What is quality 6

[25362]

Although quality is easy to talk about in generalities, it's difficult to define in specifics.

It was Aristotle who said: "Quality is not an act. It is a habit."

Think for a minute what quality, or the lack of it, means to you, objectively as well as subjectively, and let us learn from your opinions.

 

[ewh]

If I need a tool for a home project, and I'm only going to use it once, I look for an inexpensive tool, even though I perceive it to be a low quality item. Therefore, for my purpose, it is high quality. Again, I fell that quality is subjective.

 

[25362]

MintJulep, in the watch example, wouldn't the word price (or cost) be more appropriate than value ? Wouldn't price/cost be one of the various overall quality characteristics, in this case a mitigating factor ? I think there is indeed a relationship between the two.

Value, like quality, would be a translation of our satisfaction, or pleasure, obtained from the acquired product, or service, more as a chelant (borrowing from chemistry) of quality and price. Kindly recall the quality triangle mentioned by TrevorP in his post of Jan. 25.

Do you agree ?

 

[BigH]

I saw this in a Thailand shop of sartorial splendor and I think that it defines quality (I've modified it slightly):

Quality is never an accident.
It is always the result of intelligent effort.
Thre must be a passion to produce superior products.

 

[MintJulep]

25362,

I will agree that "price" or "cost" would be a better choice of words for the purpose of my previous example.

However, just as there is no relationship between quality and cost, neither is there a relationship between price and value.

You may value the sponge Bob watch given to you as a gift from your daughter more than you value a Rolex you bought for yourself from the proceeds of a fortunate stock pick.

I disagree with TrevorP that there is a correlation between quality and cost. Take the Fctory A Factory B example from above.

Factory B sends only perfect widgets to market, but scraps many imperfect ones. Thus the cost of Factory B's widget reflects the cost of all the raw material that was thrown away and the costs associated with converting that raw material to scrap.

Factory A may let the occasional bad widget out the door, but the vast majority will be good, and much less expensive.

Assuming that Factory A and Factory B have the same standard for good and bad, I prefer the less expensive widget.

 

[BJC]

It may seem logical to assume something made to a higher standard would cost more ( that is, take more dollars, material and manpower to make).
But one defination of a good engineer is to do for $1 what any one else can do for $2. Isn't part of engineer doing more for less.
Value can be intrinsic and not calculated exactly, or if it's something like a machine in a shop I define it's value exactly.

 

[HVAC68]

Economics is part of Engineering. Designing or producing a product at the most economical value or "value for money" is one of the keys for success in Engineering.

HVAC68

 

[25362]

Since we already got to the point of cost vs quality, I thought it is the right time to mention Taguchi's (successful) attitude towards this subject.

Taguchi's (Japan's) approach to quality is based in the "total cost to society". When the consumer's cost would rise while the producer's cost is falling, Taguchi adds both and gets what is called the "loss" function. When this function reaches a minimum it quantifies, in fact, the optimum quality.

Loss could be quantified in relation to time lost, wear, ultimate strength, fuel economy, health problems, etc.

Following this phylosophy, if someone steals $10, the net loss to society is apparently zero; someone has a $10 loss while the thief has a $10 gain. If, however, the manufacturer causes an additional loss to society, everyone in society has suffered some loss.

Thus, a producer who saves less money than the customer spends on repairs has done something worse than stealing from the customer...

Take the case of Japan's position in the worldwide economic competition. It is an island with a large group of people but otherwise limited natural resources.

The method they selected to survive and thrive is to import raw materials, add to their value by processing, and export high-value-added products (cars, TV sets, etc.).

Their success is due in part to the efficiency of this process. A low loss to society fits neatly within the niche of competing effectively in a value-added product area.

In Taguchi's mathematical technique the loss function entails two aspects of quality management within a factory: the variance, and the center in the distribution.

It is the engineers' job to establish the variance before the start of production, and to improve it as time goes on by off-line quality control. The second is the responsibility of the manufacturing people on a continuous basis, i.e., on-line quality control.

In short, Taguchi's strategic view of quality differs from the more traditional "conformance to specification", in the sense that it stresses a greater concern for "uniformity of products" i.e., a narrower tolerance.

It considers the relationship of variability to customer desires, and looks for a method of quantifying the value of quality by the use of a loss function and a continuous reduction in variance (factory's tightening of tolerances).

The descriptive statistics being the average and the variance. One probable quagmire in this approach lies in the measurement (repeatability and reproducibility) and interpretation of data.

The book I mentioned in a previous post brings numerous examples where Taguchi's approach was successfully applied.

As it has been already said in previous posts, the cost to society is always lower when making the product right at the factory than to find a poor quality product when the customer has it in hand.

Better yet, it is always of a lower cost to society to make it right at the point of production than to have to rework within the factory at a later time.

 

[HgTX]

Jeff Dewar of QCI International once said:

If we accept 99.9% as our goal, we'd have to accept the following conditions:

• 2 unsafe plane landings per day at O'Hare airport
• 16,000 pieces of mail lost by the USPS every hour
• 22,000 checks deducted from the wrong account
every hour
• 20,000 incorrect drug prescriptions each year
• 32,000 missed ehartbeats person per year

That puts quality goal in perspective. Thus perfection should be the aim every way every day...

This completely ignores the notion of risk assessment, of fitting the criterion to the situation. Two unsafe plane landings a day is not acceptable. Finding a cracked egg in one out of every ten cartons I pick up is acceptable. With finite resources, perfection in all things at all times is not attainable, and priorities must be established.

Hg

 

[johnwm]

Jeff Dewar of QCI International once said:

Let's not forget that Jeff Dewar sells quality management systems....

Quality in it's colloquial sense (A Rolex is a quality watch) is rather different to quality in it's technical sense (Fitness for purpose to predetermined criteria)

Good Luck
johnwm
________________________________________________________
To get the best from these forums read faq731-376 before posting

UK steam enthusiasts:

http://www.essexsteam.co.uk

 

[haggis]

This has become a great topic!

Hopefully there are other members following this thread who can relate to the following with some item or other:

The story is, my wife goes to the supermarket for the groceries and they have a limited time special. A no name circular saw for $15.95. Wishing to surprise me, bless her heart, she buys one and presents it to me. To describe this saw, it's housing is made of a light white metal casting and it's inner workings, well, only time would tell how good or bad they are. My guess was that they would be good for maybe a dozen 2 x 4's at best. Oh well, the main thing is the thought was there.

That was 28 years ago and in the mean time has seen service on 3 rec rooms, 4 decks, numerous small projects and I'll be using old faithful on the weekend. The question is, is it a good quality saw? I think so, for it's intended use that is. If I was working as a framer day in and day out, this would not have been the saw of choice although there were no disclaimers on the box like "not intended for industrial or commercial use" as some products do have today, but it was obviously not as robust as a $200 DeWalt.

If I had have used the saw in an industrial environment and burned it out just after the one year warranty, would it have been percieved as poor quality? No, I would have been guilty of abusing it. Much like using a Chevette for the duties of a Humvee.

I think johnwm's quote Quality in it's colloquial sense (A Rolex is a quality watch) is rather different to quality in it's technical sense (Fitness for purpose to predetermined criteria) sums it up.

So, did I get a good quality saw?........Yes. Did I get value for money?.....Darn right I did (wish I could estimate in cents per cut and compare it to a continously used industrial saw)

Lastly, every time I use it, I remember how I came by it. This also makes it priceless.

Haggis

 

[IRstuff]

I think that Taguchi's emphasis on reducing variance is not the end, but the means to the end.

A process that is highly variable is difficult to optimize and improve. Consider how difficult it is to fix intermittent failures. Every time I've run across that type of problem, I wind up sitting around, waiting and hoping the failure becomes permanent, so that it can be more easily found.

One of the hallmark results of control of variance is that miniscule optima can be and are found.

The apocryphal, but supposedly true, story of Mazda building engines for Ford resulted in Mazda delivering the engines which turned out to be unusable by Ford. Turned out Ford's own production lines were highly variable, resulting in mix-and-match process of selecting engines that would "fit" with equally variable transmissions. Since the Mazda-built engines had extremely low variance, Ford could not get more than a few percent of the transmissions to match. The moral of the story is that Ford wasted a lot of time and energy downstream to accommodate poor process control. They could have been more efficient and faster and more profitable.

Additionally, Mazda pointed out that since their own process was so well controlled, they could shift the design point so that the engines would be significantly more reliable. The variance in the Ford process would have never let them find that result.

TTFN

 

[MintJulep]

IRStuff,

A star for you because you correctly used "variance" instead of "tolerance".

 

[GregLocock]

OK, here's a case where we deliberately engineer an error into one system in order to control another input.

Tyres always generate a once per revolution force, as the bit where the belt overlaps is stiffer than the rest.

So, in some cases, we drill the wheels off centre, and match the low point of the eccentricity of the wheel to the high stiffness part on the tire. That's what some of those blobs of paint on your tire mean.

The same idea is used elsewhere, in many balancing or goodness of fit situations - we measure the uncontrollable system and then stuff around with the mating part to bring the assembly into spec.



Cheers

Greg Locock

 

[25362]

The Mazda-Ford story, as I was given to understand by reading the book on Taguchi's techniques, was based on the fact that Ford (Batavia, Ohio plant) contracted Mazda to make a certain portion of their front-wheel-drive automatic transmissions.

Both companies used the same blueprints. Mazda version,as warranty records showed, had a substantially lower claim rate than the Batavia version.

Ford investigated this phenomenon and found that Mazda's transmissions were made more consistently than their own. Mazda was using only 27% of the allowed tolerance range, while Batavia was using 70%.

Mazda was using a slightly more expensive and more complex grinder to finish valve outer diameters. At first glance, one may think their parts were more expensive, but knowing that the loss function was at work, the parts were actually cheaper, a fact substantiated by the lower warranty bill.

By using this information, the Batavia plant was able to improve its quality substantially and, in the first quarter of 1987, surpassed the Mazda level.

This anecdote is meant to show that a continuous reduction of variation within the allowed tolerance limits is a must i/o to provide a more desirable product to the customer. Which, in turn, results in a more competitive product with a lower associated loss to society.

 

[MintJulep]

This anecdote is meant to show that a continuous reduction of variation within the allowed tolerance limits is a must...

Yes and no. Here's the rub. Assigning tolerance limits is a design function. How much manufacturing variation can be tolerated on an individual part such that the assembly as a whole will still work properly. However, a truly well designed part takes into consideration the processes that will be made to produce it. That consideration includes knowledge of the of the variations expected to result from that process, and the distribution of thoese variations. Thus, for a properly designed part it should be expected that ALL parts produced will be within the allowed tolerances. Given that, a minor loss of control of process may be detected by a change in the shape of the distribution, even though not a single bad part has been produced.

Therefore, the only way to reduce the variation is to replace the existing process with a process with inherently lower variation. When this is done, the tolerances should be reevaluated and tightened in accordance with the new process.

The real goal needs to be continuous improvement of the process, not reduction of variation within allowed tolerance.

 

[HVAC68]

To put in simple words - redefine the tolerances. If the tolerance levels were so high that the outcome was a bad product, then, can we say that there was a problem in the initial definition of the tolerance level ???

HVAC68

 

[Theophilus]

Reduction in variance is purity.

To what degree is quality subjective? It's interesting to consider what Persig has in Zen and the Art of Motorcycle Maintenance (supposedly from Plato's Symposium, but I've not yet found it):

"And what is good, Phaedrus, and what is not good? Need we ask anyone to tell us these things?"


Jeff Mowry

http://www.industrialdesignhaus.com

Reality is no respecter of good intentions.

 

[btrueblood]

IR wrote awhile back:

"The second issue is that some of the details are conflicting. Is it even possible to make something perfectly safe? If so, can it even be built or sold at a reasonable price? Does the customer even care; since he probably is more interested in sharpness and durability in a knife than its safety? "

Geez, IR, if I start seeing safety warning labels on paring knives at the Target store, I'll know who to blame for the idea's origination.