In ultrasonic plastic welding, three process variables directly affect weld quality: amplitude, duration and force.

Amplitude has long been controlled through frequency selection, horn and booster design, and regulation of electrical input to the transducer.

But, control over the other two variables, force and time, has evolved considerably. In the past, duration could only be controlled by setting a specific weld time. Force could only be controlled by regulating air pressure in a pneumatic cylinder.

The advent of servo-driven ultrasonic welders has changed all that. With feedback from load cells, encoders and other sensors, servo-driven welders offer extraordinary command over every aspect of the joining process. Engineers can now monitor and control ram force, ram speed and weld distance.

To facilitate process optimization, these machines can export weld data in formats that can easily be imported to Excel or Minitab for analysis. The data produced by the generator, such as graphs or cycle information, is also useful for process optimization and troubleshooting.

Servo-driven ultrasonic welders offer advanced control features that are fundamentally different from those of traditional equipment. We recently completed an in-depth study to identify and measure how these newly available settings affect weld quality. These results can be used to expedite process optimization for critical assemblies.

In recent years, much research has been conducted on servo-driven ultrasonic welders. Multiple studies have shown that servo-driven welders produce stronger welds than those produced with pneumatic welders. Each study has demonstrated the benefits of servo-driven welders with regard to process consistency and performance. However, none have investigated the full range of features available, providing only small windows into the capability of this new equipment.

Distance Control

Servo-driven welders offer excellent repeatability of collapse distance from part to part. With pneumatic systems, there is a limit to how quickly air can escape the cylinder, preventing abrupt changes in velocity and limiting distance control. Dukane’s servo-driven welder can accelerate at up to 1,270 millimeters per second squared, allowing almost instantaneous velocity shifts during the weld and hold phases.

An initial study showed that the servo welder could achieve a standard deviation of 1.1 percent in measured collapse distance, compared with 3.9 percent with the pneumatic welder. Value Plastics Inc., a Colorado-based manufacturer of precision-molded couplers and components for medical devices, has been able to achieve a consistent collapse distance (standard deviation of 0.9 percent) in production using our servo-driven welder.

Clearly, servo-driven welders offer better repeatability and precision. This improved consistency will be essential as plastic assemblies become more complex.

Velocity Control

Even before the introduction of servo-driven welders, engineers recognized the need for velocity control. A consistent melting rate has a direct influence on bond strength. When the horn moves downward at a steady velocity, it produces a steady melt rate. That, in turn, creates a homogenous molecular structure and a stronger weld. Precise speed control has been shown to be a great aid in welding under less than ideal circumstances. For example, servo welders can create stronger bonds than pneumatic welders when grease is present in the joint or the energy director is damaged.

In the past, suppliers have put a lot of effort into getting consistent velocity control with pneumatic systems. However, these efforts have been in vain. It is simply not possible to get precise velocity control with a pneumatic welder. By directly controlling velocity, servo-driven welders are a clear improvement.

Studies using servo-driven welders have shown that weld velocity correlates directly with weld strength. In a study at The Ohio State University, researchers showed that by using a defined velocity profile with a slower speed during melt initiation and a faster speed in the middle and end of the weld, strength could be increased with less weld time and reduced surface marking.

Dukane has developed a new weld control to initiate melting of the plastic before collapsing the joint. Called Melt-Detect, this feature allows the press to contact the parts and turn on ultrasonic vibrations with no vertical movement until a drop in force is detected. This indicates that the plastic has started to melt and welding has begun.

Designing a Test Part

To conduct our study, Dukane created a new standard test part, the Industrial Standard Test Part (ISTeP). Previous experimentation has shown that the standard plastic I-beam test part is prone to warp and sink, which makes consistent measurements difficult to achieve. To truly test the capabilities of servo-driven welders, we needed a more precise test part.

The ISTeP was developed to facilitate precise height measurement and consistent pull testing results. Its round design prevents “zippering,” in which one side of an I-beam test part fractures first, leading to a peel effect on the remaining weld. Since peel strength is less than pull strength on ultrasonic welds, this leads to a less accurate and artificially reduced pull test value.

Additionally, the mold for the ISTeP was designed for even filling. This prevents warping and sink marks, and it ensures consistent wall thicknesses.

The mold design also includes an insert to allow molding of an unlimited number of joint designs. Already, the mold has been used to produce a shear joint, a butt joint with a 60-degree energy director, and a butt joint with a 90-degree energy director. However, any joint design can be easily and quickly implemented. For all assemblies in this study, the butt joint with a 90-degree energy director was used.

The ISTeP allows for pull testing, pressure testing via a tube opening, and measuring part height before and after welding to determine actual collapse distance.

Before we could test the capabilities of our servo-driven ultrasonic welder, we first had to determine if the new ISTeP was, indeed, better than I-beam parts for test purposes. To that end, we compared four weld setups. In all four, trigger force was 250 newtons. Weld distance was either 0.25 or 0.3 millimeter. Weld speed was either 2 or 10 millimeters per second. Amplitude was either 90 percent or 100 percent. Melt Detect was set at either 5 percent or 10 percent. Hold distance was 0.05 millimeter; hold speed was either 5 or 12.7 millimeters per second; and hold time was 0.5 second.

Part height was measured before and after welding. Because the I-beam parts did not self-locate, a fixture was used to perform these height measurements before welding.

For all experiments, weld strength was measured using a pull tester with an accuracy of ±2.5 pounds. Tooling was designed specifically for each part.

Across all weld settings, the ISTeP parts showed more consistent results for both collapse distance and pull strength. Additionally, the standard deviation of the collapse distance measurements of the I-beam parts become more consistent after practice, showing that it is more dependent on operator skill than measurements of the ISTeP parts.

Gauge R&R

Prior to completing our servo welder study, it was also necessary to conduct a gauge repeatability and reproducibility study to ensure that our means of part measurement could produce sufficiently consistent results. Ten parts were welded at low, nominal and high weld settings. (These settings were also used for the servo welder study.) Results were interpreted using Minitab.

On the low setting, trigger force was 100 newtons; weld distance was 0.25 millimeter; weld speed was 1 millimeter per second; amplitude was 80 percent; Melt Detect was set at 0; hold distance was 0; hold speed was 1 millimeter per second; and hold time was 0.

On the nominal setting, trigger force was 250 newtons; weld distance was 0.3 millimeter; weld speed was 2 millimeters per second; amplitude was 90 percent; Melt Detect was set at 5 percent; hold distance was 0.05 millimeter; hold speed was 3 millimeters per second; and hold time was 0.5 second.

On the high setting, trigger force was 400 newtons; weld distance was 0.35 millimeter; weld speed was 3 millimeters per second; amplitude was 100 percent; Melt Detect was set at 10 percent; hold distance was 0.1 millimeter; hold speed was 5 millimeters per second; and hold time was 1 second.

Two runs were completed for the gauge R&R study. In the first run, 10 welds were tested at each weld parameter set by three separate operators. After this, some changes were made to the pull-test fixture to improve consistency. The second run used only one operator, who later completed the entire run for the servo welder study independently.

The results of the first run showed good distinction between the weld sets, verifying the repeatability of the test equipment. Also, all three operators produced virtually the same results, verifying reproducibility.

Design of Experiments

To better use servo-driven ultrasonic welders, it’s important to understand how each parameter affects weld quality. Minitab was used to develop the design of experiments (DOE) and to perform the analysis.

Seventy-two runs were completed with 18 samples for each parameter set, for a total of 1,296 samples. The entire set was welded by one operator before any measurements were done. After completing the welding, the same operator carried out all the pull tests.

A Dukane servo-driven ultrasonic welder with an HMI running iQ Explorer II was used to weld the parts. The welder was equipped with a flat face, high-gain horn. A simple drop-in style fixture was used to hold the parts.

After completing the full DOE, a follow up experiment was conducted on the effect of welding speed on weld quality. Weld speeds were chosen to provide a closer look at the effect of this parameter within the range chosen for the DOE and at slower speeds. The other parameters were maintained at the nominal settings.

The complete DOE results were analyzed by Minitab, which demonstrated some exciting results. The three weld parameters with the greatest effect on weld strength were weld speed, weld distance and Melt-Detect percentage. This is particularly interesting, since precise weld speed and weld distance control are the hallmarks of servo-driven welders. Additionally, the Melt-Detect feature is exclusive to Dukane’s ultrasonic servo welders.

Specifically, the analysis shows that reduced weld speed, increased weld distance, and increased Melt- Detect percentage all improved weld quality.

Surprisingly, there was little interaction between each of these weld parameters. However, there could be some interaction in these parameters that was outside this selected value set.

A follow up experiment was conducted to take a closer look at the effect of weld speed on weld strength. Significantly slower weld speeds resulted in vastly improved weld strength results, even at the nominal weld distance and Melt-Detect levels selected for this test.


These experiments yielded several significant findings. First, Dukane’s new ISTeP design offers significant advantages over the traditional I-beam design. The ISTeP parts allow more accurate measurement, which can help develop processes for critical and precise applications.

Second, we showed that the pull-test equipment developed for this part can provide consistent results for accurate analysis of weld strength.

But perhaps the most exciting discovery is the importance of weld speed, weld distance, and Melt-Detect percentage to weld strength. Weld speed and weld distance can each be precisely controlled by servo-driven welders, and our research has demonstrated why this ability is so important to critical assemblies.

Melt-Detect is a component of Dukane’s Melt-Match technology, which allows the welder to turn on ultrasonic energy with no movement once part contact has been detected. The welder then waits at this position until a drop in force has been detected, which indicates that melting has begun. Melt-Detect can be programmed to define a percent of drop in force that must be observed, below the trigger point, before downward motion is initiated. This study has shown that this feature provides a key means of adjusting the weld for increased strength.

Finally, detailed investigation shows that the ability to weld at very slow weld speeds allows drastically greater weld strengths to be achieved.

Editor’s note: Mike DeVries of Parker Hannifin Corp. and Bob Aldaz, Loc Nguyen and Ken Holt of Dukane also contributed to this feature. Miranda Marcus is no longer with Dukane. For more information on servo-driven ultrasonic welders, contact Michael Johnston at 630-797-4902.

For more information on ultrasonic welding, read these articles: