Machine Design

Ins and Outs Of Outsourcing

On numerous occasions throughout my career, I have fought tooth and nail to prevent jobs leaving this country ("Outsourcing, like a brutal beating, is good for you," April 1).


I've argued with union members that they must be more productive to compete against lower-wage overseas workers, and I've argued with managers and CEOs about the hidden costs of outsourcing production. I have struggled with suppliers half a world away who don't speak English and can't or won't change designs, schedules, or shipping plans to work with my company's needs. I have ground my teeth in frustration over delays in making any supply-chain changes when that supply chain is 12,000 miles long.

What has been missing from your discussion so far, however, is any rational alternative to the current system. Yes, outsourcing stinks. Yes, we should take our corporate leaders to task for their shortsightedness. But please, please, don't propose that our government should step in and mandate what private capital owners can do with their capital.

As leading economist Thomas Sowell has pointed out, there are only two alternatives: Either markets made of competing, profit-motivated individuals will decide how resources are allocated, or dictators will. We don't need to contemplate the latter. It's been tried, and it is nothing short of disastrous. They may be small groups of elites in a government bureaucracy, or they may be political strongmen, but they are dictators all the same.

We may hate outsourcing, but the cure is worse than the disease if it empowers government to restrict freedom.



One reason engineers have forgotten about the impact of dimensioning and tolerancing is outsourcing. As outsourcing has become commonplace over the last 10 to 15 years, the new generation of engineers have not gotten much hands-on experience with manufacturing and assembly. Without this experience, it is difficult to know the effects of tight tolerancing. In previous times, parts were made in-house and in-house machinists asked engineers if all the tight tolerances were really needed.

Today, there is no incentive for a subcontracting machine shop to make the same phone call. They simply charge more for building to tighter tolerances, as requested on the drawings. In my experience, subcontracting shops only call when a part is impossible to make or the drawing contains conflicting information.

Older, more-experienced engineers could check drawings, but this would essentially mean design work is being done twice. As more engineers retire and outsourcing becomes more embedded, the problem will get worse.

Outsourcing may save some money on the short term but end up costing us more in the long run.



Giving your own test is a good way to evaluate job applicants ("Voodoo in the personnel department," Sept 2). I don't give psychological tests, just a simple test of the actual skills required for the job in question. I include problems ranging from trivially simple to fairly obscure and difficult. I limit the test to what I think almost anyone should be able to finish in under an hour. And it isn't how well applicants score on the test that counts. One of my best hires pretty much flunked it. The important results stem from how they handle the test and the ensuing discussion of their answers. If they can't do one of the problems, I try to see how easily they can learn how it should be done. I think it works because it gets beyond the artificial discussion of the applicant's resume and generates some spontaneous interaction similar to real work. I get some idea of what the applicant will be like on the job. And come on, Ron. Just when your column was getting really interesting, you didn't want to tell us the rest of your test results.




Principle of operation

The float-switch industry, indeed, the entire reedswitch industry, has not properly told designers how to get the longest life from reed switches ("Switch Tips," p. 50, Sept. 2). Years ago, I bought a reed switch that did not last. The design was simple: when a tank was full, the float switch engaged a relay which shut off a pump. The relay presented a low load, below the switch's rating. But switch after switch burned out after several months. The manufacturer offered no solution, only telling me the load was too high. But I had chosen their best switch and the load was lower then the switch rating. I was forced to add an intrinsically safe control relay, at a much higher cost.

I later learned that the relay produces a much higher initial load as the magnetic field is filled, then the load drops off. This is death on reed switches. It took me 20 years to trip over the answer: put a 100- resistor on the wire feeding the relay coil. It prevents initial loads from exceeding the limits of the reed switch. So simple, so elegant, so cheap. Why don't those selling such switches include a resistor or at least tell customers of this simple solution when directly operating a relay?


Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.