5 common mistakes to avoid in presentations and papers

With many conferences beginning to restart in person following the pandemic, we’ve noticed the return of a few classic mistakes in conference presentations and papers. In this page, we give a short review of the main errors that authors should avoid in order to provide sound and replicable surface texture results.

1. Never report parameter results without filtering conditions

All too often, scientific papers report surface texture parameters (usually profile parameters such as Ra or Rq) without any filtering conditions. Although ISO 4288 or ISO 21920-3 admit default conditions that can therefore be omitted, they are not always applicable, due to the small size of the measured parts.

For clarity, parameters should always be provided with the cut-off value and type of filter used. Otherwise, one can obtain any required parameter value simply by tuning the cut-off value, which makes results totally meaningless!

Moreover, with the new ISO 21920 standard, parameters are no longer averaged leading to differences in values, so the reference standard used should always be stated in the paper.

Parameter tables in MountainsMap® automatically give the filtering conditions in the header:

Read more:

   Default specifications in international standards
   Profile parameters of ISO 21920-2:2021

2. Do not limit analyses to Ra or Sa. Explore other parameters.

Surface texture is too often expressed simply as Ra or Sa parameters. Before jumping on the Ra bandwagon, ask yourself: what am I trying to achieve? What kind of surface function do I want to characterize?

If you’re looking to analyze average roughness height, use Rq instead of Ra, because it is more closely related to physical phenomena.

If you want to characterize a particular function or performance on a surface, try exploring other parameters that will likely be more “pertinent” and better correlated to the process. In particular, parameters related to height distribution (Rsk and Rku) or to material ratio (Rmr, Rdc) are usually more significant. Many other parameters offer useful insight, for example Rk, Rdq, Sdr, etc.

Read more:

   About surface function
   Parameters of the new ISO 21920
   Functional parameters of ISO 13565

3. Avoid using spacing parameters on stochastic profiles

The most common spacing parameter, Rsm, is sometimes used in contexts where it cannot provide sound results. Spacing parameters are Feature parameters that should only be used on periodic, pseudo-periodic or structured profiles and not on purely stochastic profiles. Another useful spacing parameter is AR from the motifs method (ISO 12085). In both cases, filtration should be chosen carefully in order not avoid removing wavelengths that contribute to the period.

Read more:

   How to characterize lateral features?
   Profile parameters of ISO 12085 (motifs)
   Stochastic vs deterministic surfaces

4. Avoid using default cut-offs

Users are sometimes lost when they have to select cut-offs, and this is why default values are used so frequently. But it is important to remember that filtration is an operation that affects results and even changes the conclusions of a study. The cut-off value depends on the material or process signature on the surface. It cannot be chosen only with respect to instrument capabilities or measurement length.

Users should carry out explorations to determine at which scales or which wavelengths their material or process will exhibit characteristics of interest. The chosen cut-off value may not necessarily be the default one, nor the same as that used by another user working on another application.

Read more:

   Filtration techniques
   Bandpass filters

5. Do not report Sdq or Sdr without stating surface resolution

Some parameters are calculated with respect to neighboring points, which means that depending on the spacing (lateral resolution), they provide different values.

This is indeed the basis of scale-sensitive fractal analysis (SSFA) where Sdr varies with scale. This applies, in particular, to Rdq/Sdq (mean slope) and Rdr/Sdr (developed area), and even on Spc/Svc (curvature). These parameters are very sensitive to resolution and to noise.

Read more:

   Scale-Sensitive Fractal Analysis
   Areal field parameters


   Always provide filtering conditions alongside parameter values.

   Take the time to think about which parameter is most appropriate for your application.

   Only use spacing parameters on (pseudo-)periodic or structured surfaces

   Investigate at which bandwidths your study is relevant.

   Sdr and Sdq parameters should always be accompanied with lateral resolution indications.