Our Sites

Gauging welding’s success

Robotic welding cell helps set the stage for improvement for pressure gauge manufacturer

Ashcroft Inc.’s Stratford, Conn., plant employs 400 people, and many have been making some of the company’s mechanical and electronic instrumentation here for decades. But in recent years they’ve witnessed significant changes on the plant floor.

James Hubbard, continuous improvement process engineer, and Prasad Sastry, lean and continuous improvement manager, pointed out additional space being earmarked for new product production (see Figure 1)—space created thanks to a reorganized assembly department arranged in five workcells. In three years the assembly operation has more than doubled its available capacity, and all the work is done in about 60 percent of the space.

With these stellar results, the company is moving forward with improvement efforts in other areas of the plant, including welding. Two years ago the company upgraded some of its automated welding equipment to improve quality, flexibility, and reliability. This occurred as part of an extensive plan to move away from batch processing.

That newly created space in the assembly department was just the beginning.

Some History

Ashcroft’s roots go back to the early years of steam power, which requires pressure and, for reliable and safe operation, an accurate way to measure that pressure. At an 1851 London exhibition, American inventor Edward Ashcroft met the French inventor Eugene Bourdon, who demonstrated how a curved metal tube deflected consistently under pressure. When the pressure changed, so did the amount of tube deflection, which actuated a mechanism that moved the gauge needle.

According to a history posted on Ashcroft’s website, “Edward Ashcroft was so impressed with the gauge design that within a year he acquired the patent rights to produce the French ‘Bourdon tube’ gauge in the U.S., and by 1852 founded the Ashcroft Manufacturing Company.”

Over the ensuing 160-plus years, Ashcroft became a global corporation and expanded its product lines to include switches, transducers and transmitters, thermometers, test instruments, diaphragm seals, and other offerings. It never stopped producing mechanical pressure gauges, though now they have a plethora of advanced features. But the gauges still use Bourdon tubes, and the fundamentals behind their operation haven’t changed since that fateful day at the London exhibition.

Flexible Automation, Tight Process Control

Making a mechanical pressure gauge at Ashcroft starts at a tube mill, where strips of material are formed and welded into a tube, which is then cut to length before being bent and flattened in a rotary bending machine. This creates the Bourdon tube, still at the heart of every mechanical pressure gauge. The tube is joined to a socket, and the components are heat-treated, powder-coated, assembled into the gauge case, calibrated, tested, and then sent out the door.

The gauges’ structural welds include one that joins a 3⁄16-in.-thick carbon steel back plate to a round, 7⁄8-in.-diameter socket that has a ¼-in. hole drilled in the middle to make an assembly about 3 in. high. Previously most welding on the back plate was accomplished by a custom welding machine with three gas tungsten arc torches autogenously welding simultaneously.

This hard automation was built for a specific product family, so it was expensive to adapt to low-volume parts. One such part, the gauge tip, required skilled welders to join it to the subassembly. The hand welding wasn’t terribly difficult, but it was inefficient and didn’t make best use of the company’s welding talent.

Figure 1
This space at Ashcroft’s Connecticut plant, freed up thanks to a lean manufacturing project in the mechanical gauge assembly area, is now earmarked for new products.

About three years ago, the hard automation began to show wear, because of the increased volume of parts it was handling, so the Ashcroft team opted for new equipment to improve the automation, thereby increasing flexibility and its ability to handle additional volume.

Choosing a Welding Process

The old method of hard automation was built to weld a specific range of parts, and that was that. This time the team wanted something flexible—an articulating-arm robot that could be adapted to more products. But they also wanted to maintain the level of throughput achieved by those three GTAW torches.

The full-penetration, circumferential weld between a stainless steel round socket and a carbon steel base plate is about ¼ in. from toe to toe. The weld needed to be clean, with no spatter, which was the reason that the welding team went with gas tungsten arc welding. But GTAW is slow. Could the team achieve the same level of control and weld cleanliness with gas metal arc welding (GMAW), a much faster process? Could one robot using GMAW achieve the same throughput of three GTAW torches?

This is where Cold Metal Transfer (CMT) came into play. It’s an iteration of GMAW developed by Fronius International, which describes it as a short-circuit process in which droplets detach from the wire in a new way. In conventional short-circuit GMAW, the wire advances until a short circuit occurs. At this moment the welding current rises, which allows the short circuit to open and ignite the arc again. But two issues sometimes make conventional short-circuit GMAW problematic. First, a high short-circuit current produces high heat. Second, the uncontrolled way the short circuit opens can contribute to spatter.

According to literature from Fronius, “The digital process control detects the short circuit, then helps to detach the droplet by retracting the wire ... During welding, the wire moves forward, and as soon as the short circuit happens, it is pulled back again.”

Hubbard and his team decided to go with CMT for its high level of arc control, which can virtually eliminate spatter, and its ability to control the arc characteristics to produce a strong, deep weld (see Figures 2 and 3). “Some of these gauges go through some pretty rough service,” Hubbard said. “They’re sometimes on equipment that vibrates. So we’re proud of the fact that our gauges stand up to abuse.”

This, he added, is why strong welds are so important.

To design the automated cell, Ashcroft turned to Seconn Automation, a sister company of Seconn Fabrication, a contract fabricator (and 2008 FABRICATOR® Industry Award winner) based in nearby Waterford, Conn.

When testing the cell, Seconn at first used a 0.035-in. wire, which seemed like the natural choice, considering the small weld size. “Welding sometimes is a black art,” said Skip Swift, Seconn’s vice president of automation. “No matter how much you learn, it’s the part and the resulting weld that tells you what you can and cannot do. We spent quite a lot of time in testing using 0.035-in. wire, thinking we could keep the bead as small as possible, and we could maintain the strength we needed.

“But to consume that 0.035-in. wire, we ended up putting more heat into the part, because it took longer to weld,” Swift continued. “And this is not a normal joint.” The root is a very narrow V-groove, while the rest is in a shape that resembles a fillet (see Figure 4).

“CMT allowed us to make a larger-diameter 0.045-in. [309 stainless steel] wire work like a small-diameter wire,” Swift said. “The wire didn’t have to feed into the weld at such a high speed, which gave the liner and other consumables a longer life. We were still able to get the penetration we needed, but because it filled up the weld so quickly, any issues we had with cracking and other problems went away. We wanted to get in there with a nice hot weld [for penetration], and then we wanted to fill it up quickly and move on, so we didn’t put excess heat into the part.”

Figure 2
Ashcroft replaced a three-torch GTAW system with a robotic GMAW workcell using Cold Metal Transfer. The cell has a setup in which the system recognizes the fixture for a certain job and calls up the appropriate program.

Errorproofing a Quick Changeover

Considering Ashcroft’s lean implementation, the welding team also needed the automated cell to change over quickly between jobs. To keep the pace of the previous system, and feed downstream operations exactly what they needed at the right time, engineers wanted to eliminate as much waste as possible between operations.

Seconn developed a system of six fixtures that handle all the part subassembly iterations, including the previously hand-welded gauge tip. Each fixture has its own welding program tied to it. So if the operator loads fixture No. 4, the system calls up the corresponding program automatically. “There is no programming on the part of their operators,” said Swift. “They just change the fixture, and the robot automatically calls up the part program.”

Sensors on the fixtures—specifically, a group of proximity switches, all enclosed for protection against sparks and debris—send a signal to the robot to tell it which program to call up. “We installed a set of six sensors, and each fixture lights up a specific sensor and calls up the program,” Swift said. “Then we wrote a small software program to talk to the robot controller.”

To errorproof the fixture design, the team made fixtures with pins and other components designed to clamp onto the parts for the associated job, and nothing else. This prevents the operator from choosing the right fixture but the wrong part, and vice versa.

Swift added that joint variability presented another challenge. Because of the nature of some upstream manufacturing processes, the weld joint’s position can change slightly from one part to the next. Welding robots can have touch sensing, which essentially turns the tip of the welding wire into a touch probe. But this joint is so small and precise, conventional touch probing just wouldn’t do.

“We needed to be absolutely on the money,” Swift said. “As wire comes out of the end of the GMAW gun, it has a cast to it, and it doesn’t come out in precisely the same position every time.” Moreover, the act of touch probing, which measures via ohmic resistance (touching positive to negative current), requires the robot arm to creep slowly forward toward the joint, which is too slow for this high-production application.

For this reason, the ABB robot has a dimensioning laser mounted near the welding gun. If the laser finds that a part edge isn’t within a specified window (in this application, it’s ±1 mm), the robot’s program doesn’t start. As Swift put it, “With the laser, we never miss, and we can move at much higher speeds.”

Swift summed up the entire setup this way: “We basically made the system smart enough so that it will be able to take on a variety of parts that the [hard automation] couldn’t handle before.”

More to Come

Sources added all this will open the door to even more improvements in the welding department, including smaller batch sizes and less work-in-process. This may well include an additional robot in the welding area, an application that will involve dissimilar-metal welding with joints requiring a homogenous material makeup. All this is happening with significant lean transformation projects in other areas that produce electrical switches and thermometer product lines.

The lean journey continues and, in reality, never stops. “I don’t think you can ever be as lean as you can be,” Hubbard said. “You make it as good as you can make it, let it run, then go back and see how you can make it even better.”

Figure 3
The lean team that worked on Ashcroft’s welding project includes, from left to right, Prasad Sastry, lean and continuous improvement manager; Maki Tillman, continuous improvement specialist; James Hubbard, continuous improvement process engineer; Kevin Rodgerson, fabrication cell leader; and Christopher Costello, continuous improvement process technician.

Lean Assembly

Years ago the mechanical pressure gauge assembly area at Ashcroft’s plant in Stratford, Conn., looked dramatically different. Back then one assembly person received a work order and brought the gauge through the entire assembly process. This seemed like a logical approach, because Ashcroft’s pressure gauges can be customized for the application; there can be thousands, even millions, of variations. How could assembly possibly be standardized with so many different products?

Still, this “workshop” style assembly was a major problem. Final assembly was a major constraint, and it affected the throughput of the entire plant. According to Prasad Sastry, lean and continuous improvement manager, this is why his team decided to work on assembly first, where the company would see the biggest improvement in the shortest amount of time.

The assembly operation needed flexibility, and this came in the form of workcells. “After studying the data and performing a demand analysis, we discovered that all the product variations in the mechanical gauge line could be grouped into five product families,” said Sastry. “They are grouped by different pressure ranges, different types of connections, and similar attributes.”

Cross-trained employees work in U-shaped assembly cells, each with five or six stations. During busy times, each station is manned; during less busy times, a person moves between several stations. The team identified gauge calibration as the constraint, simply because of the nature of the process. There’s no simple way to shorten calibration time. Being the constraint, gauge calibration always has a dedicated technician, because that sets the production drumbeat for the entire cell. (Eli Goldratt, the man behind the theory of constraints, would approve.)

So how has overall assembly lead time improved? “It used to take four to six days to complete a unit,” Sastry said. “Now, as soon as a gauge lands at the first assembly operation, it is finished within three hours.”

About the Author
The Fabricator

Tim Heston

Senior Editor

2135 Point Blvd

Elgin, IL 60123

815-381-1314

Tim Heston, The Fabricator's senior editor, has covered the metal fabrication industry since 1998, starting his career at the American Welding Society's Welding Journal. Since then he has covered the full range of metal fabrication processes, from stamping, bending, and cutting to grinding and polishing. He joined The Fabricator's staff in October 2007.