Dear Gemba Coach,
We applied SMED to increase the flexibility in our production cells. Now our dies are crashing like crazy and disrupting production because maintenance can’t keep up! Got any recommendations or experiences in applying lean in the tool and die shop?
Yes, this is not unexpected. Let’s reason it through. Why do we keep dies glued to the machine and do long batches? Because dies are hard to settle in until they make good parts.
Die changeover is measured from last good part to first good part. When you start changing dies more often, two things happen:
- First, you get a lot more start-up bad parts as you have to get the production right more often.
- Second, you handle the dies more and they can get damaged in the process.
“Die problems are not die problems – they are human problems. They are mental problems.”In any case, it’s quite clear that changing dies more often is more hassle. Or is it?
In fact, the fragility of the dies has always been there, but as long as we’re doing long batches, it’s hidden deep in the process. Rather than paying for the weakness of the dies, we prefer to pay the cash layout of having parts in inventory and not meeting spot customer demand.
What you’re facing is a cost trade-off. Do you prefer to put your money in:
- Inventories and late shipments
- Die maintenance and operations
Of course, in the company, this will not appear so starkly because these costs come out of different budgets. Inventory is still often considered as an asset, so the plant will be rewarded for having produced a lot of parts – even if we don’t sell them. Late deliveries are a bummer but don’t come out of any particular budget. On the other hand, spending money on operator training or die maintenance is a cold hard cash outlay, right now. Guess what most managers will prefer?
Changing dies more often is about learning to learn: it will teach you what you need to learn – if you’re listening.
Dies as Pets
Before we go any further in solving the problem, we need to look into it in more detail. The deeper question is “who needs to learn what?” No one would crash a die on purpose, so:
Die crash = something we don’t know
A learning curve is structured around mental steps:
- Unconscious incompetence: We don’t know what we don’t know – and often convince ourselves (and tell others) we don’t need to know more than we already do.
- Conscious incompetence: We’ve figured out what it is we need to know that we currently don’t know, and it’s rarely a happy prospect.
- Unconscious competence: Usually, just by opening a book, talking to an expert, looking into real cases, scratching our heads, things get better as we remember our basics and discover new ideas.
- Conscious competence: We know how to explain what we do better – and so can communicate it to others and standardize that knowledge.
Dies are crashing. Dies can’t learn. Only humans can do that. The first step is connecting every die with a person by creating a table with these headings:
|Most frequent problem
Treat your die as a pet. It’s ill, we need to figure out what’s wrong with it. That means watching it attentively and taking note of everything that could be wrong in its lifestyle:
- Manpower: How does every person who touches the die actually handle it? Do they understand its soft spots? It’s cranky habits? Do they know what finishing touch to apply after each movement to make sure the die feels happy?
- Materials: Is the material being pushed through the die appropriately? Are each of the die’s components, connections, and transport jigs OK? Are we close to the cliff edge by using dies in dodgy conditions or are we fine?
- Machine: Is each machine applied to the die for its maintenance (and, yes, that means the cleaning cloth as well) operating as it should be? When did we last check?
- Method: Is the theory to set the die in the press correct? Have we checked how we set the machine and thought about conditions of pressure and temperature? Are we reflecting on physical processes to understand what is going on?
Die problems are not die problems – they are human problems. They are mental problems. The place to start is your troubleshooting methodology. Get your guys together and ask them to share their sequence for troubleshooting – where they start to look and then where do they go, and so on to see how they intuitively prioritize problems.
Then get them to compare sequences and explain why they go about it one way or another. Bear in mind that when two experts disagree, 1) they’re probably both right (but they’ve generalized specific conditions differently), 2) tempers may rise as thinking stirs up emotions and experts feel quickly challenged. So be sure we have frequent breaks for chit chat and sharing a coffee, a soda, or a soothing tisane (whatever works in your culture).
And then get everyone to leave with a commitment to a study topic – not a corrective action, just something to check. And a date to meet again.
Your question is really not about dies but how to handle technicians when things get rough. Our cultural reflex is to get them to diagnose the problem; then organize the issues; then come up with corrective action plans; and then monitor to see that these plans are followed. This usually makes things worse.
Experts need to be listened to, feel they are heard, and that their unique experience is valued. They’re rarely good at communicating what they know and the frustration of not making themselves understood makes them crabby. The first step to solving your die crash problem is to realize this is a social problem, not a SMED problem.
How can you create a social setting conducing to technicians being interested in why the dies crash rather than having to correct it immediately or cover it up? Start with “why?” Performance will follow.