Adopting new technology takes a change in shop mindset
October 14, 2008
Modern press brakes add intelligence to the machine control and bring programming offline.
Adaptive bending, either through laser sensing (pictured here) or internal devices, can help overcome problems with material variation. Photo courtesy of LVD Strippit, Akron, N.Y.
The press brake remains one of those last bastions of manual operation. Even the smallest job shops may use a welding robot, perhaps a small-footprint sheet metal cutting cell with automated material handling. But it's not uncommon to see most if not all brakes in the bending area programmed on the floor, which means much of the time, press brake rams aren't moving.
Nevertheless, with a little upgrading press brakes can have uptimes comparable to their cutting-center counterparts. Many shops start upgrading by buying repeatable press brakes with graphical, user-friendly controls with built-in intelligence, drawing from extensive material databanks and effectively automating a lot of press brake moves that before required manual calculation.
As Steve Claude, president of Santa Fe Springs, Calif.-based Durma explained, "The operator doesn't have to calculate the bend sequence. [The control] will calculate it for him and show him how to bend it before he even bends it."
Peter Kruger, product manager for Cincinnati Incorporated, put it this way: "In the controller there are user preferences. If you use a grade of mild steel, you select it and the material thickness, and it knows the particular mild steel has a springback of 5 degrees, so the controller automatically adds an extra 5 degrees to the part to achieve the right angle."
Shops may invest in precision-ground tooling and perhaps tool changeout systems, like hydraulic clamping bars. (A high-end press brake with nonprecision tooling is a bit like a Lamborghini with bicycle tires; it looks pretty, but it won't go anywhere.)
Beyond this come advancements on the software side that account for behavior of material that's actually at the machine, along with software that takes programming off the shop floor. Adaptive features can account for material variations during the bend, through strain gauges or other internal sensing, or with external laser sensing (seeFigure 1). Machines can account for deflection and crowning through mechanical or hydraulic devices in the lower beams. And offline programming can reduce setup time to mere minutes.
All this sounds impressive. But according to sources, shop managers should take care or risk making a big investment only to find it little used on the shop floor. The new technology requires not only a willingness to try new things, but a shift in thinking about how work actually flows on the shop floor.
Controllers have vast material databases to calculate springback and deflection before bending a particular grade and thickness (see Figure 2).But what if the metal's tensile strength varies significantly? What if a bend is placed next to a hole, and the heat-affected zones from laser or plasma cutting left some hard spots in the metal?
"Modern press brakes in the field all demonstrate very high repeatability and accuracy," said John Kemp, bending product manager for Bystronic, Hauppauge, N.Y. "Today it's become more of a challenge adapting to material changes than just achieving high repeatability. With software and some hardware built into the machine, we're better able to compensate for differences built into the material to move us toward that first-angle-correct ideal. Being able to compensate for machine deflection errors and calculation errors on-the-fly has come a long way."
Material grades have a range of thicknesses, "and with the price of steel today, many suppliers are going down to the lower limits of that range," Kemp explained. "And when you're bending with a very small V die, a small change in thickness can have a big effect on an angle." A 0.002-inch thickness change in a 3⁄8-in. V die, for instance, can change the resulting angle by more than a degree.
Tensile strength variations also enter the picture. "We have a tensile strength tolerance [range] from the mill, but we don't know what the exact tensile strength is. So if the machine doesn't compensate for a change in pressure because of that tensile strength, we'll see that as an angle discrepancy in the finished bend," Kemp said.
Say an operator or setup person programs a bend operation, entering the thickness, the type of tooling, material, and assumed tensile strength. The machine may calculate 40 tons to bend the part, which means the brake needs 40 tons worth of crowning and frame-deflection compensation. As the bend starts, the machine, through its sensors, can detect a change in tensile strength. "Up to 20 percent variation in tensile strength is not uncommon," Kemp explained. "So now, your tonnage requirement goes from 40 to 48 tons. If the machine doesn't change the crowning and frame deflection on-the-fly, then you will see an underbend, because you didn't apply enough crowning to the part. Many may attribute this to a difference in springback, but in most cases springback is only a very small part of the problem."
The deflection adjustment happens through various methods in the brake's lower beam, either through mechanical or hydraulic methods. With certain systems, the machine monitors tonnage continuously during a cycle, and when the machine senses a required change, crowning cylinders can engage on-the-fly and make the appropriate adjustment as the ram descends to bottom dead center.
Sources concede that adaptive bending doesn't always get a warm welcome on the floor. Operators may view the system as slowing them down, and there may be a reason that they're in a hurry. A press brake, programmed on the floor, has a lot of downtime, so when the ram is finally moving, operators want to move as quickly as possible.
Click here to view image larger An operator calls up a program created offline. Photo courtesy of Durma USA, Santa Fe Springs, Calif.
But what if operators didn't have that downtime to worry about? Enter offline programming (see Figures 3 and 4).
"Nobody would accept 60 percent nonproductive time on a laser cutting machine," Kemp said, "but for most people it's perfectly acceptable on a press brake."
"With the use of offline programming, the operator does not have to do the programming, which means the press brake remains in constant production," explained Duane Gibbs, TruTops software supervisor for Farmington, Conn.-based TRUMPF Inc. "Meanwhile the offline programmer simultaneously creates future work orders."
Most important, he said, software works with complex parts previously handled by only the most experienced press brake operators. "Consider today's complex parts—six, eight, 12 bends for computer chassis and other components. We can bring those off the machine, into the office," Gibbs added. "This can change the business incredibly."
Offline bend programming has its share of hurdles. For one, a designer may put an inside radius on a 3-D model, but it may not be the same radius that can be obtained with available tooling. If those bend radii are off, so is the size of the blank—so bending software must account for these complexities.[image4]
Kemp described a six-bend sequence to explain. "The error will usually end up on one flange. So if the entire part is 0.050 inch too big, and you bend the part with a standard program, you'll end up with one flange that's 0.050 inch too big. The operator tweaks the program so all the flanges are 0.006 inch too big, just to bring it within tolerance. And all this takes time and expertise."
But what if the press brake program was developed first offline, directly from a 3-D drawing, and before the blank size is determined? This way, the actual properties of available press brake tools can be taken into account, so the inside radius specified is the inside radius that can be performed on the machine. Once the bend program is finished, the part is unfolded flat and placed into a nest for the laser, punch, or plasma machine.
"The idea," Kemp explained, "is to give the operator a proper blank size so he doesn't have to deal with differences at the press brake."
As an added benefit, software produces a data history that can give a shop a leg up when quoting jobs. As TRUMPF's Gibbs explained, "Programmers are able to work with concept parts from start to finish. This gives them an edge when performing part feasibility studies prior to actually bending any parts at the machine level. Before the job is sent to production, the fabricator already knows if they own the tooling to make the part as well as the time required to produce the part."
An emerging benefit is what's called the batching or nesting of tools, where programs are made in such a way that the maximum number of bends can be made with one tool setup arrangement.
Cincinnati's Kruger explained, "If a certain number of parts require different tooling lengths, they can be staged accordingly. If you write programs using those setups, you can put those parts together in a batch, call them up, and run them accordingly."
Rick Pawell, a manager at Amada's verification and proposal planning division, explained it this way: "Let's say I've got 15 parts to run across this bending area tomorrow. That could mean 15 tool changes, which could take up to 30 minutes to do. That's seven hours of tool changes in an eight-hour day—not much productivity. But based on these 15 jobs, what's the fewest number of setup requirements? For these 15 tool changes, only four tool changes may be needed."
Truth be told, offline bending hasn't been embraced like offline software for cutting. Sources cited many reasons. For some, it's a manpower issue. Kruger mentioned one company where the off-line bending was handled by the engineering group, which happened to be two very overworked people. Kruger recalled, "They both said, 'I've got enough on my plate!'"
Resistance might come from the operator, because offline programming significantly changes his routine. Bending can be hard work, especially for large parts. "The press brake operator has to lift the part for every bend," explained Paul LeTang, senior applications engineer-forming, LVD Strippit, Akron, N.Y. Sending preprogrammed runs to him on the floor "basically takes away his break. For a typical press brake operator, half the day is spent programming while the other half is spent running on the shop floor." Regardless, he added that offline programming "is the wave of the future, but shop owners need to address issues with operators. They may perceive [the change] badly."
Durma's Claude attributed the resistance to offline programming to, among other things, the complexity of bending. "When you speak of a laser or turret punch, programming is relatively simple. You have X and Y coordinates, and any machine can handle that X and Y coordinate without exception and with precision. Bending is different. Bending has more variables, so consequently it's less appealing. Many times a shop may talk about [offline programming] at the beginning of an order, but the controller makes it a lot easier for the operator to write a program, so offline programming falls by the wayside."
A person programming offline should know his way around a brake. "I've seen a lot of companies make the mistake of having somebody with no press brake experience trying to write these programs," Bystronic's Kemp said. "The press brake operators don't like the programs they make, so they refuse to use it, and that defeats the whole purpose."
Strippit's LeTang added, "In some of the best shops I've seen, they take the best setup people from the floor and move them into a programming office." People without bending experience programming offline may spur excessive operator tweaking at the machine. "As soon as an operator starts opening up a program and tweaking it, you've let the cat out of the bag. He might spend some serious time changing things he doesn't like.
"Nevertheless," he added, "you're always better off giving more information to the operator than just the blueprint."
In fact, some changes may be unavoidable and necessary, Kemp added. "If you ask 10 different press brake operators how to bend a part, they may give you 10 different answers, and none of them would be wrong." For this reason, some of the latest controls allow operators to tweak certain things to obtain a bend sequence that's right for them.
Offline programming and adaptive bending could be seen as complementary technologies. Offline programming means operators don't spend shop floor time programming, and adaptive bending means operators don't spend time with angle corrections. The result is that setup times plummet, brakes spend more time bending good parts, and throughput increases. And as lot sizes decrease, reducing setup becomes all the more important.
Using both could make the operator's life easier too. By itself, offline programming eliminates press brake downtime—and those operator breaks. Adaptive bending by itself may slow an operator who's trying to get through a certain number of jobs during a shift. Put the two together, and the story changes.
It's a bit like the tortoise and the hare. The "hare" operators without off-line programming and adaptive bending can speed through work faster, then stop, set up, and reprogram the machine, taking a break before starting the strenuous work again. Tortoise operators work differently. They receive programs produced offline. The brake performs adaptive bending, which may slow the bending slightly (though the newest technology is minimizing such slowdowns). Regardless, every part is a good part.
And like the tortoise, that operator wins the throughput race.
Strippit's LeTang added that economic conditions have made conditions ripe for upgrades like offline programming. "Higher-volume jobs have disappeared or moved to China," he said. "We're now addressing smaller batch sizes, so making the part on the first try becomes more important. If you make 50,000 of something, it doesn't cost that much more to make 50,001. But if you're making one of something, and you scrap the first part and make two, you've doubled your cost."
He added that the technology essentially has flattened piece-part costs, so piece-part costs for a run of one is pretty close (though not identical) to a run of 100. This happens because setup time plummets and so becomes less of a factor in calculating the price of a job. "All the operator has to do is put the tools in according to predetermined instructions," LeTang said, "so you have setup times that are often less than a minute."
Kruger recalled one job shop manager who visited his office to obtain a trial version of his company's offline bending software. On the way north back to Minnesota, riding in the passenger seat, the manager loaded the software into his laptop, programmed a batch of parts, and was ready to start bending once he got back to the shop.
The thing is, the shop was in a Hutterite colony, similar to Amish or Mennonite enclaves that dot the Midwest and Middle Atlantic.
"They embraced the technology like you wouldn't believe," Kruger said.
The shop didn't have the latest and greatest equipment, and the company itself sits amid a culture embracing traditional, rural living. But that didn't stop the shop owners from thinking differently.