Chapter 53: The Automotive Industry's Migration |
A Brief Summary |
By 2005, the automotive industry had reached an inflection point. The barcode---a technology that had become synonymous with industrial efficiency---was no longer capable of fulfilling the industry's primary requirement: complete, end-to-end traceability of every critical component. The solution that had served the sector so well for two decades, Code 39, was giving way to more information-dense symbologies like GS1-128 and Data Matrix. This chapter explores why this migration occurred, the technical limitations of Code 39 that drove the change, and the broader implications for manufacturing and supply chain management. |
|
Introduction |
To understand the automotive industry's migration away from Code 39, we must first appreciate what Code 39 is and why it became so deeply entrenched in manufacturing in the first place. |
Code 39, also known as Code 3 of 9 or Alpha39, was developed by Intermec in 1974 . It was the first barcode symbology that could encode both letters and numbers, a capability that set it apart from earlier numeric-only systems like the UPC (Universal Product Code). This alphanumeric capacity was revolutionary for industrial applications because part numbers, serial numbers, batch codes, and supplier identifiers all contain letters. A barcode that could represent these characters natively was an obvious choice for tracking physical goods through complex manufacturing processes. |
The symbology's structure is simple and elegant: each character is represented by nine elements---five bars and four spaces---of which three are wide and six are narrow. This 'three of nine' pattern is where the symbology gets its name. Unlike many modern barcodes, Code 39 is variable-length, meaning it can encode data strings of any practical length without a fixed character limit. It is also discrete, meaning each character is independent and separated by an intercharacter gap, making it relatively tolerant of printing imperfections. |
Perhaps most importantly, Code 39 was open and not encumbered by patents. This meant any manufacturer could implement it without licensing fees, and any scanner manufacturer could build readers capable of decoding it. As a result, Code 39 became the de facto standard for industrial barcode labeling worldwide. |
The Automotive Industry Action Group (AIAG), founded in 1981, formally adopted Code 39 as the automotive industry's barcode standard in 1984 . This was a defining moment. The AIAG published the first industry-wide barcode standards---B-1 for barcode symbology and B-3 for shipping and parts identification labels---and Code 39 was at their heart. For the next two decades, Code 39 labels would be the primary means of tracking parts moving through the automotive supply chain. |
|
The Challenge of Increasing Information Density |
The automotive industry's reliance on Code 39 was, for many years, a success story. It worked. Suppliers could print labels, OEMs (Original Equipment Manufacturers) could scan them, and parts could be tracked from the factory floor to the assembly line. But by the early 2000s, the cracks in this system were becoming visible. |
The fundamental limitation of Code 39 is its low data density . This is a direct consequence of its design. Every character in Code 39 is made up of five bars and four spaces, with three wide elements and six narrow ones. This means a typical Code 39 barcode takes up a significant amount of horizontal space for each character it encodes. |
For comparison, Code 128---a later symbology developed to address Code 39's density limitations---encodes data much more efficiently. Code 128 uses variable-width bars and spaces and includes multiple character sets that allow it to pack more information into a smaller area. The result is that a Code 128 barcode can be 30% to 50% shorter than a Code 39 barcode containing the same data. |
|
This density problem became critical as the amount of information that needed to be encoded on a label grew. In the 1980s and early 1990s, a simple part number might have sufficed. But as supply chains became more global and regulatory requirements more stringent, manufacturers needed to encode more data. A typical automotive component label now needed to include: |
- Part number |
- Serial number |
- Date of manufacture |
- Batch or lot number |
- Supplier code |
- Quantity |
- Country of origin |
- Quality certification information |
|
A Code 39 label containing all this information could become extremely long. For instance, a serial number field alone might be limited to a maximum of nine characters per automotive industry standards , but when combined with part numbers, dates, batches, and supplier codes, the total data string could easily exceed thirty characters. Some application standards recommended that Code 39 symbols should not exceed twenty characters and should never exceed thirty, primarily because very long Code 39 labels are impractical to print, label, and scan reliably . |
The physical reality of manufacturing exacerbated this problem. Many automotive components are small, irregularly shaped, or have limited space for a label. An engine block, for example, has large flat areas where a label can be placed, but many smaller components---sensors, fasteners, brackets, connectors---do not. If a Code 39 label is too long, there is simply no place to put it. And even if there is room for a long label, it is more likely to be damaged or obscured during handling, making it unreadable. |
|
The Regulatory Push for Traceability |
The migration away from Code 39 was not driven solely by technical limitations. A more powerful force was at work: regulatory and industry-mandated traceability requirements. |
Traceability is the ability to track a product or component through every stage of its lifecycle, from raw material to finished product to end-of-life disposal. In the automotive industry, traceability is not optional; it is a matter of safety, quality assurance, and liability. |
Consider a faulty batch of brake calipers. If a manufacturing defect is discovered, the manufacturer needs to know exactly which calipers were affected, what production batch they came from, what date they were made, and what vehicles they were installed in. This information allows a targeted recall rather than a blanket one, which is significantly less expensive and less damaging to brand reputation. |
The automotive industry formalized these requirements through standards like AIAG B-17, which provides guidelines for 2D direct part marking in automotive applications . Direct part marking---engraving or etching a barcode directly onto a metal or plastic part---became increasingly common because labels could be removed, damaged, or fall off, compromising traceability. Data Matrix codes, with their small size and tolerance for damage, were particularly well-suited for direct part marking. |
At the same time, the rail industry was developing similar standards for component traceability. Companies like Knorr-Bremse, a major rail systems supplier, began requiring serialization and batch management for all critical parts. Their supplier guidelines specifically recommend using GS1 DataMatrix for application on products due to its high data density and durability . These standards are not isolated; they reflect a broader shift across transportation manufacturing toward more rigorous identification requirements. |
The US Department of Defense (DoD) also contributed to this trend with its UID (Unique Identification) program, which mandates permanent marking of all items with a unique identifier. MIL-STD-130, the DoD's standard for marking, specifies Data Matrix as the preferred symbology for direct part marking . The aerospace industry, through ATA Spec 2000, similarly moved toward Data Matrix for unique item identification. These military and aerospace standards have a trickle-down effect on the automotive industry; suppliers who serve multiple sectors must adopt the most stringent requirements common to their customer base. |
|
The Rise of GS1-128 and Data Matrix |
As the limitations of Code 39 became apparent, two alternative symbologies emerged to fill the gap: GS1-128 and Data Matrix. |
GS1-128 |
GS1-128 is a specific implementation of the Code 128 symbology that adheres to GS1 standards for application identifiers. Application identifiers (AIs) are standardized prefixes that define the meaning of the data that follows. For example, the AI '01' indicates a Global Trade Item Number (GTIN), '10' indicates a batch or lot number, and '21' indicates a serial number. |
GS1-128 combines the density advantages of Code 128 with the structured data format of GS1 standards. This means a single barcode can encode multiple pieces of information---serial number, date, batch, quantity, and more---in a way that is unambiguous and machine-readable. A scanner can decode the entire string and automatically parse it into its constituent data fields. |
For the automotive industry, GS1-128 provided an immediate solution to the density problem. A Code 39 label that would stretch across a foot of space could be condensed into a much more manageable GS1-128 label. The use of application identifiers also eliminated ambiguity. If a label contained a part number and a date, the AI told the scanner which was which, removing the risk of misinterpretation. |
GS1-128 also supports variable-length data fields with a maximum length of 30 characters total per symbol , which provided manufacturers with the flexibility to encode more complex data strings than Code 39 allowed. This was particularly important for serial numbers, which often needed to include a manufacturer code prefix plus a unique sequential number. |
|
Data Matrix |
While GS1-128 addressed the horizontal density problem, it did not solve the size limitation for very small parts. For components with minimal surface area, a linear barcode---even a dense one like Code 128 or GS1-128---simply cannot fit. |
This is where Data Matrix came into its own. Data Matrix is a two-dimensional (2D) matrix barcode that encodes data in a grid of black and white cells arranged in a square or rectangular pattern. It is read in two dimensions, meaning the scanner captures the entire pattern at once rather than scanning along a single line. |
The density advantage of Data Matrix is enormous. A 2D barcode of a given size can hold much more data than a linear barcode of the same physical dimensions. Furthermore, Data Matrix includes sophisticated error correction that allows the code to be read even if a significant percentage of its cells are damaged. This makes it particularly valuable for direct part marking, where the code is etched or engraved onto a metal surface and may be subject to wear, scratching, or contamination. |
Automotive standards increasingly recommended Data Matrix for direct part marking applications . For example, an engine block may be laser-engraved with a Data Matrix code that encodes the part number, serial number, and production date. The code is small enough to fit on a flange or boss, durable enough to survive the engine's lifetime, and readable by a handheld scanner at any point in the manufacturing process. |
The industry standards bodies, including AIAG and GS1, formalized the transition. The GS1 DataMatrix guideline, for instance, standardizes the application of Data Matrix codes within the GS1 framework, enabling interoperability across supply chains. This ensures that a Data Matrix code marked by a supplier in China can be reliably read and interpreted by an OEM in Germany. |
|
The Migration in Practice |
The migration from Code 39 to GS1-128 and Data Matrix was not a single event. It happened gradually, over the course of several years, as the industry adapted to new requirements and new technologies. To understand how this migration unfolded, we can look at several specific application areas. |
Transportation and Shipping Labels |
In the late 1990s and early 2000s, the automotive industry's transportation marking standards began to evolve. The European automotive supplier industry, for example, used a unified transport label that specified Code 39 as the barcode symbology for many years. By 2001, the standard was updated to reflect changing industry practices, with some versions beginning to support both Code 39 and Code 128 . |
The shift was driven by the need to include more information on shipping labels, including transport unit identifiers, hazardous materials warnings, and carrier handling instructions. These fields required more data capacity than Code 39 could reasonably provide. A single transport label might need to contain a part number, a container serial number, a quantity, a purchase order number, a carrier routing code, and a delivery destination. With Code 39, this information would require either a very long label or multiple separate labels. With GS1-128, all of it could be encoded in a single, compact barcode . |
This migration was facilitated by the availability of new barcode printers and scanners that natively supported both Code 39 and Code 128. Rather than forcing suppliers to replace all their equipment overnight, the industry transitioned gradually, with labels often including both Code 39 and Code 128 symbols during the transition period. |
|
Component Traceability |
The most critical application of barcodes in automotive manufacturing is component traceability. Every part that goes into a vehicle must be identifiable and traceable. |
A master's thesis study from Lund University examined the marking standards and technologies used by SWEP International, a manufacturer of brazed plate heat exchangers . The study, conducted in 2010, found that while SWEP primarily used Code 39 for internal product labeling, customers were increasingly requesting Code 128 and Data Matrix. The American automotive industry organization AIAG had developed standards that could be used with both Code 39 and Code 128, but Data Matrix was requested for its ability to encode more information in a smaller space, making it suitable for small products and direct part marking. |
The study also noted that RFID was not yet economically viable for SWEP's application. The cost of RFID tags and the fact that the products undergo a brazing process that would destroy the tags made RFID impractical. This highlights a key point: the migration was not about pursuing the latest technology for its own sake. It was about finding practical, economically viable solutions to real business problems. |
Similarly, the rail industry's adoption of serialization and batch management standards shows how similar principles apply to other transportation sectors. Knorr-Bremse's supplier guidelines specify that serial numbers and batch numbers must be encoded in a suitable barcode, with GS1 DataMatrix being the preferred application method . They also recommend that serial numbers be 8 to 12 digits long and include a manufacturer code prefix. These are practical guidelines that balance the need for uniqueness and traceability with the realities of barcode printing and scanning. |
|
The Impact on Manufacturing Operations |
The migration to GS1-128 and Data Matrix had a significant impact on manufacturing operations. The immediate effect was the need to upgrade or replace labeling equipment. Barcode printers had to support the new symbologies, and scanners had to be capable of reading them. The cost of this upgrade was significant, but it was offset by the operational benefits. |
Scanning accuracy improved. The structured data format of GS1-128 meant less manual data entry and fewer errors. A worker scanning a GS1-128 label could capture all of the relevant data in a single scan, whereas previously they might have needed to scan multiple Code 39 labels or manually enter some data. This reduction in manual entry decreased the likelihood of transcription errors. |
The use of Data Matrix for direct part marking enabled new levels of traceability. Parts could be marked directly at the point of manufacture, and the mark would stay with the part through every subsequent process: heat treatment, machining, assembly, testing, installation, and service. This cradle-to-grave traceability was a significant step forward for quality assurance and recalls. |
Manufacturing execution systems (MES) were upgraded to handle the new symbologies. These systems track the progress of parts through the factory floor, and they rely on barcode scans to update the status of each part in real time. The ability to capture more data from a single scan enabled MES to make more intelligent decisions, such as routing a part to a specific assembly line based on its serial number or automatically logging test results against the appropriate part record. |
Modern automotive manufacturing relies heavily on these systems. For example, in battery pack production for electric vehicles, every cell, module, and pack must be tracked with unique identifiers. The manufacturing execution system (MES) scans barcodes at each production step, linking each subcomponent to its position in the pack and recording key process parameters . A single Data Matrix code on a cell might encode not just the serial number, but also the cell's capacity, internal resistance, and manufacturing date. This is the kind of detail that would be impossible to encode in a Code 39 label of practical size. |
The industrial marking technology has evolved in parallel with these requirements. Marking solutions now include high-resolution laser printers that can create extremely small Data Matrix codes on metal surfaces, thermal transfer printers for labels, and durable inkjet systems for larger packaging . These systems are engineered for reliability and speed, with uptime requirements as high as 8000 hours per year on some production lines. |
|
Technical Characteristics and Their Impact on Application |
To appreciate why Code 39 was eventually supplanted, it is useful to examine its technical characteristics side by side with the alternatives. |
Data Density and Capacity |
Code 39 is a low-density symbology. Each character takes up a relatively large amount of space, and the intercharacter gaps add additional overhead. This makes Code 39 suitable for labels that have plenty of room---large boxes, shipping containers, and large components---but problematic for small items. |
The practical effect of low density is that Code 39 labels cannot encode long data strings. The industry-standard recommendation is that a Code 39 symbol should not exceed 20 characters and must not exceed 30 characters . This is a hard constraint. If your application requires encoding a part number, serial number, batch number, and date, you may quickly exceed 30 characters. |
By contrast, Code 128 (and thus GS1-128) is a high-density symbology. It uses variable-length codes and three character sets to pack more data into a smaller space. A Code 128 symbol of the same physical size as a Code 39 symbol can hold roughly 50% more data. Data Matrix, as a 2D symbology, has an even higher density, enabling large amounts of data to be encoded in a very small area. |
|
Character Set |
Code 39 encodes uppercase letters (A-Z), numbers (0-9), and a limited set of special characters: space, hyphen (-), period (.), dollar sign ($), slash (/), plus (+), and percent (%) . It cannot encode lowercase letters directly, nor can it encode the full ASCII character set without using a special extension mode that is rarely implemented. |
This limitation is not a problem for many applications, but it can be a constraint when encoding data that includes special characters. The automotive industry's parts identification standards explicitly avoid using characters that are not supported by Code 39 when data fields may need to be encoded in both linear and 2D symbologies . |
Code 128, by contrast, supports the full ASCII character set and the extended Latin character set. Data Matrix supports even more, including extended character encodings for international languages. This makes these symbologies more suitable for global supply chains, where part numbers and descriptions may include special characters or accented letters. |
|
Self-Checking and Error Correction |
Code 39 is considered self-checking, meaning that a single erroneously interpreted bar cannot generate another valid character . However, it does not include a mandatory check digit, unlike Code 128, which requires a check digit as part of the symbol structure. |
Data Matrix has even more powerful error correction. It can recover the original data even if a significant percentage of the symbol's cells are damaged, scratched, or obscured. This is particularly important for direct part marking, where the code is subject to wear. A mark on a metal part may be scratched, worn, or contaminated with oil or dirt. Data Matrix can still be read even with these imperfections. |
The tolerance of Data Matrix for damage translates directly into improved reliability in real-world conditions. Scanners can read a partially damaged Data Matrix code, reducing the number of 'no-read' events that would disrupt the production line. |
|
Readability |
One of the strengths of Code 39 is that it can be decoded by virtually any barcode reader . Its simplicity and age mean that it is supported by every barcode scanner ever made, including the simplest and cheapest devices. This ubiquity was a key reason for its success. |
However, the availability of cheap scanners that can read Code 39 does not mean they can read damaged Code 39 labels. A worn or smudged Code 39 label may be unreadable, especially if the label is small or the printing quality is poor. The lack of error correction means the scanner must read every bar and space correctly. |
By contrast, Data Matrix scanners are more sophisticated and require a camera-based reader (an imager) rather than a simple laser scanner. While this means Data Matrix readers are more expensive, they are also more capable. They can read codes from any orientation, read codes on curved surfaces, and read codes that are partially damaged. |
The tire industry provides a clear illustration of the tradeoffs between different data carriers. A white paper from the Global Data Service Organisation (GDSO) maps different data carriers to different use cases over a tire's lifecycle . While RFID offers cradle-to-grave traceability, printed barcodes like GS1-128 and Data Matrix remain essential for specific stages of production, supply chain, and logistics. The paper notes that different data carriers enable different use cases; there is no one-size-fits-all solution. |
|
Hardware Compatibility |
The migration to new symbologies required investment in new hardware. Laser scanners, which were the standard for Code 39, cannot read Data Matrix codes because they scan along a single line. Data Matrix requires an imager---a small camera that captures the entire 2D pattern at once. |
Many barcode scanners marketed as '2D imagers' support both linear and 2D symbologies. This compatibility made the transition easier for manufacturers, because they could gradually upgrade their scanning equipment while still maintaining the ability to scan existing Code 39 labels. However, the cost of 2D imagers was initially higher than laser scanners, which created a financial barrier to adoption. |
Today, the cost difference is minimal, and 2D imagers are the standard in most industrial applications. The additional capability of reading Data Matrix and QR codes, combined with the ability to read damaged labels and to capture images for quality assurance, justifies the small premium. |
The Consequences of the Migration |
The migration from Code 39 to GS1-128 and Data Matrix had far-reaching consequences for the automotive industry. |
|
Improved Quality and Recall Management |
The most important consequence was improved traceability. Automakers now know, with a high degree of certainty, what components are in every vehicle they build, and what vehicles contain every component they produce. This has transformed recall management. A recall for a defective brake caliper, for instance, can be precisely targeted to the vehicles that received calipers from a specific production run. This reduces the number of vehicles affected, the cost of the recall, and the damage to brand reputation. |
Supply Chain Efficiency |
The use of GS1-128 and Data Matrix labels also improved supply chain efficiency. Suppliers print labels that comply with the OEM's requirements, and OEMs scan these labels automatically as they receive shipments. The data from the label is automatically entered into the inventory management system, reducing manual data entry and eliminating the errors that come with it. |
The structured data format of GS1-128, with its application identifiers, enables automated handling. A scanner reading a label can parse the data and route the parts to the correct location without human intervention. This is particularly important for large, complex facilities where hundreds of different parts arrive each day. |
|
Reduction in Errors |
The migration also reduced errors on the factory floor. Workers scanning labels to verify the correct part or to log a serial number are less likely to make a mistake than workers manually entering data. The scan is quick, accurate, and can be verified in real-time by the manufacturing execution system. |
For example, in an automotive assembly plant, a worker might scan a Data Matrix code on a sensor and the system automatically verifies that this is the correct sensor for the vehicle model being built. If it is wrong, the system issues a warning, preventing the assembly of the wrong part. This kind of error prevention is critical for quality and safety. |
|
Better Aging and Decommissioning Tracking |
Finally, the ability to encode detailed production dates and batch information enables better tracking of part age and shelf life. Some parts, such as rubber seals and batteries, have finite lifespans. A clear, machine-readable batch code enables system-level expiration tracking, preventing the installation of aged parts. In the case of electric vehicle batteries, the manufacturing execution system records every process parameter for each battery pack, enabling a detailed lifecycle history that can improve warranty management and end-of-life recycling . |
|
A Detailed Summary |
The automotive industry's migration from Code 39 to GS1-128 and Data Matrix was a response to real, practical needs. Code 39 served the industry well for two decades, but it was ultimately not capable of meeting the growing demand for data capacity and durability in a manufacturing environment characterized by ever-increasing complexity and regulatory oversight. |
The low data density of Code 39 meant that labels containing multiple data fields---serial number, date, batch, part number, supplier code---were impractically long. The physical constraints of automotive manufacturing, where many components are small and irregularly shaped, made these long labels difficult to apply and prone to damage. The lack of error correction meant that damaged labels were often unreadable, disrupting production and reducing the reliability of the traceability system. |
GS1-128 addressed the density issue by packing more information into a smaller space while supporting structured data fields with application identifiers. This made shipping labels and component labels more compact and more informative. A single scan could capture all of the relevant data, reducing manual entry and improving accuracy. |
Data Matrix went further, offering extremely high density, robust error correction, and the ability to be marked directly onto parts. This enabled cradle-to-grave traceability for critical components, improved recall management, and supported the trend toward direct part marking that is now standard in many industries. |
The automotive industry's adoption of these newer symbologies was not instantaneous. It was a gradual process that unfolded over several years, driven by changing customer requirements, new standards from bodies like AIAG and GS1, and the availability of more capable hardware. By 2005, the transition was sufficiently advanced that most automakers required GS1-128 or Data Matrix for traceability purposes, acknowledging that Code 39 labels simply could not hold enough traceability data . |
The migration had a lasting impact on the industry. It improved quality control, enabled more effective recall management, reduced errors in supply chain operations, and laid the groundwork for the data-driven manufacturing that is now standard in the automotive sector and beyond. |
|
For the technical reader, it is useful to remember this transition as an example of how practical constraints---space, durability, and data capacity---drive technological change. Code 39 was not a failure. It was a success that enabled a generation of manufacturing efficiency, and its legacy lives on in the many systems that still rely on it for basic identification. But as the industry's requirements evolved, so too did the technology. The migration to GS1-128 and Data Matrix was not a repudiation of Code 39, but rather its natural successor: a response to the demand for more information, in less space, with greater durability and reliability. |
cs@easiersoft.com