By Jerry Fulkerson
I’ve had a chance to tour several machining facilities in the last month or so. That’s always a fun time for me because exploring new ways of doing things is built into every engineer’s genetic code. As someone who has spent much of my career working with cutting tools, I’m especially drawn to what tooling is being used and how well it’s performing. That’s crucial because perishable tooling performance is one of the areas where an engineer can make large improvements over a short time and at minimal cost. Finding the correct tool and fine tuning that tool to its optimum machining parameters directly affects not only tooling cost but also OEE by improving part quality, machine cycle time and up time of the cell.
So, I was surprised to find that, of the seven plants that I toured, only two had permanent standards in place for tracking cutting tool performance at the machine. Of the other five, one tracked nothing, three had evidence that there was a procedure for tracking specific tool tests and one, according to my host, had the ability to track overall tooling cost for the factory. That’s all useful information, but woefully incomplete in today’s continuous improvement environment.
If optimization of the process is your goal then baseline data from your current state is essential. It’s not enough to do a quick snapshot based on current tool performance over the life of three or four tools. There are too many variables and they can change too quickly for that. Tool life and failure mode can be affected by coolant condition, chemical composition and physical properties of the raw material, condition of the machine, condition of the work holding and a host of other factors. The current state of your tooling must be assessed over a long period of time so that the data isn’t skewed by short term performance spikes with an assignable cause.
The best way to do this is to log every tool change at the machine noting piece count, failure mode and any unusual machining conditions or problems. Most companies do this with a simple pre-printed log sheet that hangs on or near the machine. If you’re working in a paperless environment (that’s a good engineer, go get yourself a donut) then the log can be a spreadsheet or, in a best case world, a relational database which gives exceptional ability to customize queries and reports.
Besides providing long term baseline data for tool tests, a well-planned tool performance monitoring program also gives the machine operators, the supervisors and the engineers an effective visual management tool for quickly identifying problems. For example, let’s say that a particular tool has over a period of time shown that it typically produces 400 parts and the failure mode is normal flank wear that results in increasing the surface finish to an unacceptable level. Now, let’s say that an operator coming in glances at the log entries from previous shifts and notices that the last six tool changes have happened after 175 to 250 parts have been produced and that the failure mode was a chipped edge. Since the feed rate and surface speed haven’t changed, that’s an unusual situation and he asks his supervisor for guidance.
After going through the trouble shooting chart hanging on the work bench (y’all have one of those, right) the operator and supervisor working as a team determine that the chip formation has changed. Instead of breaking up into nice 6’s and 9’s, the chip is stringy and hanging onto the part long enough that it’s sometimes getting caught between the cutting edge of the tool and the work piece. Zero surface footage is a problem and in this case it results in the chipped edge. To contain the problem, the operator quarantines the parts produced since the problem appeared and the supervisor arranges to pull the material to have it tested for compliance. Then he replaces it with material from a different heat to see if the issue improves. With the new material, chip formation and tool performance return to normal and production resumes.
So, what happened here? If the tool in question was a negative trigon, then you lost two to three edges on the insert and the time it takes for two to three tool changes. Even if you add in the probable loss of one part per chipped edge, the total monetary damage isn’t too great.
What did we gain by having a monitoring procedure in place? First, we saved time. Without the log it might have taken another three or four shifts to notice that there was a serious problem. That means more lost production, more scrap parts, more lost tooling and the potential that the tool holder might eventually be damaged from the chipped inserts. But, there’s something else we gained that is huge. If that material does have a chemistry or physical properties problem then we insulated the customer from receiving a manufacturing lot of non-conforming parts. Trust me when I say that the last thing you want to do is call a customer and tell them that there is a raw material problem. The containment costs and corrective action from that call will haunt you for a very long time.
But, let’s return to the original topic, optimization of the process. After the best machining parameters have been established through experimentation and documenting the results using the log, optimization is going to lead to tool testing between competing tool vendors. Here’s an example. Going back to the tool that we just discussed, let’s say that we are currently using a negative trigon to turn the outside diameter of the part. It’s basically a micro grain C-5 carbide substrate with a relatively new multi-layer coating. Brian, our favorite cutting tool salesman, stops by with a pack of his new Ultra Coat Nano Cermet inserts and swears that he’ll get twice the tool life and drop our cycle time by 30%. We like Brian, but we have to prove his claims before we switch, especially since the insert costs 20% more.
In years gone past, the decision would have been simpler because cermet and carbide had their own niches. Cermet liked light cuts, high surface speeds, low feed rates, no interruptions and a stable thermal condition. Carbide couldn’t run as fast or give equal surface finishes, but it was tough and more dependable in harsh machining environments. Times have changed. Cermets are more durable now and less picky about machining conditions. Coating and substrate improvements allow carbide to produce better and more consistent surface finishes and to approach cermet speeds. The lines are starting to blur and formal tool testing is absolutely required.
If we have a tool monitoring standard in place, we’re ready for that. We don’t have to scratch something on a piece of paper and then quickly train the operators how to collect data on the existing tool because we have reams of historical data at our fingertips. We casually walk out to the machine, make a notation on the log sheet that a test is in progress, hand the inserts to the operator, change the machining parameters to Brian’s instructions and move onto the next task while we wait for the results. Sixty edges later, we will either know which insert is the best choice or whether further testing is warranted.
That’s a simplification of course. The parts produced with the new insert will probably have to go through the Quality Department for a dimensional layout, we may need to complete a statistical capability study and it’s possible the that parts produced will have to be held for customer approval, but those things are standard anytime a process change occurs. The key thing is we are doing a controlled test in a structured manner and we were able to initiate it without doing anything on the fly.
In this example the test was initiated because a salesman brought in a new tool that he was excited about and that’s a good reason to test. But, it’s not the only reason. In today’s highly competitive manufacturing environment it’s imperative to constantly be testing somewhere in your machining operation. Tooling changes and improves too quickly to ignore that need. If you don’t find that “hot” new tool when it hits the market, I promise you that your competitors will and you’ll be scrambling to play catch up. Seeing the backside of your competition as they pull away from you is never a good sight. That’s especially true if the new tooling not only reduces tooling costs but also results in increased productivity.
P.S. The type of tool testing I described here is suited to the high volume production typically found in automotive, aerospace or medical applications. The folks who live and work in a low volume, high setup job shop environment have a much more difficult time doing controlled tests. I’ll touch on the procedure for doing that in another post.
Next up – A three part series on specifying and buying machine tools