MD+DI Online is part of the Informa Markets Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Validating the Medical Device Laser Welding Process

A proven method for systematically improving process and product quality, validation helps to reduce scrap and field failures while enhancing a medical device’s competitiveness.

By David W. Steinmeier and Lisa Schaller

In the medical device world, laser welding encompasses a wide range of applications and part sizes. However, when laser-welding process cannot be fully verified, FDA requires that manufacturers validate them. This article illustrates the necessary steps that manufacturers must take to validate their laser-welding process and highlights the steps they should consider to conduct successful validations.

Why Validate?

While the terms validation and verification are often used interchangeably, they have very different meanings. Validation ensures that the right product was made, and verification ensures that the product was made right.

Manufacturers should validate their laser-welding process for four reasons:

  • For those manufacturers that practice Six Sigma techniques, there is no laser-weld monitor or checker on the market today that can separate bad welds from good welds to a Six Sigma confidence level without destroying the product. Therefore, validating the laser-welding process is the only way to determine weld quality.
  • For medical device manufacturers, FDA mandates that manufacturing processes that cannot be fully verified must be validated.1
  • Improving process yields by reducing product scrap and field failures far outweighs the cost of validating the laser-welding process.
  • Validation is a good marketing tool. Manufacturers capable of proving their laser-weld quality level to their customers have a substantial advantage over the competition.

As illustrated in FIGURE 1, the validation process consists of six main components: the validation plan, equipment installation qualification, equipment operational qualification, process qualification, process validation, and product performance qualification. Each component contains its own protocol, data, and report, and manufacturers may implement them in different ways to accommodate their design and manufacturing processes. A common modification incorporates the equipment installation qualification step with the equipment operational qualification step. Another variation includes a design of experiment (DoE) process as part of the equipment operational qualification step instead of a more traditional method embedded within the process qualification step. In either case, the goal is the same: to consistently produce a product that meets the intended use.

Laser-Welding Validation: A Case Study

While this article analyzes a minimally invasive surgical tool to illustrate the basic laser-welding validation process, this process applies to all laser-welding applications. The surgical tool shown in FIGURE 2 consists of two parts: a tip and a shaft, both of which are made from Type 304L stainless steel. The tip must be inserted into the shaft without being cocked and without damaging the shaft sidewall.

Example of a laser-welded surgical tool.

In addition, the insertion process must minimize the weld-junction gap between the tip and shaft because a gap can cause voids and expulsion. Previous studies involving this surgical tool showed that the laser-welding process can tolerate a maximum weld-junction gap of 0.05 mm. Because the mechanism used to insert the tip into the shaft is part of the laser-welding system, the ability to consistently assemble the tip and shaft with a weld-junction gap of less than 0.05 mm must be validated. For this application, the required fit-up study was included in the equipment operational qualification.

Validation Plan. All validation projects must start with an overall game plan known as the validation plan. This plan must address each of the six elements shown in FIGURE 1. A well-conceived validation plan is a roadmap to success.

In addition to the elements shown in FIGURE 1, the validation plan defines postvalidation process controls and monitoring, which ensure that the validated laser-welding process meets all the specified requirements. The plan also defines the conditions under which the welding process must be revalidated. Such conditions are met when the process transfers, changes, or deviates from the validated state.2,3

Equipment Installation Qualification. This step involves setting up the equipment in accordance with the manufacturer’s installation specifications and verifying the equipment calibration. Equipment manuals contain the key installation information required to ensure proper equipment operation. Sample equipment installation qualification elements required to ensure the functionality of the laser-welding equipment include the mains voltage range, the mains frequency range, the airflow space for adequate cooling, the room-temperature range, the minimum bend radius on the fiber optic cable that connects the laser power supply to the focusing head, and the argon cover-gas flow system.

Calibration checks should be performed at the beginning of the laser-welding validation process. These checks can be as simple as verifying the information on a calibration certificate from the laser-welding equipment supplier or performing a calibration check at the beginning and end of the validation process.

Equipment Operational Qualification. This component verifies that the laser-welding system meets the manufacturer’s performance specifications. It also establishes procedures and documentation for calibrating, cleaning, operating, and maintaining the system and for training operators. Equipment manuals contain key equipment capability specifications.

For laser-welding systems, the most important welding equipment parameters are weld power, pulse duration, spot size, and pulse repetition frequency, which control the weld-spot overlap and weld-spot location in all three of the system’s axes. Manufacturers should verify and document in the equipment operational qualification report that the entire welding system produces the programmed laser-welding parameter power over the projected operating ranges on a repeatable basis.

Because many laser welds are sensitive to the weld-junction gap and the weld-spot location on the weld junction in the x, y, and z axes, gauge R&R (or ANOVA gauge repeatability and reproducibility) studies must be conducted on the laser-welding system to establish its ability to produce the correct weld-junction gap and weld-spot location. If available, historical data from similar laser welding processes should be used to set the limits. Otherwise, manufacturers should determine limit values using the DoE process. In the case of the surgical tool, the maximum range limits for both the weld-junction gap and the weld-spot location were derived from historical data.

The results of both gauge R&R studies successfully established the fit-up and laser weld-spot positioning capability. The weld-junction gap passed, the maximum range limit value was 0.050 mm, and the actual weld-junction fit-up range was 0.025 mm < 0.05 mm. The worst-case range value of the weld-spot location passed, the maximum range limit value was 0.30 mm, the actual x axis (shaft axis) range was 0.09 mm < 0.30 mm, the actual y axis (shaft diameter) range was 0.06 mm < 0.30 mm, and the actual z axis (laser focus) was 0.09 mm < 0.30 mm.

Process Qualification. This step follows the equipment operational qualification step and contains seven subcomponents, as shown in FIGURE 3.

Selecting the Process Qualification Weld-Quality Metrics. This process begins with selecting the weld-quality metrics, which should represent the stresses and physical limitations to which the final laser-welded product is subjected by the end user. This surgical tool is pressurized during the procedure and must be free of voids, even if the voids do not leak. For this example, two weld-quality metrics were selected: minimum burst pressure and absence of voids.

Measuring burst pressure is time-consuming, expensive, and destructive. Therefore, it was decided to add two nondestructive metrics that might correlate with the burst pressure: weld-spot location in relation to the weld-junction centerline and weld width.

Conducting the Pre-DoE Study. The purpose of conducting a pre-DoE study is to select the variable input factors, fix certain input factors, identify input factors that represent experimental ‘noise,’ and determine the range for each variable input factor.

To determine which input factors to include as variable input factors in the DoE, manufacturers should examine existing laser-welding production processes. In this case study, the variable input factors selected included peak power, pulse duration or pulse width, percent of weld-spot overlap, and weld-spot location in relation to the weld-junction centerline.

Next, manufacturers should select which input factors to fix or hold constant. In this case study, the rotational speed of the tip/shaft assembly and the pulse repetition rate control the percent weld-spot overlap. Because the easiest parameter to vary is the pulse repetition rate frequency, the rotational speed should be fixed. TABLE I presents the fixed input factors.

Uncontrolled input factors representing experimental ‘noise’ include variations in the weld-junction gap and the weld-spot location in relation to the weld junction. The measured limits of both noise sources are listed in the equipment operational qualification section.

Manufacturers must know which variable input factors are important for producing the desired output responses. Therefore, a series of trial-and-error miniexperiments must be conducted to determine the range for each variable input factor that produces both ‘cold’ and ‘hot’ welds. Cold laser welds produce voids or incomplete weld flow, while hot laser welds produce unacceptable weld splash and distortions in the weld flow.

Using these definitions for cold and hot laser welds resulted in the variable input factor ranges shown in TABLE II. Overlap is one variable in which the pulse-repetition frequency controls the overlap percent.

Conducting the DoE. A D-optimal model with four variable input factors, six replicates, and three output responses was used. The D-optimal model provides two-order interactions with excellent model strength and requires fewer parts than the use of a full factorial model. The three output responses included burst pressure, weld width, and void length.

TABLE III presents the ANOVA results for the three output responses. The burst pressure and weld-width model results are very strong, with error values of 13.77% and 11.21%, respectively. The weld-spot location is the primary input factor affecting the burst pressure, while the pulse duration is the primary input factor affecting the weld width. With an Error value of 84.56%, the void-length model is meaningless. Correlation studies showed no correlation between the weld width and the burst pressure. Therefore, weld width and void length cannot be used as nondestructive process qualification weld-quality metrics.

Optimizing the Welding Parameters. The surgical tool was pressurized during product performance qualification to 20.68 MPa. An optimized target goal of 96.53 MPa and no voids was specified. Illustrated in FIGURE 4, the marginal means graph for burst pressure shows that achieving the optimized value is possible.

Using DoE expert prediction software produces the optimized set of laser-welding parameters for the surgical tool, as shown in TABLE IV.

Determining the Lot Run and Sample Size. As shown in FIGURE 5, the most significant input factor controlling the burst pressure is the distance between the center of the weld spot (D4 in the figure) and the weld-junction centerline (D0). Therefore, the process qualification acceptance metric uses the weld-spot location deviation for qualifying the laser-welding process. To prepare for the process qualification confirmation run, it is necessary to determine the confirmation run sample size and the minimum K values, which represent the acceptance criteria.4–7 Statistical software such as Minitab can be used to calculate the sample size and K values.

The variable data used during this step included a production lot run size of 2000 pieces, an acceptable quality level of 0.023% (5 Sigma = Cpk = 1.67), a rejectable quality level of 0.15% (5 Sigma = Cpk = 1.67), and—to absolutely prevent weld voids—a maximum and minimum deviation of the weld spot from the weld-junction centerline of ±0.10 mm. The historical maximum radius of the weld-spot from the weld-junction centerline is 0.015 mm. Thus, given the foregoing parameters, Minitab 15 calculated a minimum sample size of 30 welded pieces for each operator and an acceptance K value of 3.2.

Conducting the Confirmation Run. Three operators made 30 samples each using the test conditions defined in TABLE V and the fixed parameters listed in TABLE I. The confirmation run was conducted at two different power levels to test the use of a wide weld window during production. X represents the operator identification code.

Before each laser-welded sample was removed from the laser-welding system, the deviation of the weld-spot location from the weld-junction centerline was measured and recorded using the measurements shown in FIGURE 5. First, D1 and D2 were measured and recorded. Then, the weld width was calculated and recorded, so that D3 = D1 + D2. Next, the weld-location error was calculated and recorded, so that D4 = D1 – (D3/2). D4 was used for determining process qualification acceptance or rejection. After the weld-spot location was measured for each sample, the samples were submitted for process validation testing.

Applying the Process Qualification Acceptance Criteria. For each operator, the average and standard deviation of the weld-spot location was calculated, and the Z test statistics—the Z lower spec limit (Z.LSL) and Z upper spec limit (Z.USL)—were determined using the formulas Z.LSL = [(average weld-location error) + (0.10 mm)] / standard deviation and Z.USL = [(0.10 mm) – (average weld-location error)] / standard deviation. TABLE VI shows the calculated Z.LSL and Z.USL values for each operator.

All three operators successfully laser welded tip/shaft assemblies that contained no voids and passed the process qualification acceptance criteria for the weld-spot location. The worst-case weld-spot location error across all three operators was ±0.025 mm, which is four times less than the upper/lower limit range of ±0.10 mm. Thus, the laser-welding process was qualified, but it still required process validation testing.

Process Validation. This step establishes that the welding process consistently produces a part or product meeting its predetermined specification. Process validation metrics must represent the stresses encountered during product use and must be different from the process qualification metrics. Process validation involves correlating process qualification data with process validation data.

The first test used to achieve a process validation weld-quality metric was a leak test based on helium gas. To pass the test, the surgical tool could not exhibit any leaks when subjected to pressure of 6.89 MPa over a 20-sec period. The device remained leak-free if no helium gas bubbles surrounded or emanated from the laser weld.

The second process validation weld-quality metric was derived by performing an outer-diameter test across the entire shaft length of the surgical tool. Since a laser weld does not produce a flat surface, the outer diameter across the laser-spot weld width can vary depending on the laser power and duration. All process qualification samples passed both process validation weld-quality tests.

To ensure that the weld-spot location was controlled properly, as defined in the validation process, the worst-case weld-spot location had to be derived from the weld-junction measurement obtained within a single lot. To ensure that the actual Z.LSL and Z.USL values remained above the critical K value of 3.2, the distance between the center of the weld spot (D4 in FIGURE 5) and the weld-junction centerline (D0) had to be ±0.095 mm.

Product Performance Qualification. This step documents that the finished product meets all requirements for functionality and safety. Product performance qualification incorporates a series of environmental tests that are used to simulate the operating environment of the finished product. Environmental tests include, but are not limited to, life cycling, temperature, vibration, humidity, impact, and shipping. If no failures of any type occur, the product is considered to be validated. Should weld failures occur during the product performance qualification step, the basic product design must be revisited for weldability and the laser welding process should be revalidated.

In this case study, the first product performance qualification metric was derived from a pressure test using argon gas. During a surgical procedure, the surgical tool must not leak at an operating pressure of 20.68 MPa. The device is determined to be leak-free if no argon gas escapes from any part of the assembly. The second product performance qualification metric was derived from a tip-temperature test. The tip temperature must be lower than –80°C during the surgical procedure. All process qualification samples passed both tests.


Validation is a proven method for systematically improving process and product quality. It is also important for reducing product scrap and field failures, as well as for enhancing the competitiveness of a medical device.


  1. 21CFR-820, “Code of Federal Regulations Title 21: Quality System Regulation,” Chapter 1, Subchapter H—Medical Devices, Quality System Regulation (Silver Spring, MD: FDA, 2014).
  2. 21 CFR Part 820.75(c), “Code of Federal Regulations Title 21: Quality System Regulation,” Chapter 1, Subchapter H—Medical Devices, Quality System Regulation (Silver Spring, MD: FDA, 2014).
  3. 21 CFR Part 820.75(b)(2), “Code of Federal Regulations Title 21: Quality System Regulation,” Chapter 1, Subchapter H—Medical Devices, Quality System Regulation (Silver Spring, MD: FDA, 2014).
  4. D Steinmeier, “Weld Quality Validation—Sample Size Selection” [online] (Arcadia, CA: microJoining Solutions, 2009 [cited 11 September 2014]; available from internet:
  5. 21 CFR Part 820.250, “Code of Federal Regulations Title 21: Quality System Regulation,” Chapter 1, Subchapter H—Medical Devices, Quality System Regulation (Silver Spring, MD: FDA, 2014).
  6. EG Schilling, “Sampling by Variables, Quality Control Handbook,” 3rd ed. (New York: McGraw-Hill, 1974), Tables 25-18 and 25-19.
  7. ISO 3951, “Sampling Procedures and Charts for Inspection by Variables for Percent Nonconforming,” 2nd ed. (Geneva: International Organization for Standardization, 1989), 18–25.

David W. Steinmeier is principal and consult at Arcadia, CA–based microJoining Solutions. Specializing in medical device technologies, he has manufacturing process expertise in packaging and assembling miniature and microminiature electronic and electromechanical products for the aerospace, automotive sensor, commercial electronics, and electromechanical industries. He also has extensive experience in developing, optimizing, and validating laser and resistance welding processes. Contact him at [email protected].

Lisa Schaller is the associate director, quality at Austin, TX–based HealthTronics Inc. A quality expert in the medical device industry, she has extensive experience in process validation, supplier management, and lean manufacturing techniques. She received an MS in biomedical engineering from the University of Akron and a BS in mechanical engineering from Ohio Northern University. She can be reached at [email protected].


TAGS: News
Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.