Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations IDS on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Quality impeding Reliability. What are your stories? 2

Status
Not open for further replies.

DHambley

Electrical
Dec 7, 2006
246
Without tainting the thread with my own views, let's look at a few responces first. Do you have a story about quality getting in the way of reliability in aerospace products?
 
Replies continue below

Recommended for you

You better fully define what you mean by quality.

If you mean 'better made components' then, no I don't think I've experienced them causing a problem (I'm excluding possible situations where there was a design problem that poor quality may have been hiding).

If you mean 'QA systems' such as ISO-9001 then maybe. I've heard the view espoused that because of the effort required to 'certify' design changes, or the like, that improvements sometimes aren't incorporated.

Posting guidelines faq731-376 (probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484
 
hgldr,

Many, many years ago, I was involved in a project for a European customer who insisted that we set up a QA process. The end result was that all the people who knew how to do the work wrote quality procedures, and we brought in contract people to do the actual work. Quality suffered.

If you are working with an organization controlled by rigid, procedure driven people, your QA process will be just another set of procedures to follow, and there will be no improvement in quality.

Critter.gif
JHG
 
On the other hand what about the hundred years of experience suggesting that better quality leads to better reliability? Or do you buy Yugos in preference to Hondas?




Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376
 
GregLocock,

Better quality leads to better reliability. A lot of QA processes (not all) lead to reduced quality, with all the attendant consequences.

I read the OP's use of "quality" as meaning QA. A strict literal intepretation of his question does not make sense.

During WWII, the Germans had problems controlling their Fritz_X glider bombs in combat. The controls worked fine on the ground. It was eventually found that a French slave worker had taken apart some coax cables, snipped the conductors inside, then reassembled everything. The resulting electronics worked on the ground, but did not work in the air with the airplane vibrating. One could argue that this was a quality program with the objective of reduced reliability.

I cannot find any mention of this in Wikipedia. I got it from an aircraft history at home. I will have to look it up.

Critter.gif
JHG
 
Brazing job.

100,000 parts.

Didn’t get full flow on maybe 1,300. Still functional but appearance not suitable.

Reran them and all looked good.

What we missed was that the braze alloy had Zinc in it. Every time it was heated it lost Zinc and the remelt temperature was maybe 50F higher.

Customer went to use the parts in a machine set for the proper temperature which turned out to be too low for the reworked parts.


Thomas J. Walz
Carbide Processors, Inc.

Good engineering starts with a Grainger Catalog.
 
I wanted comments first before explaining the question more clearly. Quality in my industry is not the same as the word "quality" to most english speaking people. In street language, quality is synonymous with goodness, runs-forever, reliable, lasts-a-long-time, etc, etc. In Industry however, Quality basically means that every unit is built exactly the same. The epoxies are mixed exactly the same, at a controlled temperature, the oxide coatings are all controlled to within a few millionths of an inch, all the same.

If the original design and analysis is wrong, if a shaft is running beyond its torque rating, if a transistor is running too hot, then all of the units will be built exactly the same; with a transistor which is going to fail and a shaft which is going to break.

KENAT's comment reminded me of what I see very often, "...because of the effort required to 'certify' design changes, or the like, that improvements sometimes aren't incorporated."
Sadly in aerospace, certification is so incredibly expensive, that faulty products are often sold even though the manufacturer knows that a simple fix will improve it. The FAA and ISO don't allow simple fixes though. I bet those O-rings on the Challenger had 20lb of paperwork with the word "Quality" stamped on each sheet.
 
So what is your problem or question?

"The FAA and ISO don't allow simple fixes though" Rubbish. The FAA doesn't allow "fixes" based on wishful thinking, arm-waving and management BS. While the cert process isn't perfect, its a lot better than not having any process.

 
So, in your situation, how is "Quality impeding Reliability"? I can see how it would affect production, but if product of consistant "quality" is being produced I don't understand how it affects "reliability". Or are you referring to "Poor Quality" as opposed to customer requirements?

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]
 
Personally I use Dr. Deming’s definition that quality is exactly what the customer says it is.

We differentiate quality, reliability, consistency and many forms of each as distinct attributes of the same process. Remember the story of the blind men and the elephant. One felt it and determined it was a wall, another felt a tree trunk, a third felt a rope and a fourth felt a snake while the fifth felt a huge leaf.

Your question seems like playing with words. Much as in high school we devised elaborate math formulas to prove that 1+1 =0 and other nonsense.

Personally I am really impressed by the aircraft certification process. Many tens of thousands of parts all interacting and the planes rarely fall out of the sky. It is still a very safe method of transportation.


Thomas J. Walz
Carbide Processors, Inc.

Good engineering starts with a Grainger Catalog.
 
Seems like the OP hasn't used 'reliability' in his Demming definition of 'quality'. Anyway looks like a classic thread where anecdote will be presented as data.





Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376
 
SWComposites, your comment that it's "Rubbish" that the FAA doesn't allow simple fixes shows me that you really have not seen how expensive re-certification is. Your bias that simple fixes are defined as "arm waving" may point to an actual problem of what goes on in your particular office.

In the orignal question of this thread I was looking for actual examples from actual projects, and there ARE many stories.



 
12 posts already and only one example. Engineers I know who actually work in aerospace have many examples. If you are in that industry, I know you've seen similar stories. Here's a few from three different companies:

Inductor core cracking on an aircraft battery charger due to excessive core loss. Engineering found a simple alteration of the inductor was necessary. QA sais can't do. Re-certification would cost over $20k though. Fix was never implemented.

Diodes within a rocket thruster subject to voltages beyond their rating. Fortunately, this would only occur during a worse case condition not seen with typical conditions yet. QA would require expensive recertification of the entire unit if the fix were installed. Simple change to a different diode never took place.

Unstable oscillations from a voltage supply to software unit in a satellite. Engineering found a simple chip capacitor added solves it. Change was tested in lab over temperature extremes and CAD analysis. QA sais no. Change would require vibration tests, certified thermal chamber tests, re-writing documents like build procedures, certification compliance, etc...too expensive and time-consuming. Units were shipped with unstable supply lines. They were all built the same. QA was satisfied.

If you are an engineer in aerospace, I know you've seen similar stories.
 
Phil Crosby is considered "Old hat " by some QA gurus these days, but a comment he made in the 70's still resonates among some QA people.

"Tip:
Be very careful about where you apply zero defects. If what you're doing contributes towards a mission critical or complex goal, you'd better adopt a zero defects approach, or things could quickly unravel.


However, if you fanatically follow a zero defects approach in areas which don't need it, you'll most likely be wasting resources. One of the most important of these resources is time, and this is where people are accused of time-destroying "perfectionism.""
B.E.

 
Dear hgldr,

Originally it looked like you were after a philosophical discussion about the relationship between quality and reliability. Now it looks like you are talking about quality specifications that are not perfect. To me those are two entirely different topics.


Thomas J. Walz
Carbide Processors, Inc.

Good engineering starts with a Grainger Catalog.
 
Seems part of the problem is your QA departments operating outside their field of responsibility/expertise.

In my time in Aerospace/Defense it was up to Engineering and the Customer Desk officer to decide the types of decisions you're talking about.

It was Engineering's Names that went on the Certificates of Design, not QA.

It was up to QA to ensure units built actually met the data pack referenced by the Certificate Of Design.

Posting guidelines faq731-376 (probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484
 
hgldr,

I have worked with people who have aggressively demanded consistency, in my opinion, frustrating efforts to improve processes.

You need to distinguish between QA procedures, and actually achieving quality. QA is just like any other job. It is futile if you do it badly.

Critter.gif
JHG
 
hgldr said:
...

Inductor core cracking ...

Diodes within a rocket thruster ...

Unstable oscillations from a voltage supply ...

Are these conflicts with your QA department, or are they conflicts with management? QA requires you to follow procedures and submit documentation. It is management that decides that it is not worth the effort.


Critter.gif
JHG
 
This thread should be retitled to "Does management greed impede reliability?" Sadly the answer is yes in too many cases.

As KENAT stated, QA should have nothing to do with certification; that is an engineering responsibility. Engineering should discuss the certification requirements with the customer/FAA and determine the required data/testing/etc.

Sounds like in the cases quoted, insufficient testing was originally done, and now management doesn't want to pay to fix the problem. Hard to have sympathy.

Perhaps the real question should be "why wasn't the customer/FAA informed about the defective conditions?"


 
Status
Not open for further replies.

Part and Inventory Search

Sponsor