Comparative Design of Experiment to pick better Trimming Blade Supplier



Hi all,

I am a newbie here so please bear with me if I am asking the obvious stuffs:eek:.

I am currently interning for a medical device manufacturing firm and I am involved in this project that I do not know where to start.

Basically, we are trying to evaluate film cutting blades from two different suppliers:

Blades from supplier A: we are currently buying from this supplier. We have to change them every one or two week as they either get broken or blunt when there is production everyday (two shifts per day, each shift 12 hours).

Blades from supplier B: Less expensive of the two.

The problem is how do we design an experiment to compare a characteristic that would only show overtime like: sharpness or resistance to crack?

We can actually do a real simulation by putting both blades through a whole week of full force production run and then compare the state of the blades and products eventually. However, that would mean wasting a lot of materials and time.

I could not get the critical info about the blades such as blade sharpness index, just the physical dimensions. Their performance is judged based sole on whether products meet specifications and any other visual observation of the blades like cracks.

I have two major questions:
1) How to compare the sharpness of the two blades when they are brand new/after they are used for several runs?
2) Supposedly initially one blade is sharper than the other, does it mean it would continue to retain its sharpness over a longer time duration?

I am an industrial engineering student in year 2 and have no knowledge and experience in basic DOE yet. Any clue on this?



Hi thm219 and welcome to the forum,

Well for blade evaluation / optimisation, most of the characteristics or responses would only show over time. The benefit of a DOE would be you could minimise the number of runs.

I think that your film cutting process variables may be optimised with a DOE. Evaluating blades from different suppliers is different - loads of suppliers available and each supplier may offer a basic cheap blade or slightly more expensive blades with treatments to greatly prolong blade lifetime. Such as low-friction or tough-coat surface treatment or solid carbide micrograin.


Forum Moderator
I have two major questions:
1) How to compare the sharpness of the two blades when they are brand new/after they are used for several runs?
2) Supposedly initially one blade is sharper than the other, does it mean it would continue to retain its sharpness over a longer time duration
1) Do you have any means for measuring sharpness directly? Same for nicks and cracks? If you do, you could measure these directly at specific intervals and predict the time/cycles to failure using measures of the current blades at failure as a baseline.
2) No. The useful life is more dependent on the material properties. You could potentially have a very sharp, yet soft material that wears out quickly, and a less sharp, yet very hard material that wears out slowly.

Bev D

Heretical Statistician
Super Moderator
Miner is correct.
You say this is for a medical device company - depending on what tier you are in and what product perfromance characteristics are effected by the blade, as well as the level of documentation on the blade requirements, you may be restricted in what you can and can't do for validation. Subject to your regulatory reviewer of course.

What I have done in similar cases where it woudl take a long time to get life data and might waste product in the process is to:
  • demonstrate via independent laboratory testing that the material properties of the two blades are equivalent. You'll need more than 1 pair - I typically use a minimum of 3 of each.
  • measure an appropriate number of parts made with several new blades form the current supplier and several from the new supplier - perform a t-test, or a paired t test if there is substantial lot to lot differnces in teh material that effect the 'cutability'.
  • submit the 'equivalency of the change based on the material verification testing and the 'time zero' validation testing.
  • Request to perform 'concurrent' validation over time, shipping any product that meets current product specifications.

there have been cases where the wear is of no interest to the regulatory agency so I dont' request 'concurrent' validation, but do perfrom it any way to safeguard the company.

statistics cannot replace matter how much we wish it could

Steve Prevette

Deming Disciple
Super Moderator
A few thoughts:

Take a look at for some ideas on blade "goodness".

Sometimes sharper is NOT better - may have more of a tendency to slip/cut too easily and cause injuries.

One text to consider is taking a stack of materials that the workers must cut through, take a week-old knife and measure the force to penetrate a certain distance into the stack.


Involved In Discussions
Perhaps you could develope and run a surrogate test to determine relative wear resistance and durability. I'd envision testing them against a bench grinder (perhaps outfitted with a drum of material similar to what you cut) in doing so, you might be able to quickly replicate lots of use very quickly and not consuming actual items. As for how to test if they're still sharp, maybe the method suggested above about force and depth of cutting might work but I'd suggest running several replicates of the test as who knows what the repeatability of such a test is.

Top Bottom