When joining plastic parts, different applications require different methods.

When welding thermoplastic parts, there are a number of different technologies to choose from, including ultrasonic, hot-plate, vibration, laser and spin. Hot-gas welding, in which heated air or nitrogen is used to melt a plastic welding rod, and induction welding, in which a ferromagnetic insert is placed between two parts and energized with an electromagnetic field in order to melt the material, are two other options. However, the former is slow and dependent on operator skill, and thus largely relegated to field work and repairs. The latter is expensive and complicated thanks to the ferromagnetic insert.

The true workhorse technology for joining plastic parts is ultrasonic welding, thanks to its fast cycle times and low cost. One drawback is that it can only be used with small or medium-sized parts. However, there is no lack of small and medium thermoplastic parts in this world, so the process is alive and well.

Spin welding is also inexpensive, but limited in that one of the pieces needs to be circular in shape. Hot-plate and vibration welds can be used for large parts or those that incorporate irregularly shaped joints. But the systems that produce them are expensive either to operate or buy.

Laser welding, a relative newcomer to the game, is expensive and limited in terms of the number of materials with which it can be used. However, it provides the most precise welds available, with no flash or particulate matter. The technique can be used on a wide variety of joint configurations.

Squeak and Squeeze

Although an exact breakdown is impossible, the consensus among equipment manufacturers is that 50 to 70 percent of plastic welding is done ultrasonically. First developed in the 1960s, it has become the "go to" method for joining plastic parts. Unless there is a compelling reason to do otherwise, most assemblers will go ultrasonic by default.

In terms of the mechanics of the process, creating a weld involves the use of a titanium or aluminum horn. This horn both clamps the two pieces together and vibrates vertically to melt the material at the joint through heat generated by molecular vibration and friction. Vibration frequencies range from 15,000 to 72,000 hertz, with most applications using around 20,000 hertz or frequencies in the 30,000- to 40,000-hertz range. Higher frequencies are used for more delicate assemblies, since they require smaller vibration amplitudes, which are easier on parts. These amplitudes are very small-from 5 to 150 micrometers-and cycle times are usually 1 second or less. The process can be loud, sometimes requiring soundproofing. But it creates neither fumes nor smoke, so no ventilation is necessary.

In some cases, especially those involving small consumer products, three or four welds can be performed with a single composite horn comprised of a drive horn and multiple extenders, giving throughput an additional boost. One drawback to this approach is that it is impossible to monitor each individual weld parameter, making it hard to guarantee quality. Although, ultrasonic welding creates some flash and particulate matter, the vibrations are fine enough that it isn't excessive. If a weld needs to be aesthetically appealing, it can be designed with a "flash trap," or a recessed groove, to hide any excess melt.

With regard to cost, most of the other methods don't even come close. An entry-level ultrasonic welder (or in the words of Herrmann Ultrasonics product manager Vasko Naumovski, a basic "squeeze and squeak" model) can be had for as little as $9,000. Refurbished models from companies like Sonitek (Milford, CT), which sells both new and used welders, can cost $7,000 or less, depending on the vintage.

In recent years, ultrasonic welders have become increasingly complex. Digitally controlled machines, like Herrmann Ultrasonics' (Schaumburg, IL) computer numeric controlled (CNC) Dialog welder or the K20-XT from Stapla Ultrasonics Corp. (Wilmington, MA), can operate in a number of modes, including weld depth, time, peak power, absolute distance and energy. Even at the top end, though, ultrasonic welders can be had for around $40,000. That's competitive with the cheapest vibration and hot-plate welders, and far less than laser systems, which cost $100,000 and up. Spin welders are competitive on price, but again, they are limited in the types of parts they can weld. The mechanics of spin welding are also more aggressive, creating more flash and particulate matter around the weld.

In terms of part size, opinions vary and it depends on the materials, but parts with a 6-by-6-inch footprint are about as large as you can go.

"Basically, if the parts fit in a shoebox they can be ultrasonically welded," says Naumovski. Otherwise, he says, you may need to go with another method, invest in a more complicated machine, like his company's fully enclosed, multihead Ultraline model or at the least take extra care when setting process parameters.

Ultrasonic performance is also affected by the distance between the horn and the joint, the configuration of the joint, and the basic geometry of the parts being welded. A "near-field" joint, for example, is one in which the horn comes within 0.25 inch of the weld site. This is the ideal and makes for a quick, easy weld. With a "far-field" joint, the horn can have problems transmitting vibrations to the work area. Ultrasonic joints can be made with the horn several inches from the joint. But they require far more power. The same is true if a part contains voids or curves that absorb the ultrasonic energy instead of transmitting it to the joint.

A further complication arises if the parts being welded are made of a soft plastic, or a semicrystalline thermoplastic like polypropylene or polyethylene. In the latter cases, the springlike molecular structures that make up the material tend to absorb the ultrasonic vibrations, blocking them from the weld site. Amorphous thermoplastics like polycarbonate or rigid PVC, on the other hand, make for much easier welding. With these materials, the constituent molecules are randomly jumbled together like a plate of spaghetti and transmit ultrasonic energy much more efficiently.

This is not to say welding with semicrystalline thermoplastics can't be done. In fact, Jeff Frantz, director of applications and acoustic engineering for Branson Ultrasonics Corp. (Danbury, CT) emphasizes that "thousands" of semicrystalline assemblies are joined using ultrasonic energy. It's just that designers have to pay a little more attention to the details.

Finally, when welding ultrasonically, the joints need to be almost entirely flat, so all points are affected equally by the ultrasonic energy. According to Dan Bolduc, applications manager at Sonitek, height variations of as much as 0.5 inch are possible using composite horns. Any more, though, and an assembler will likely need to shop around for another method.

For all these reasons, product designers should solicit input from process engineers early on, so questions like joint design and part geometry can be settled before committing time and money to one particular form. Adapting part geometry to process needs not only ensures good, strong welds, but welds that can be performed in the shortest possible time.

"We have actually caught a lot of issues before the mold is cut," says Bolduc. "The mold builders for some customers may suggest that they see us early on just to make sure the customer is following the right path."

Not that the process is completely unforgiving. Bolduc says his company has had a great deal of success with customers interested in refining their assembly processes.

"Often, customers will be upgrading from, say, adhesives, and it turns out the part can be ultrasonically welded just fine," he says. "Maybe we just need to change the process parameters, although there comes a point where you can only go so far."

With ultrasonic welding, assemblers have a variety of options for joint design, from simple butt joints to tongue-and-groove or step joints with or without melt recess grooves. Many joints incorporate "energy directors," V-shaped ridges protruding from the part that concentrate the ultrasonic energy, rapidly initiating softening and melting at the joint surface. Branson Ultrasonics has also pioneered the use of a textured surface opposite an energy director to further facilitate melting, help reduce the formation of particulates and form even stronger welds.

Do the Twist

Of course, not all assemblies fit in a shoebox. For this reason, there are other methods for welding plastic parts.

Because of cost considerations, spin welding is the next logical option for many assemblers. With this method, one part is held motionless in a fixture while the other is spun at high speed. The spinning part is pressed against the fixed part, creating friction in the bond area that causes the plastics to melt. When the spinning is stopped, the parts are held together under pressure as they cool, creating the weld.

As with ultrasonic welding, spin welding is fast and energy efficient. Cycle times are usually just 1 or 2 seconds. The method also works with pretty much any thermoplastic. In fact, spin welding is particularly useful in welding semicrystalline thermoplastics since their molecular structures have less of a damping effect than with ultrasonics.

Spin welds are strong, but because of the dramatic movement involved in creating the weld, the process produces more flash and particulate matter. As is the case with ultrasonics, a number of joint designs are available, including tongue-and-groove configurations, shear joints and scarf joints, all of which can help in creating hermetic seals. Flash traps can be used to create more aesthetically pleasing welds.

There is theoretically no limit on the size of the parts that can be joined. Spin welds have been performed with parts that are more than 3 feet across, although cases like these are the exception. Typical applications include float cylinders made from softer materials, and tube-and-housing assemblies in the medical field. According to Forward Technology (Cokato, MN) product manager David Kralovetz, a classic spin welding application involves joining the inner liner to the outer body in plastic coffee cups.

Despite the aggressive motion of the process, parts with walls as thin as 0.4 millimeter have been welded using this method. Maximum wall thicknesses are around 7 millimeters. Although at least one of the parts must be circular, the other, fixed part is not limited by this restriction. When designing for spin-weld applications, it is often necessary to incorporate a "drive feature," a means by which the welder will grip the spinning part. With smaller parts, friction alone may do the trick. However, larger parts will require ribs or a toothed surface that mates with the driving fixture.

Historically, another drawback to spin welding has been the inability to control the final orientation of the two parts after the weld is set. In fact, inertia spin welders using pneumatic motors afford no control whatsoever with regard to registration. Modern direct-drive, servomotor spin welders, on the other hand, offer good registration control-some to within a fraction of a degree-but at a price. A servomotor driven spin welder will cost anywhere from $35,000 to $50,000 and up. An inertia spin welder, on the other hand, can be had for as little as $15,000 or even $10,000.

Good Vibrations

All well and good, you say. But what if you've got a large assembly to weld and one of the parts doesn't happen to be circular-say an automotive instrument panel? In that case, you might be in the market for a vibration welder.

At first glance, this might seem the same thing as ultrasonic welding. In fact, the two are very different. Whereas an ultrasonic welder creates high-frequency vertical movements, a vibration welder generates friction at the joint interface through horizontal movements, with a frequency between 120 and 300 hertz. Usually, these movements are linear-the parts simply moves back and forth-although a limited number of applications call for orbital movements. The movements in vibration welding are also much more dramatic, with amplitudes ranging from 0.75 to 5 millimeters.

The upshot of all this shakin' and rattlin' is a process that is tough on the parts being welded and requires some very expensive equipment. A vibration welder includes a lower part fixture and a large upper platen and fixture, which holds the second part and is put in motion by a combination of springs and electromagnets. This platen-fixture arrangement can weigh anywhere from 2 to 100 pounds or more. Controlling all this action requires a robust framework, to say the least. In contrast to ultrasonic and spin welders, vibration welders are all fully enclosed because of the "god awful" amount of noise they make, as Kralovetz puts it.

This creates problems when designing automated assembly systems, because parts cannot be easily positioned and removed with a conveyor or indexing table. This movement also takes it toll on parts, so much so that delicate assemblies or assemblies with thin, unsupported walls can be shaken to pieces by the process.

"If you're putting finishing nails into a piece of trim, you're not going to use a sledge hammer," Naumovski says with regard to fine assemblies and vibration welding.

Nonetheless, vibration welding, which accounts for about 15 percent of the welding market, does have its place. Again, it can be used with large parts, it works well with most all thermoplastics, and like spin welding, it is not so adversely affected by semicrystalline materials.

In contrast to ultrasonic welding, energy transmission when welding with vibration is relatively straightforward, making part geometry less critical. And although the joint needs to in one plane, it can be angled as much as 15 degrees. Cycle times are slightly longer than with the spin and ultrasonic processes, but still as little as 8 seconds for smaller parts. Throughput rates can be improved by welding multiple parts with a single platen, an option that obviously doesn't exist with spin welding.

It is for these reasons that with some large parts and large production runs-like those in the automobile industry, for example-vibration welding is worth the investment.

The Heat Is On

Because of its drawbacks, hot-plate welding might initially seem more trouble than it's worth. The machines are not especially cheap, starting out at around $35,000 for a basic model. Operating costs are also fairly high, thanks to the energy involved and the fact that each assembly requires its own heating platen, which must be routinely cleaned of accumulated melt. Beyond that, the process is relatively slow. Low-temp welds, in which the heating platen is set at 500 F or less, have cycle times of 30 seconds to a minute; high-temp welds, in which the heating platen is set at above 500 F, generally have cycle times of 15 to 45 seconds.

As if that were not enough, the process will create fumes and smoke, especially at the higher temperature range, and tooling changeovers can be problematic. This is because the platens are large, have to be allowed to cool before being removed, and then have to be brought back up to temperature before beginning production.

Still, hot-plate welding does have one big advantage over the competition: It is very flexible and easily lends itself to welding multiple assemblies.

Do you need to weld two large, irregularly shaped parts with a joint that includes complex curves and other geometries, say, some parts for a dish or clothes washer? No problem. Is the part made from a semicrystalline resin that isn't lending itself to either ultrasonic or vibration welding? No problem. Do the parts include thin fragile structures that would go to pieces in a vibration welder? No problem.

Hot-plate joining also creates a smooth bead with virtually no particulate. As a result, it often finds applications in medical equipment assembly.

In hot-plate welding, the two parts are pressed against the heated platen to melt the plastic. The platen is then removed, and the two parts are pressed together to make the weld. Depending on the application, the platen can be vertical or horizontal. A vertical orientation makes it easier to load and unload parts manually; a horizontal platen lends itself more easily to automated processes. A horizontal orientation is also best if there are loose components in the lower half of the assembly prior to welding.

Low-temp welding works well with polypropylene and creates minimal fumes, making it acceptable in a clean room environment. On the downside, a Teflon coating is often necessary to prevent resin and fillers from sticking to the platen surface.

High-temp welding is used for welding polyethylene and nylon. It usually creates smoke and fumes that need to be ventilated. Fillers that build up on the platen can be scrubbed off with a wire brush, although automatic cleaning systems are available with some systems.

Noncontact welding refers to those procedures when the thermoplastic parts are held close to the heat platen, but don't quite make contact. The process is extremely rare and only used when constant resin or filler buildup on the heat platen makes it the only viable option.

Ultimately, although less than 10 percent of welds are created using the hot-plate method, the process is there when you need it for those troublesome assemblies that just don't want to go together using any other kind of weld.

Beam Me Together, Scotty

Finally, there's laser.

First, the downside: It's expensive, and although in theory any thermoplastic can be laser welded, the process is currently used with only a limited number of materials.

That said, although laser welding makes up only a tiny fraction of today's welding activity, it will inevitably grow in popularity. There are a number of reasons for this. For one thing, laser welding is extremely precise, and entirely flash and particulate free. It is also energy efficient and fast, with cycle times of 2 seconds or less. It is a noncontact process that leaves no marks on the finished product. It can also reach recesses that are inaccessible to other welding methods.

At its most basic, a laser welder uses a glass plate to hold two parts together, and then directs the laser through the plate and a transparent upper part until it strikes an absorbent layer. When this layer heats up, it melts the surrounding plastic and creates the weld. This process is known as transmission welding.

Originally, the absorbent layer would be comprised of the lower piece of the assembly fabricated from tinted plastic, while the upper part had to be laser transparent. This, not surprisingly, put severe limitations on laser welding's use. In recent years, however, new near-infrared absorbers from companies like BASF (Florham Park, NJ) and Gentex Corp. (Zeeland, MI) have made it possible to perform welds on parts that are clear or only lightly tinted. These absorbers can be applied to the joint area or incorporated into the plastic itself, opening the door for the use of laser welding in a number of new medical and automotive applications.

With laser spot welding, the laser is focused to a single point, which is then directed along a preprogrammed path to create the weld. According to Jerry Zybko, general manager for Leister Technologies (Itasca, IL), the ideal spot size for this kind of welding is 1 to 2 millimeters, although spot sizes can vary from 0.6 to 2.5 millimeters, depending on the application. The advantage to this method is flexibility. Almost any welding path can be programmed into the welding machine, which can direct the beam using robotics, a moving stage, or a system of mirrors and servomotors. Once the programs have been entered into the controller, changeovers from one part assembly to another can be performed with the push of a button.

Less flexible, but faster is simultaneous line welding, in which the laser light is directed, or collimated, along a straight line. Typical weld dimensions are 1 to 2 millimeters by 30 millimeters, with a cycle time of 1 to 2 seconds. Multiple lasers can be used to create square or rectangular contours. Optics also exist that can create curved lines.

With mask welding-a proprietary technique developed by Leister Technologies-the laser configuration is similar to that with line welding, only the line sweeps across the entire part, which has been masked so that only those portions left exposed will melt to create a weld. The technique can be used to create extremely precise and complex weld patterns. Applications have included sensors and microfluidic components in medical diagnostic devices. Weld lines as narrow as 100 micrometers have been successfully made. Masks are produced via a photolithographic process involving metal-coated glass.

Laser welding has also been used in automotive sensor applications, and it is beginning to make inroads in automotive taillight assembly. In each case, the process is able to justify its cost by producing precise welds without any particulate matter or flash.

Finally, Leister Technologies has recently developed what it calls its Globo Welding process. The system looks a little like an oversized ballpoint pen, in which the laser passes through a clear ball about 1 inch in diameter. As the ball passes over the weld area, it applies pressure and focuses the laser energy at the same time. According to Zybko, there are two main benefits of the Globo system. First, it facilitates the welding of 3D shapes, such as tail light lenses. Second, it provides an alternative to using glass or other clamping methods, because the necessary clamping is provided by the ball. This ensures that the parts are in intimate contact when and where the laser is emitted.