Despite the great increase in the awareness of lubrication technology over the last few years, there remain many misconceptions about lubrication, and the testing of lube oils for machinery. We are launching a regular article here on our website, where we tackle some of these urban legends, as well as comment on the latest lubrication related trends.
— Current Case File —
Advanced Particle counting- when results are too Consistent
Q: I see that BTS offers advanced particle counting using the Laser Net Fines particle counter. I notice occasionally that my particle numbers in the larger size ranges, particularly > 70 microns, are sometimes identical. That’s statistically impossible! What is going on?
A: The Laser Net Fines (LNF) automatic particle counter is very different than traditional laser particle counters in how it detects particles. A powerful laser produces a short burst of light to produce an image that is magnified four times onto a video chip and each particle is captured as shown in Figure 1. Each particle measured is represented as a collection of pixels and can be displayed to differentiate different shapes/types. Essentially the system takes a series of pictures, and quantifies the shadows, similar to traditional light blocking sensor diodes used on older particle counters. The similarity ends there though. BTS is able to generate a morphology report to distinguish the type of particles generated and is useful as a screening tool. It still needs refinement however, before we consider it as an automatic wear debris analyzer.
 |
Figure 1 - LNF Particle Resolution onto CCD |
Your central question is why the abnormal repeatability? This will occur when particles of size interest are normalized to 100 ml volume. This only occurs when there are little to no particles present. Here is an example of the process:
As stated earlier, the laser produces a burst of light 2 µsec in duration at a frequency of 30 times a second (Hz). This pulse will “freeze” the particles in the flow cell (a snapshot), which will be magnified 4 times and focused on a CCD video chip. Each picture or frame corresponds to a small volume of oil equal to the viewing area multiplied by the cell thickness. Figure 2 is a representation of the flow cell where the particles are photographed.
 |
Figure 1 - Flow cell |
The calculation of sample volume that passed through the flow cell is a function of the number of frames (pictures) that the LNF has taken.
One frame has a volume of:
1200 µm X 1600 µm X 100 µm (Cell length X Width X height) = 1.92 X 108 µm3 or [1.92 X 10-4 cm3] of volume per frame
However 1 mL of volume is equivalent to 1 cm3
Therefore Number of frames per 1 mL of fluid is:
1 cm3 / [1.92 X 10-4 cm3]/frame= 5208 frames in 1 mL of fluid
If 3.1 minutes is the selected processing time (typically used by BTS):
186 seconds X 30 frames/sec = 5580 frames
However, We know that there are 5208 frames in one mL therefore the LNF has seen slightly more that 1 mL of fluid. It should be noted that some frames are lost for other processing tasks and other frames are rejected due to flow lines and other errors. Typically the volume is 1.0 mL for 3.1 minutes of processing time.
Finally, particle counts are normalized to 100ml for each size range (consistent with particle count standards). If only one particle was observed in the cell in the >70 micron range, and the volume used was 1 mL then for the calculation would be 100 particles per 100 mL. If the volume observed is 0.65 mL? (Often used for contaminated oils) The factor would be as follows:
1 particle/0.65mL = X particles/ 100 mL
Therefore the number of particles is 100 /0.65 = 154 particles.
If two particles were observed the results would be 308 particles per 100 mL. Therefore values may repeat when using trend analysis due to the small number of particles observed by the LNF at the higher size ranges.
- Chuck Goudreau
Reference: LNF Manual - Personal correspondence with D. Anderson, Spectro Inc.
If you have a question about oil or oil testing, email it here.