In Part 1, we talked about the “why” of quick change.  In Part 2, we started the discussion of “how” to get to quick change.  In particular, we talked about gathering some information about the current state of setups and changeovers, e.g., how long they take and how they’re carried out.  In Part 3, we’re going to talk about what to do with that information.

As I mentioned, in the last post, we talked about the sort of information you should gather about your changeovers and setups.  In particular, we recommended that you track how long they take and what goes on (especially, what goes wrong) during the changeovers.  We also recommended that someone watch a few changeovers to get a general idea of the process.  Let’s talk about that second one first.

You can observe with a stopwatch in one hand and a clipboard in the other, if you like, but I recommend against it.  True, there might be some times when you want to see how long a given activity takes but your wrist watch will do.  We’re not looking to cut seconds here.  Rather, we’re looking at the big picture with respect to the changeover.  In particular, we’re looking for external steps that are being carried out as internal steps.  External steps are those activities that can be carried out before the changeover starts or after it’s completed.  They are steps that can be carried out while the machine is in operation.

Let’s say that an operator shuts down the machine and then goes to get the tooling for the next job.  That step of going to get the tooling should be done as an external step.  The tooling for the next job should be staged next to the machine before the previous job ends.  I’ve also seen operators take the previous job’s tooling back to storage before installing the next job’s tooling.  The operator should wait until after the next job is running to return the tooling.  So opportunities to move current internal steps outside the changeover, i.e., to make them external steps, is the first thing you’re looking for.  (If you want to time these steps to get an idea of how much time you’ll reduce the changeover, that’s fine.)

The second things you’ll be looking for are needless steps and delays.  The operator has to get tools from a toolbox that’s 50 yards away.  The operator has to look all over the shop for die clamps.  The operator gets called away to help someone else on a different changeover or with an operating problem.  The first part is taken to the quality office, where it sits for 30 minutes before being approved.  Again, if you want to time these delays, go for it, but, whether they’re a minute long or three hours long…they’re needless delays and you’ll want to identify and eliminate them.

A third element of the changeover that you’ll pay attention to is the work done to get a good part after the tooling is in place.  In many changeovers, this is what takes the most time…actually changing the tooling goes pretty quickly, then getting a good part takes another two hours.  Once again, there’s no harm in tracking the time all this takes but you’ll also want to gather info from the operator that will help you understand why he or she can’t get a good part right away.

When you have your observations, sit down with the folks who actually do the setups and develop the new procedure that:

1. Moves as many activities as possible to external steps, e.g., getting tooling, putting away tooling,
2. Reduces travel and delays,
3. Reduces the time it takes to get a good first part.

Then, of course, you roll out your new procedure.  As you do, keep these points in mind:

1. Make sure that continued measurement of changeover time is part of the new procedure and that it continues to be carried out.
2. Go observe a few changeovers using the new procedure to assure that, you know, the new procedure is actually being used. This is important; associates often return to old habits and procedures simply because they are familiar with them.
3. Talk with the folks carrying out the new procedure to see if they have any further ideas for improvement.  It’s infrequent that a perfect procedure is developed the first time.