Intro: Welcome to the E6S-Methods podcast with Jacob and Aaron, your source for expert training, coaching, consulting, and leadership in Lean, Six Sigma, and continuous improvement methods. In this episode number 66, “In the Eye of the Cash-holder” Part 2, we cover the Taguchi loss function. Variation is evil to the whole of society. Here we go. http://bit.ly/E6S-066
***In the Eye of the Cash-holder***
Objection 1: In spec is good enough. Variation doesn't matter.
Counter 1: Any variation from target results in depleted function of your customer. Although customer specs are wide, customers often have to account for supplier variation in some other way, and end up releasing lower quality goods into the market. Variation is evil according to taguchi, not just to customers, but to society as a whole.
I Process Variation relative to Specifications: How Variation is Evil to Society
a. Taguchi Loss Function;
i. Variation from the center/target results in a parabolic “loss,” increasing by the square (x^2) with each incremental move away from center.
ii. When near the spec limits, the losses more closely resemble “total loss” than minor loss. The closer to target, the lower the loss.
iii. Any departure from the nominal value results in a loss!
i. Loss at a point: L(y) = k*(y-m)^2 where,
k = loss coefficient
y = measured value
m = target value
ii. Average Loss of a sample set: L = k*(s^2 + (pm - m)^2) where,
s = standard deviation of sample
pm = process mean
iii. Total Loss = Avg. Loss * number of samples
d. Application: A company that manufactures parts that require a large amount of machining grew tired of the high costs of tooling.
i. To avoid premature replacement of these expensive tools, (maximizing tool life):
1. set the machine to run at the high-end of the specification limits.
2. tools wear down until the products end up on the low-end of the spec limits.
3. the machine produced parts that fell just inside of the specs, both ends. Starting high and drifting down, then shooting up sharply. (saw-tooth pattern)
4. Process variation takes up entire spec range, rather than optimizing for minimal loss to produce the highest quality part possible.
a. Products may fall within spec, but will not produce close to the nominal.
b. Several of these "good parts" may not assemble well, may require recall, or may come back under warranty.
5. The Taguchi loss would be very high.
a. Phone batteries that lose charge too quickly
b. Unpredictable internet or satellite signal
c. Low potency antibiotics
d. Gaskets that lose their seal
ii. Is the savings of tool life worth the cost of poor products? Would it be better to replace the tool more often to reduce part-to-part variation? (penny-wise, pound foolish)
e. Products that barely meet specifications are shipped and fail after customer purchase.
i. This causes negative feedback from customers, lost loyalty & brand
1. losses to the manufacturer
2. ultimately society
ii. loss function can determine costs to fix defects before shipment vs. after purchase (society). https://controls.engin.umich.edu/wiki/index.php/Taguchi_quality_loss_function_and_specification_tolerance_design
f. Variation is not free; not even when all outcomes appear to meet specifications.
i. Consumes resources, money, and time.
1. Investigating OOS needlessly
ii. Close enough is not good enough; nor is being satisfied with meeting specifications. http://www.ivtnetwork.com/sites/default/files/Meeting%20Specifications.pdf
Outro: Thanks for listening to episode 66 of the E6S-Methods Podcast. Stay tuned for episode number 67 for introduction to the wonderful world of capability indices, “In the Eye of the Cash-holder” Part 3. If you would like to be a guest on podcast, contact us through our website or tweet us @e6sindustries. Join our discussions on LinkedIn. Subscribe to past and future episodes on iTunes or stream us live on-demand with Stitcher Radio. Leave a 5-star review while you’re there. Find outlines and graphics for all shows and more at www.E6S-Methods.com. “Journey Through Success”