SMED 3️⃣ External Work
Single-minute exchange of Die. Powerful methodology to reduce the changeover time. The critical part of this methodology is to move works that need to be done internally to externally. But a big mistake is that we focus so much on internal work, and external work is forgotten. Just because we defined something as external does not mean the work has disappeared from the plant.
External work requires more attention than internal work.
“Before everything else, getting ready is the secret of success.”
- Henry Ford
In SMED methodology, the internal work gets the first attention of Kaizen. Why? It is because the amount of internal work directly impacts the time that a machine needs to stop for the changeover. The shorter the internal work, the faster the impact. Yet, on daily operations, the success of the changeover largely depends on the successful external work. If the team has successfully prepared the area for the changeover, internal work will go well. It is easy and sensational to report that “the changeover time has reduced from two hours to thirty minutes.” But if we created the thirty minutes of external work, then that also needs to be reported and discussed systematically. Just like internal work has standardized work, external work should have standardized work.
I knew I hadn’t touched the closet when I determined to clean the room. So I moved everything out. Now, the closet is empty and clean. It was just that the room was a huge mess. Just moving the external work out will create such a mess.
One of the tricky questions of external work is who will perform the tasks.
An operator is a candidate. Yet, if we assign the standardized work to the operator, it must have “free” time to conduct the external work. “Free” is a nice word. The truth is that the operator needs to be wasting time waiting for the machine, which is not good. The operator should always be focused on the cycle works. And when I make such a statement, a mathematician will always say, “external work is 15 minutes (900 seconds). Our batch size is 1000. It’s less than a second. They can do the external work.” Yes, the calculation is correct. But it is difficult, almost impossible to do. Nobody will do the 1/1000 work and then rush back to the machine to complete the cycle work and return to external. If we assign external work, we must ensure we have a mechanism for accumulating buffers without sacrificing quality.
Another approach is to have a set-up operator. (The name of this position is different among factories.) If cost allows such a position, this sounds like a great idea. But this approach requires some attention. Such a position will likely cover multiple machines, and this operator cannot perform two external works simultaneously. Each external work will be performed in a sequence. Then, a machine downtime became a headache since every downtime could change the changeover sequence among machines. There are many places where they believe in 80% OEE, which means one out of five machines is always down. The potential change in the sequence is enormous.
The obvious answer is close to the machine. But there are some additional considerations required. And this depends on who.
If an operator is responsible for external work, then it is necessary for some space for the buffer. And this buffer should always be visible to the operator. If not, the operator will face painful reluctance between cycle & external work. It’s a disrespectful way of designing work. Therefore, if we assign it to the operator, both works must be visible. This means two works should not be on another side of a machine.
On the other hand, if we choose to have external work on a setup person, the responsible machines must be grouped and easy to access. I have been to places where they assigned the machines at the four corners of the building. The thinking was that those were the same technology. Yet, distancing responsibility will create some blind spots. It is not a fair condition.
Also, there is some thought that since external work is external, it is okay to have waste, such as walking long distances, searching for materials, etc. Such thinking will lead to chaos. Even if it is external work, we should not have “waste.”
The ideal timing to do external work is immediately after the previous changeover. Once internal work is done, do the external work for the next. By keeping this sequence, the space for external work should always be occupied with the next tooling, which will become a good visualization for management. There are situations, such as consecutive changeovers for the setup person, which prevent external work from being done immediately after the changeover. But those are exceptional cases. If we choose not to perform the external work continuously, we need a mechanism to highlight the timing to start the external work. But this is tricky. What if the responsible person can’t begin the external work immediately? Some flexibility is required, but the general rule should be to perform the external work immediately after the changeover.
Please refer to my previous post.
Factories that use SMED properly will focus on external work. And differences in external work will appear clearly in layout, people, and management. Unless you see some differences, internal work will fail.