SMED - almost funy

There was a real event that prompted me to write this article. The case took place in the year 2012 in Warsaw in a pharmaceutical company, where we should deal with a blistering machine, which packs tablets into blisters and blisters into single packs. During the SMED workshop somebody said the sentence that can be cited now as an anecdote, and it was spoken by a novice Lean specialist.

What is SMED?

SMED - Single Minute Exchange of Die, or rather Single Digits Exchange of Die would be better, it is a changeover in the time of single minutes.

Instead of arguing whether it should be a single minutes or a minute, one should assume that SMED is a process of continuous improvement in regards to changeover time. Practice shows that in the case of processes that were not subjected to SMED workshop, one can count on a changeover time reduction of at least 30%, often even above 50%.

SMED stages according to Shingeo Shingo:

Separation of activities into external and internal ones.
Transformation of selected internal activities into external activities.
Continuous improvement of all steps needed to carry out the changeover.

The changeover time has been treated in the past as an unknown or constant, while it may be: controlled, measured, managed and improved. To the point 1. I would add from himself - removing unnecessary activities.

However, I would like to return to the anecdote. Then, in Warsaw, the following sentence was pronounced: "Our changeover is not long, just the setup takes a long time".

Well, what is funny about that? Here we would have to come back to the definition of the changeover time. The changeover time is the time that elapses between the completion of the last part of the previous batch (before a changeover) and the completion of the first correct part from a new batch (after a changeover). Let us clarify this definition with example. If we "changeover" something within 20 minutes and then it takes 40 minutes to set it up, of course the changeover takes 60 minutes. There is also the question of correct part. We can talk about the end of a changeover only in the case of the production of parts without defects. The production time for defective parts should be included in the setup time, which is included in the changeover time.

Separation of activities into external and internal ones.

External activities are those that can be performed while the machine is running, e.g. substitution of new production parts, notification of technicians. Internal activities, in turn, must be done while machine is down, e.g. mold replacement, stamp replacement etc. Important - what is now an internal activity may in the future become an external activity. We strive to ensure that as many necessary activities as possible should be or should become external activities.

How to start the assessment of activities type?

Video recording(s) would be invaluable. The rule is following: as many video recordings as there are people doing a changeover. If one cannot make a video recording, first of all one has to try to convince employees to do so and involve them into SMED workshop. If it fails, then the laborious taking of notes remains, necessarily with reference to the time, e.g. 10:30 - start of changeover, 10:32 - turning off the machine, etc. The accuracy of the time recording depends on the changeover time. The shorter the changeover time, the more accurately we have to record the time. When you have a video, you do not care about accuracy, as there are programs that can show even hundredths of a second.

How to record data analysis?

Work Combination Sheet, Line Balance Sheet, Gantt diagram can be used. I will not pick on here. Let everyone use what they like, let them do it on the computer or on a sheet of paper. The most appealing to me is the Gantt chart as presented in Fig. 1. I choose the computer or paper version depending on the group I work with or (which happened) I do the analysis alone. Fig. 1 shows the changeover performed by one person. There are individual activities in the lines. The horizontal axis is the timeline. External activities are marked in blue. The time unit is 1 min.

Transformation of selected internal activities into external activities.

In this step, one should approach each internal activity critically. One needs to ask the following questions for each such activity:

Can this activity be transformed into an external activity?
What needs to change for this activity to become an external one?

It is very likely that you will meet with resistance, somebody could say that it has always been this way. In such a discussion, examples of other changeovers with non-standard solutions would be invaluable. It can be very helpful to lead the discusion in form of brainstorm to generate out-of-the-box ways to transform an internal activities into an external ones.

Example: let us assume that we are doing a changeover of a mold in a packing machine which is packing yellow cheese. The heated mold causes the bottom foil to take the shape of this form and cheese can be packed into the bottom foil shaped in this way, which is then closed with the top foil.

What type of activity is mold heating? It will probably be an internal activity, because we put the mold into the machine, connect the heating element and wait for it to heat up, time passes and the machine does not produce.
How to transform this activity into an external activity? Well, it is not possible! How about a little out-of-the-box thinking? One could heat the mold outside the machine. In Formula 1, the tires are heated before assembling, why cannot this be done with the mold? It remains to take care of safety so that the hot mold does not burn the employee who is replacing it. In this simple way, we can save from a few to several minutes.

Continuous improvement of all steps needed to perform the changeover.

There are many ways to improve activities needed to perform changeover. I strongly advise against rushing, because the point is not to do something faster, but it would have to be done smarter.

A few things to pay attention to:

How many tools does one need for a changeover?
How much time does one spend looking for the right tools and parts? Maybe one needs to organize the parts warehouse and the trolleys used for a changeover?
Could there be fewer tools, e.g. by unifying joints?
Could there be joints which would not require tools?
Could there be used power tools?

Have you noticed who is involved in the changeover? The most common situations are:

A changeover done by technicians (maintenance workers) only.
Less often: by technicians and operators (i.e. people working on the station subjected to a changeover).
Quite rarely: only by operators.

In case 1, we irretrievably lose a resource that cannot be recovered - time. Operators do nothing or do something that does not increase productivity, such as keeping a close eye, but nothing will be faster doing that.

Everything must be done to involve the operators. I understand. Soon someone will say that they are not trained, that there is a risk of an accident, etc. ... And who said they had to do complicated and risky operations? Maybe they could do them in the future? They can start with simple operations i.e .:

part replacement,
documentation replacement,
program replacement,
shutdown of parts of the machine etc.

What do we gain having a shorter changeover times?

What to do with the time obtained by reducing the changeover time? The answer seems to be obvious - one can increase productivity, that is, one can produce more in the same unit of time using the same resources. One can also significantly expand the product range (more variants at the same time, e.g. in one shift) and this in turn will contribute to a reduction of WIP (Work In Process). On the one hand, production batches can be reduced and, on the other hand, parts in buffers associated with a changeover can be reduced too.

Final thoughts

In SMED, I look at time as Michelangelo looked at his sculptural material. Apparently, Michelangelo, when asked how he made the sculpture of David, said something like this: "I took a block of marble and removed unnecessary fragments." Time is my sculptural material. I take such a solid block of time and remove unnecessary fragments. Sometimes I move some of the leftover fragments to another place so that an effective composition is created. It means optimal from the perspective of time.

There were several SMED workshops in my career, but the most important ones are following:

Blister machine - the one from the anecdote.
Labeller on the bottle filling line - lots of fun, looking for parts etc.
4-stroke laser welding robot, where over an hour changeover time was shortened to about 30 minutes. The success of the changeover was important to convince a few people to the analytical approach and to Lean in general.
Foil extrusion line - a changeover of 6-meter rolls needed for texture. It was an important changeover, because I showed a younger colleague who was just starting out in Lean, how to use knowledge in the field of SMED in practice and this is how I made him a believer in Lean.

Each changeover was from a different fairy tale, in a different industry, in different companies, but we managed to apply everything that Shingeo Shingo said and in each case the improvement in the changeover time was 30% or more.

I encourage you to check your knowledge in the SMED test. Please do not worry about the lack of a set of points, as some questions clearly relate to the elements of my original SMED training and contain several neologisms.

PS
What would you say to the publication of an e-book that would describe in more details SMED topic and include some full case studies with SMED analysis and solutions applied?

Author: Adam Cetera (LeanSigma.pl)
Creation date: 2018-11-05
Modification date: 2021-09-20

Leave a comment below if you would like to add something?