Printed Circuit Boards (PCBs) have become ubiquitous in today’s technology-driven world, powering the functionality of countless appliances. Protective coatings play a crucial role in enhancing PCB durability and longevity. In this study, we show that hyperspectral cameras are suitable for monitoring the thickness of the coating and its distribution uniformity.


NIR = NEAR INFRARED (900 – 1700 nm)



NIR = NEAR INFRARED (900 – 1700 nm)


Challenges PCP Quality Inspection

The manufacturing of PCBs requires the application of a protective coating. Different protective coatings are used depending on the application and where the PCB is placed in the appliances. Typically, the coating is sprayed. Some coatings have UV fluorescence, which is one way to detect its distribution. However, it does not give information about the thickness of the coating, which should be optimized between 25 and 75 um. Additionally, not all coatings emit under UV light, further complicating the inspection process.

How Hyperspectral Imaging Can Ensure Precise Protective Coating Thickness?

We tested the suitability of a Specim FX17 to monitor the thickness of the protective coating on PCBs. The Specim FX17 is a line scan hyperspectral camera that operates in the near-infrared (NIR) range of 900-1700 nm.

We investigated two types of coating.

  • Plastic spray PRF 202
  • Electrolube SCC3UL, which is a UV lacquer

Those were applied on layers (1 – 5) on five PCBs.

For each measurement, there were:

  • 5 PCBs with 1, 2, 3, 4, 5 layers (from top to bottom)
  • 1 reference sample (top right) without any coating.
Figure 1: PCB samples on the tray

Spectral Analysis

In Fig.2, the effect of the coating on the spectra is presented. It appears that:

  1. Both coatings have unique spectral signatures, which means they can be sorted from each other
  2. The thicker the coating, the deeper the related spectral peaks are. Based on this, we built a regression model to separate the samples based on the coating layers.
Figure 2: False RGB of the PCBs with effect on spectra (reflectance data).

Modeling – PLS regression

We built a PLS regression model for each coating type to quantify the number of layers. The regression variable was called “coating,” ranging from 0 to 5 (corresponding to the number of coating layers). Notice that with SpecimINSIGHT software, the regression can be built on samples only, regardless of the background.

Figure 3: Predictive performance of the “Coatings” regression models.


This study shows that Specim FX17 is an ideal tool for monitoring the homogeneity and thickness of protective coatings on PCBs. It is also a relevant camera for accurately quantifying the number of coating layers applied.

Given their widespread use in various appliances, PCBs’ quality is critical. Hyperspectral imaging offers a cutting-edge solution for accurately measuring the thickness and uniformity of these coatings, thereby enhancing their overall reliability and performance.


This technical note was prepared by Specim, Spectral Imaging Ltd., and is for generic guidance only. We retain all rights to modify the content.

Related products:

Specim FX17