The Problem Child in Automation | Part 1 of The Problem Child series -

The Problem Child in Automation | Part 1 of The Problem Child series

May 30th, 2021 / bshah

By: Peng-Sang Cau | VP, LS Emerging Markets and Symphoni™

Creating the Problem Child

For those of us who work in automation, we’ve known for decades who the problem child is: assembly manufacturing.

We all know a true problem child when we see one. He’s not just noisy or fussy or ill-mannered. He’s not just the kid who throws temper tantrums on the floor of the grocery store because you didn’t buy him a whole pint of ice cream. No, what defines a problem child is that his terrible behavior is persistent and seemingly impossible to fix. The problem child isn’t just the opposite of a model citizen. He’s the kid who can never become one—or so are made to believe.

For those of us who work in automation, we’ve known for decades who the problem child is: assembly manufacturing. As his siblings grew up and made money, the assembly process remained immature—like a college dropout who spends his days emptying the fridge and playing video games in the basement. For so many businesses, assembly line automation is rigid, conventional, wasteful, and far too expensive.

But why?

In the first of this three-part series, I want to answer that. Why did the rest of the manufacturing world embrace automation while assembly processes lagged behind? Why are assembly processes so challenging for pre-engineered automation solutions? Why did we give up on the problem child?

Over the last 30 years, the evolution of automated manufacturing systems has led to the proliferation of pre-engineered work cells around the world. This in turn has led to significant growth of automated systems in every facet of manufacturing. Virtually all redundant tasks have been automated. Beyond the need for capacity and quality, buyers have turned to pre-engineered systems because they:

Offer the lowest total cost of ownership;
Increase machine utilization per square meter of factory space;
Shorten build and delivery lead times;
Decrease technical risk by acquiring equipment with a proven performance track record;
Can redeploy the automation to produce multiple or different product types.

The benefits of pre-engineered automation have therefore transformed processes such as machining, injection molding, labeling, packaging, case packing and palletizing. The reason that assembly-related processes have been left out comes down to the four key characteristics of pre-engineered automation solutions in manufacturing.

First, pre-engineered solutions generally focus on common processes: cutting, molding, filling, weighing, labeling, and bonding. These processes are associated with myriad products across many industries. As a result, suppliers are motivated to come up with solutions that can be deployed broadly with the highest margins.

Second, they tend to be associated with a single value-added process. For example, CNC machines are dedicated to cutting metal, while labelers only apply labels to products. So long as it is clearly defined and understood, such a singular focus makes it easier to design a standardized solution, including hardware and software, that can be used over and over again.

Third, pre-engineered solutions generally have a single input brought to a single entry point. For example, in an injection molding machine, plastic resin is supplied to the injection unit on the molding equipment. Any pre-entry processes are not considered part of the pre-engineered system—it is assumed that the required input will be present at the right location and in the right format. These processes may be pre-engineered solutions in and of themselves, but they are independent.

Lastly, the solutions must be redeployable. With hardware, most of the system never changes, but by using interchangeable hardware, or by enabling certain functionality, it can be customized. For example, cartoners are generally designed to handle a range of size and shape of cartons. The base machine remains the same but there are a range of standard options to address glue flap vs. tuck flap, single load vs. dual load, etc. Meanwhile, with software—whether it’s the user interface or under the hood—the recipe drives the solution. The software under the hood remains the same for all products, while the user interface allows the user to adjust what happens within the system specific to the product. For example, dispense time, pressure, and temperature can all be adjusted to deliver a different amount of glue, but still utilize the same gluing system.

All of these characteristics have created our problem child. Here’s how.

With assembly manufacturing, there is no single value-add process or a single input and single entry point. Yes, many processes may be commonly understood, but the way in which they are applied is usually specific to the product in part due to the nature and number of parts that are being assembled. Furthermore, assembly applications have unique characteristics that at first glance do not make them candidates for pre-engineered solutions: production rate, controls logic, and process order of operations. Because of this, the vast majority of automated assembly systems are custom-built and viable only at higher production rates. And the more flexible these systems become, the slower their output rate.

A robot packaged in a pre-engineered cell is powerful and flexible. But the physical features that make it flexible also tend to limit its speed. Consider its work envelope, which is made to be as broad as possible for maximum freedom with respect to workstation design and functionality. The greater the distance the robot must cover, the more time it needs to complete its cycle. Likewise, for a robot to do a wide range of repetitive, force-bearing manufacturing operations, it must have the requisite strength and durability. But strength and durability require mass, and mass comes at the expense of speed. The inverse is also true: a delta style “spider” robot has low-mass linkages that allow it to move at very high speeds, but it cannot do more than relatively light work.

The control methodology used in conventional pre-engineered systems further restricts their throughput. Operations are invariably carried out in sequential, indexed motion. Consider a pick-and-place operation: the pallet arrives in the station, the robot moves to and from the pallet, and the pallet releases for travel downstream. Each step is carried out in sequence, and no motion can start until the previous motion is completed. The actual time involved in placing the part at its destination (where actual value is added) amounts to a tiny portion of the overall cycle time. The rest is non-valued-added time.

Finally, assembly processes change from product to product in three major ways: the types of processes, the number of processes, and the order in which the processes are executed. Pre-engineered solutions generally don’t execute more than one process; they focus on a single value-add. You can add or subtract single processes, or vary the order of operations. You can implement change tooling, like end-of-arm tools on robots, to accommodate a different part geometry. But, the production rate remains limited. If you add more than one value-add process with a pre-engineered solution, processing time increases drastically. You must sacrifice flexibility or speed. You can’t have both.

Until now, the assembly buyer has had to rely on custom automation companies and systems integrators. These systems, although reliable and repeatable, are costly, and they lack the flexibility and adaptability of a pre-engineered system. As products are modified or new products are launched, more custom systems are needed, and the buyer is caught in a repeating cycle.

It’s time for that cycle to end, and time for the problem child in automation to get booted from the basement. It’s time for Symphoni.

Read Part 2 of this series to find out how we overcame the false choice between speed and flexibility.