In God We Trust. All Others Must Bring Data.
Post by Mike Gugger
There is an issue with a process – you can’t hold tolerance or the surface finish is less than desirable. Someone – usually the machine operator – decides that he/she will change the speed to correct the problem, but the results have either no effect, a negative effect or a small, but inconsistent, positive effect. So then he/she tries a feed change, a new tool or maybe a coolant change. Still only producing a marginal effect, and not necessarily positive.
Next, the machine operator calls in the “go to” guy. You know who I am talking about – that person that has been at your company for years, seen it all, tries this, changes that and adjusts something else. But still no significant impact! Meanwhile, all this time, part after part has been made that either does not meet spec, needing secondary processing of some kind, or is outright scrap. All the adjustments and changes made haven’t been documented and things are so far off from the initial situation that no one really knows what is going on any more.
In God We Trust. All Others Must Bring Data.
From operators to set-up to supervision to manufacturing engineering – everyone’s decisions have to be based on data and facts. To work “seat of the pants”, or to try “what has worked before” without the data and facts to back it up is simply shooting in the dark. You might hit something, but it won’t be a full solution and you won’t know why it worked.
Recall your high school science class discussion on the scientific method. Can you remember? The scientific method states that you should hypothesize, gather data, test and evaluate. Now take a large leap forward to the work of Dr. W. Edwards Deming. Deming’s improvement cycle – PDCA: Plan, Do, Check, Act. Sound familiar? The scientific method and the PDCA cycle are, essentially, the same thing.
Use your deductive tools to evaluate what is going on. First step is to ask why… why is that surface finish so poor? Why is that feature bouncing in and out of tolerance? Why is…
Then form a hypothesis, “I think the speed feed, tool geometry combination can’t give me the surface finish required.” Calculate the theoretical surface finish (ask me if you’re not quite sure how to calculate) and then determine if it is even possible to get to the required finish. Then, and only then, after a strong evaluation of the data and the facts, should you attempt to verify your hypothesis.
Verifying your hypothesis is the D or “do” stage in Deming’s PDCA cycle. Make the change you have information on, and do so in a manner that you believe will provide the most effective results; not just some helter-skelter decision, but one based on knowledge and understanding. If it works great! If it moves the needle in the right direction, good! But what else? If it doesn’t work, or makes things worse, change things back to the original state and develop another course of action.
Sadly, I see it all the time… a scatter shot approach to machining issues. Don’t do it! First evaluate the data and facts then work through the process. “Oh we don’t have time for that!” Really?! But you have time to try six or seven different things in a hap hazard manner?!
Do it right the first time. Scrap and time will be saved and this new problem solving method will become the standard approach.
