But most companies don’t realise that some of their existing technology can be utilised even before a new CNC machine has been delivered. By Bryan Jacobs, Marketing Communications Manager, CGTech.
When a company purchases a new CNC machine tool, the process is usually chronological: a specification of the requirements is created, a machine is selected, and after negotiation a delivery date is set. Once the facility is prepped and the machine is delivered, the shop then begins the process of creating new part programs and first part prove-outs are run on the new machine. By the time the post-processor is fine-tuned and production parts are being shipped, many valuable months have passed.
There are better ways. By using proven g-code driven NC verification, simulation and optimisation software, manufactures can gain the confidence that the first part will be good, often before a new machine has even been delivered. Better yet, simulation can help ensure that the machine specifications are correct. Making a mistake during the specification phase can turn out to be very expensive.
Shops must determine what features are important to simulate with a new complex multi-axis machine tool. To illustrate, following are two examples of shops using simulation well before receiving a new machine.
Proving out the process prior to machine delivery
Aerospace Manufacturing Group (AMG) operates a state-of-the-art manufacturing facility in California. They are supported by the latest technology in hardware, software and quality control, giving them the capabilities to manufacture the most complex parts in the aerospace industry. Their investment in technology makes them a trusted source worldwide, and their products can be found on commercial and military aircraft for companies like Boeing, Lockheed Martin, and Northrup Grumman
AMG recently added a Scharmann Ecospeed F to their existing line up of manufacturing capabilities. The Ecospeed F specialises in high-speed machining of medium sized aluminium structural components. It has a horizontal spindle, vertical pallets and a high speed Sprint Z3 parallel kinematic machining head and 150 tools. “It’s only been in and running for a couple weeks now,” said John Gates, AMG’s Director of Engineering. AMG purchased the Ecospeed F for a specific job with a family of parts. Although not at liberty to say what that job is, parts range in size from 68”x52” to 60”x200”. The new machine cuts metal very efficiently, and feeds up to 2,000” a minute. If there was any problem in the code with a tool path or possible impact the operator would never be able to stop it. A 10” cube of aluminium can be reduced to nothing but chips in two minutes with the new Ecospeed.
The NC programmers at AMG use Vericut simulation software. They use it daily to simulate all of their machines and it was instrumental in getting the new Ecospeed F up and running as quickly as possible. “The new machine and control are quite a bit more sophisticated,” explains Gates. “That’s really saying something considering the technology we have on our floor. I needed to write the post for the Ecospeed while it was being built, and before it was installed in the shop. We couldn’t afford to wait until it was in-house before getting to work.” Through the use of the simulation software Gates was able to debug 95 percent of his post before even seeing the machine in person. He estimates that they were up and running at least a month ahead of where they would have been without it.
“No program ever goes to any machine without completely going through Vericut. The code I send through Vericut is the exact code that goes out to the machine,” said Gates.
Vericut shows material removal at the work piece level and simulates entire machine tools as they appear on a shop floor. The program also simulates NC machine controls and supports advanced control functions to reduce the possibility of a machine crash. Machine simulation detects collisions and near-misses between all machine tool components such as axis slides, heads, turrets, rotary tables, spindles, tool changers, fixtures, work pieces, cutting tools, and other user-defined objects. A user can set up near-miss zones around the components to check for close calls, and detect over-travel errors.
“You can zoom, rotate, pause, whatever I want to do,” explains Gates. “To put it in perspective, if the average shop has a piece of 12”x12”x4” aluminum and you screw it up, that sucks, but you can go buy another one and move on. We are working on a piece of aluminum that is 260” long, 60” wide and 4.5” thick. It costs thousands and thousands of dollars, not to mention the months that it will take to get another piece like it. We can’t afford to make mistakes like that.”
Using simulation to determine machine specifications
The R&D engineers at a large aircraft engine supplier are experienced at developing simple solutions to complex challenges. One such challenge involved developing the manufacturing process for the leading edge of the fan blades used in the fastest-selling engine in company history.
This engine powers the most advanced commercial aircraft in the world. The engine uses super high bypass, composite materials and specialised coatings, and a special fan blade that blends form and function. The blades' composite material gives the engine a rare combination of unprecedented power and low noise and is combined with a titanium leading edge for extra protection. This creates a lightweight and durable combination.
The R&D engineers were tasked with developing a cost-effective solution to manufacturing the titanium leading edge of this revolutionary new fan blade. To realise this project, a group of engineers, designers and programmers were assembled. They had at the time a 5-axis milling machine available. Traditionally, the leading edge is made using a patented grinding process developed by an outside vendor (single source). The engineers sought to make the leading edge using a simple 5-axis milling approach, but needed a way to prove to management that it could work. The solution was to show management the proposed process by creating a video of a simulation.
To prove that a milling process could work for this leading edge, every element was tested in the virtual environment. The 5-axis machine available in the shop at that time had a 41” work envelope so the first step was to determine if the leading edge would even be possible in the current machine. Every piece was validated in the simulation: the complete machine, all the fixtures, and many new tools that would have to be designed.
The final leading edge is a 3’ tall V-shape with walls down to .010” in thickness. The inside of the V is up to 5” deep but about 5/16 to 3/4” wide. The inside and outside of the V part is cut in small levels from top to bottom, with each level completing a roughing and finishing pass before proceeding to the next one. There are more than 200 levels. Multiple tool lengths are used to ensure they always utilise the shortest, most rigid cutter.
Their biggest programming challenge is that the CAM software does not know anything about the remaining material when generating 5-axis motions. They used simulation to detect the holder and tool collisions inside the V. Some of the holders come as close as .010” to the part wall. When a collision was detected, they then created a new surface in the CAM software, reprogrammed the 5 axis motion to move the holder away from the wall, and verified it again. This process was repeated hundreds of times.
The simulation ran 16 hours a day for 2 ½ years to work out the details of the process. It would have been impossible to test new processes on an actual machine. It gave the engineers the freedom to try every idea, no matter how impossible it might be. Only after the process was proven in a virtual environment was it was taken to the actual machine. This process has worked so well that they purchased new machines with the goal of shaving the machining time of each part down even further. Again, the process was first proven in a simulation.
To develop the requirements for a new machine, the engineer started with models of a new machine most closely matching his requirements. He then used simulation software to alter dimensions of the machine to his specifications. Once his process was again proven virtually, the models were sent to the machine tool builder. The machine was configured to the exact specifications.
Simulation allowed them to test the latest machine technology, without having the actual machine in place. Without simulation, they would have been forced to buy the identical machine that they proved the process on. Even with many the changes to the process, there has not been a single machine collision, and only a minimum amount of scrapped material.
Not all simulation is equal
Every CAM system offers simulation, and most do so by licensing a faceted material removal engine from a third party supplier. The simulated cut stock from these systems is a bunch of triangles. This might be fine when there are flat surfaces, but multi-axis motion through material can result in a whole lot of triangles. The more you cut, the more triangles get created. The animation speed is typically fast initially but will slow as material is removed – especially when the triangle tolerance is set small enough to detect small gouges that can catastrophically ruin the work piece. On long part programs it is even possible to run out of memory and fail because each block increases the amount of RAM used. Vericut’s unique algorithm provides fast, accurate results. Performance does not degrade with increased cuts, so Vericut can process programs with millions of cuts and virtually any type of material removal technique. The amount of memory used is constant from start to end.
Another problem with the faceted material removal method is that analytic measurement accuracy is associated with simulation tolerances. The larger the tolerance, the faster the simulation, but you lose accuracy on the cut stock. OpenGL can be used to create shading tricks to mask imperfections, but measurements will still return results that don’t accurately reflect the nominal values in the part program. For example, interpolated hole sizes will depend on tolerance and the measured diameter will be different depending on where you pick.
As worldwide competition increases and customers demand more from manufacturers, it is critical that shops have a full arsenal of tools to ensure that the parts they produce are right the first time and new machines are up-and-running as quickly as possible. Simulating CAM output to view basic workpiece material removal is not enough. In order to survive, shops must operate as efficiently as possible; modern simulation and optimisation software has become a valuable tool to minimise the cost and time of production while increasing product quality. It has evolved into an important process that protects and frees up CNC machines, eliminates the prove-out step and even helps companies select the right CNC machine for the job.
