air may exist.

If no flow lines exist, it may be necessary to purchase a unit with a built-in pump that can extract a sample from sumps and tanks. While these models have a drain line, they may not reach a suitable disposal point in the system, so a small waste container may be necessary for the flush and sample fluid. You should avoid tipping this fluid back into the system unless the container is immaculately maintained and there is no cross-contamination risk.

Roll-Off Cleanliness and Process Quality Control

If you need a particle counter to monitor roll-off cleanliness, or for the quality control of process fluids and machine tool fluids, then a very different set of parameters exist. A portable online unit would be the most appropriate because it can be carried to different sites. However, because flushing - by its very nature - usually involves a lot of turbulence and, thus, creates aeration problems, many OEMs in the fluid power business prefer pore blockage devices. In fact, because water-based glycols used in the offshore business generally preclude the use of optical APCs, pore blockage devices usually dominate.

However, the use of a pore blockage device in this instance does not constitute a particle count, and the unit should only be considered as a flushing guide to indicate when a sample can be drawn for a microscopic patch particle count. For these purposes, make sure the unit is sufficiently portable, rugged and can print out data (or at least store it to upload at a later time). The requirement here is similar to a Condition Monitoring unit outlined above, but the storing of data is somewhat different to a trend of a particular machine. In this instance, the proprietary programs often provided with a unit are usually sufficient; but if there is a requirement to store the data elsewhere, make sure output data is available.

In other applications, such as on test stands and manufacturing rigs for quality control purposes, OEMs often use a particle measuring device to determine if the fluid is sufficiently clean to meet the standards (i.e., to flush cylinder heads or fill sealed-for-life sub-assemblies). This requires a totally automated unit, so control of the instrument via PLC or PC is crucial, and the communication protocol is another important consideration. Furthermore, given the number of tests these units can deliver in a short time, extended reliability and service interval are important. In the civilian aerospace industry where Skydrol fluids are used, a special unit is required with special seals and materials.

And as with any unit - irrespective of its planned use - you should consider the location of your service center, its ability to quickly repair or service the equipment, or to loan out equipment if necessary. Also consider the operational costs, the frequency of calibration and whether a calibration or validation can be done on site by a competent technician.

It is only after considering all of these factors that an engineer can decide on the particle counter best suited to its needs.