"Apart from the data on the multipass filter test, what other ISO standards exist for the evaluation of filter performance that I can use to objectively select a filter?"

The ISO standard for multipass testing (ISO 16889) has changed to require filter manufacturers to publish their data for the beta ratios across a wide size range, that is, at 2, 20, 75, 100, 200 and 1,000 microns.

Historically, filter beta ratios were published at either 75 or 200 with an absolute rating (per ISO 4572). The new standard gives a better interpretation of the filter's overall performance, and can be better compared against competitor offerings.

In addition to the new multipass data, consider the following data:

ISO 2941 - 1974 - Verification of collapse/burst resistance.

ISO 2942 - 1994 - Verification of fabrication integrity and determination of the first bubble point.

ISO 2943 - 1998 - Verification of material compatibility with fluids.

ISO 3723 - 1976 - Method for end load test.

ISO 3724 - 1976 - Verification of flow fatigue characteristics.

ISO 3968 - 1981 - Evaluation of pressure drop versus flow characteristics.

These additional tests are useful for comparison where the installation may be an unusual lubricant, an extreme contaminant-sensitive workpiece, a high-pressure fluid power system or a safety-critical application. When selecting for a noncritical system with no unusual requirements, evaluation of filter suppliers' multipass data, in addition to other commercial criteria, will normally suffice.

Baffles are an important component in a hydraulic reservoir. They allow the fluid time to cool, deaerate and to settle out water and dirt. A good rule of thumb for residence time in a reservoir is 3 to 5 times the pump output. If the system is highly contaminated, residence time may be 10 times the GPM (gallons/minute) of the pump.

Submit your tip.

As little as 50 ppm of sodium in used crankcase oil can mean more than one gallon of coolant has leaked into a 10-gallon sump.

Gas chromatography, FTIR and glycol test kits (Schiff's reagent methods) can't detect glycol reaction products. This can lead to a false negative or an understatement of concentration.

Aldehydes in new oil can produce a false positive reading when using the Schiff's reagent test for glycol.

It is best to look for sodium and potassium because boron compounds tend to boil off, and because some engine oils are formulated with boron-containing additives.

Some aftermarket antifreeze formulations don't use organo-metallic additives (those containing sodium, potassium and boron).

Be sure to baseline coolant metals for purposes of detecting coolant leaks.

One study of more than 100,000 engine oil samples (on- and off-road applications) found 8.6 percent showed glycol contamination. In another study of 37,000 long-haul truck engine oil samples (11,000 trucks), 16 percent had trace amounts of glycol while 1.5 percent had significant amounts.

Source: Noria Effective Contamination Control training course.

"We have a filter that appears to have been in service for nearly two years without any indication of blocking, is this normal?"

While contaminant ingression varies according to environmental and production/maintenance activity, some filter manufacturers specify a life of six months. It is generally a good strategy to take an upstream and downstream particle count to ascertain if the filter is still working properly; if not, replace the element.

Consider the maintenance history and try to establish an average life for previous elements as this will give some indication if two years is abnormal.

However, there are a number of ways in which the filter may fail without tripping the indicator, and this may lead the unwary to believe the filter is good value. Without a particle count program, it is almost impossible to determine a failed filter. Such failures as a collapsed element, split pleat, poor adhesion on the seam, a sticking by-pass valve or even a faulty indicator are typically not detectable without a particle count program to monitor changing contaminant levels.