To Stop or Not to Stop, That Is the Question

Author – Tim Smith

And to go on from there, “Is it nobler in the mind to suffer the slings and arrows of outrageous fortune.”

Shakespeare could well have been describing the “western” use of Andon screens, lights, monitors, flags, whatever.

The term Andon comes from Japanese paper lanterns used for lighting since the Edo period. As such, it is one of the different traditional paper-covered lanterns and lights, others being Bonbori (雪洞) or Chōchin (提灯). In manufacturing, the term Andon (Japanese: アンドン or あんどん or 行灯) refers to a system which notifies managerial, maintenance, and other workers of a quality or processing problem. The Andon was used as an integral part of the Toyota System whereby the line stops and the team triages, identifies and solves the root cause of the stoppage. The understanding was that a new line may stop more than run in the beginning, but as time progressed and all the root cause issues were dealt with the line would ultimately see very few stoppages. The Toyota system embraced even multiple Andon lines or cords of different colour, each indicating a severity level. But in most western manufacturing operations the Andon signals a stoppage where only the symptom is dealt with so the line can start again. In the West, a popular way to combat production problems is traditional firefighting: find somebody that can fix the symptoms so the work can continue. However, in most cases, the usage of the Andon is poor to nonexistent. Worse though is the ongoing unspoken attitude that the line should NEVER stop. This is a decades old western concept. It is ingrained into our way of operations and it’s difficult to break. A bigger challenge is that, where Toyota would have one team lead for every 4 to 5 operators, we usually see one team lead responsible for 10, 20 or 30 operators along with ongoing engineering issues and filling the coffee machine. You get my point. We cannot solve all the manufacturing world’s problems at one time, but what we can do is make use of an excellent tool by making sure that the tool selected can help mitigate stoppages by assisting in determining root causes. Most packaged Andon systems are basically an alert to a stoppage. They are normally only transient in function and do not employ collection and storage of ongoing operational data. A disco light would be more helpful. In fact, the majority of such systems are basically ignored after a number of months. The value of the Andon as a feature of an operations system must do more than just alert to a stoppage. Ideally, every station at the value stream has its own Andon. This way, the system knows not only that there has been a problem, but where the problem occurred. An Andon system has the most impact on unbuffered and limited buffered lines where each station will affect the other stations downstream. But Andons in a discrete manufacturing environment are critical because a breakdown and an individual machine for  a length of time will negatively impact upstream and down stream operations across a number of departments or value streams, wherever the machine plays a critical part. System related Andons can report on so much more, status of the different processes in the system, and an actual and target production value. It may show defect quantities, efficiency, or material supply issues.

One of the important advantages of Andons is the fact that leadership and Maintenance aren’t required to constantly monitor the line which allows them to do other work (like quality checks) until someone calls for assistance.

If a stoppage occurs or a threshold is crossed, then the system should:

  1. Alert that there is a stoppage or throughput issue.
  2. Request additional information from the operator if required.
  3. Associate the stoppage to a correct down state or production condition.
  4. Initiate an e-alert to the appropriate department as to the down state reason, be it hydraulic, electrical, mechanical, process or materials failure, or threshold alert. Determine the known category to which the stoppage or threshold alert belongs to, such as machine, job, shop, maintenance, or system.
  5. Return to normal operation once the stoppage is cleared or threshold corrected.

The manufacturing operations management system has recorded all of the related data, so the line can start again.

But now the work can start on determining the root cause of the issue. Since timely and accurate information was autonomously collected and the operator was polled for their insight, the issue can be reviewed to determine whether it is chronic or acute. Wherever it is categorized determines the type of response to the issue and the correct engineering group to perform the analysis. A solution is proposed and executed and whatever the outcome, it is initiated, and the ongoing data collection will provide insight as to whether proposed solution corrected the issue or not. If the solution corrected the issue it becomes part of the SOP.

The Andon cannot be used as an isolated alert to a stoppage and be effective. The Andon is only effective if it is one part of a holistic operations system considering the machines, operators, work orders, operations, and product output.

I have seen lots of money and time wasted on expensive Andon lights that nobody wanted, nobody had time to think through, and nobody ever looked at. Unfortunately, most of the alerts in a closed loop Andon system are not handled by an automatic system, but by a human. This is where most (Western) Andon systems fail. If a system depends on an operator to initiate a stoppage or condition or make lengthy commentary as to the source of an issue. Such responses get relegated as non-priority and at best are objective.

A stellar Andon system is a part of a larger operations system which monitors and manages the shop floor. Then stoppage and threshold alerts are autonomous, and an operator only needs to briefly interact to classify an unknown condition which take a couple of seconds. Even if he doesn’t the time and duration are recorded and will enable subsequent analysis and response.

Take care in assessing a system appropriate to your operations. It should be much more than blinking lights and never dependent upon an operator initiating a stoppage or condition.