Background: Raymond P. Lambeck (retired) was director of engineering for pumps, motors, and transmissions for Sperry Vickers, a major manufacturer of hydraulic equipment based in Troy, MI. This article appeared as a series in the Jan. 6, 20, and Feb.6 Machne Design magazine, a Penton Media Inc. publication and is sent to Eng-Tips.com with the permission of Machine Design.

--------------- as it appeared in Eng-Tips -------------------------

…"I cut this article out of Machine Design early in my career and it has been a great source of perspective and wisdom. Machine Design graciously gave me permission to send it to Eng-Tips.

It is difficult to make a product foolproof since fools are so ingenious, but any product incapable of being produced and operated by "real" people should not be turned loose in the real world.

We have a right to expect that our designs will be tooled and processed in a way that will meet the basic drawing requirements; also, that our customers will observe our published ratings and read our application manual. We have that right, but we should know better. Therefore, the slightest deviation from the straight and narrow should not result in total disaster. A little loss of performance? Perhaps, but not for every design transgression. Thus, Law #1 implies that we will plan to probe the design boundaries during the development program to establish our true design margins, and that we will extend the margins when feasible.

This law agrees with Murphy, who says that if you leave it alone it will only get worse. It also follows a dictum from Indiana basketball coach Bobby Knight: "Either be a good leader or a good follower, and, if you can't be either, for heaven's sake get out of the way." If it works in basketball, it should work for engineering.

Besides, the laws of probability tell us that three logical or seemingly appropriate decisions will yield an 87.5% expectation of success - even if each decision has only a 50% probability of being the best that could be made. All of which should remind us that our real enemy is indecision.

The odds on success per decision per decision can be increased materially if each action has a sound engineering reason. The converse to this is that we should steadfastly refuse to accept a solution founded solely on unexplained phenomena. When witchcraft and black magic are used, the whole solution can disappear the first time someone sticks a pin in our doll. In any case, don't just keep running the same thing and hoping for the best.

Priority 1 - What will produce the maximum gain for minimum cost in the shortest time?

Priority 2 - What is the most positive absolute solution with no question of adequacy?

While directing initial attention to Priority 1, it is well to start work toward Priority 2 just in case. If we miss them on the roundabouts, we'll catch them on the swings. While time and cost cannot be ignored in the second case, they are subordinate to absolute certainty. Above all, the product must work. No shortcoming will ever be forgiven.

Until the problem is solved, always have an alternative in mind and be working on it. This follows the theory that when 2 + 2 refuse to make 4, try 3.

If there is any choice, overpower the problem rather than trying to finesse it. In poker, one Smith and Wesson beats four aces. In engineering, one positive fix beats four probables.

As a corollary of this law, no change should be considered truly significant unless it makes an improvement (in stress, bearing pressure, deflection, etc.) of at least 25% - and 50% is better, considering the scatter band of material properties. This faithfully follows the theory that if a little bit goes a long way, try a lot.

Confronted with a failed part, any metallurgist worth his salt can always find something wrong. It is tempting to accept this finding as the full answer, but minor changes in heat-treating or metallurgical quality rarely affect physical properties be as much as 25%, so we often need to look further. Every engineer knows that even a very good design can suffer a random failure. He sometimes forgets that even a bad design can sometimes enjoy a random success, but the fundamental weakness remains.

(Note: This runs contrary to one of Murphy's Laws which says: "Never replicate a successful experiment.")

Running a test first and rationalizing the results later does not constitute a proof; it only illustrates a potential theory. The theory is not proven until you can turn the phenomenon on and off at will. Going back to Law #2, do not settle for one random success. It may never happen again.

Once the drawings are changed, once the parts are ordered, once the test plan is written, once the failed parts are sent to metallurgy, it is tempting to sit back and relax knowing that someone else momentarily has the ball. Don't do it! You are still accountable. Check schedules. Find shortcuts. Clear roadblocks. Buy coffee. Be a squeaky wheel.

And don't make excuses. Your friends don't need them and your enemies won't believe you. Apply all that energy to finding a way.

There is always a way.

If there has been an unexpected problem or the product has developed leprosy, you might as well admit it. The customer (or management) will discover it eventually anyhow. It will be nicer if they hear it from you.

The only antidote is to have a rational explanation available and a viable plan of attack. This may require some quick thinking, but that is what you are paid for.

Never start a job unless you intend to finish it.

This follows on Lambeck's negative postulate, which says: "Just because something has a beginning, it doesn't mean it has an end."

Many engineers, being both born perfectionists and avid experimenters, will tinker with a product far beyond the Law of Diminishing Returns. Some, in fact, will tinker with it until it no longer works and they have forgotten the road back.

A restatement of Law #8 might read: "Good enough is perfect."

An adequate product delivered to the field beats two nearly perfect products hung-up in the lab.

A memo is good for only two purposes: Disseminating useful information or recording an agreement. When it asks a question (Have you stopped beating your wife?) or voices a complaint (Why haven't you stopped beating your wife?), all it does is set off a negative reaction...followed by a return memo. There is nothing as futile and unproductive as a memorandum war.

Do your own thinking. Be your own person. Rely on technical judgement. And speak up! 'Nuff said.