At 12:26 PM 3/1/2005, Jim Jarvis wrote:
>Telrex: I think there was a period when lighter, stronger
>aluminum was available, and was used in production. This
>resulted in different physical resonance conditions, which
>in turn resulted in premature failures.
>
>Which brings us back to Titanex. Antenna designers aren't
>necessarily mechanical gurus. As many have suggested, there
>appears to be a vibration induced fatigue causing the failures.
>
>I'm sure they're neither the first nor last to experience
>this problem. It'll be interesting to see how they handle
>Bruce's failure @ zf2nt.
The whole issue of fatigue failure and work hardening (they're different
aspects of similar phenomena..changes in the crystalline structure of the
metal) is better understood these days, AND, as Jim pointed out a few posts
ago, structural design has advanced significantly.
However, one shouldn't assume that a manufacturer of some product is
actually going to make use of that information and/or modeling
capability. For one thing, it's expensive. A "seat" for most of the
popular finite element programs that support analysis for things like
fatigue and vibration probably costs several tens of thousands of
dollars/year, just for the software. Then, you need to pay some engineer
who knows how to build the models and interpret the results. This is
substantially more complex than building NEC models (many, many more nodes,
it's nonlinear, etc.).
For doing a basic statics kind of analysis (will it break under static
load) might be a few thousand bucks, because you can use standard cookbook
formulas for tubes, etc. This is comparable to the engineering analysis
you'd get from a tower manufacturer for submitting to the local regulators.
However, when you get into a dynamics analysis, now you're talking a much
more complex model and simulation problem. Maybe something around $10K-50K
for a simple vertical mast with stepped diameters and guys, by the time
you've worked out all the boundary conditions, material properties, and
environments. There are some big firms that have standard models for
products that they analyze all the time (things like tall power poles with
wires hanging off them), so the time/cost to analyze for a new set of
dimensions, etc. isn't all that high, but the initial model definition
wasn't cheap.
And, if you want to throw time-varying material properties (like fatigue or
work hardening) into the mix, it gets remarkably involved (as in PhD
dissertation scale). Usually, you'd use some "safety factors" or "design
margin" or conservative design: for instance, on steel, you could insure
that you're below the fatigue failure stress, so you could then ignore
fatigue effects. For aluminum (which has no fatigue threshold) or other
exotic alloys (for which material properties (over life) might be less well
understood), you're probably going to fall back on the "make it wicked
strong" or the "build it and try it" approach.
You also have the ever present "workmanship" question if the structural
design is sensitive to it. Composites (e.g. fiberglass/epoxy or carbon
fiber/epoxy) are notorious for having widely varying properties in the
final product depending on how it happened to be laid up and
cured. Aluminum is tricky to weld, and the final weld strength might be
quite variable.
Just like most other ham products, I'll bet that not a whole lot of
mechanical engineering goes into something that is sold with small margins
and small volumes. You build a couple prototypes, do some static analysis,
maybe do a proof test to validate the analysis (does the deflection match
the model with a known static load, does it fail at the expected load), and
hope for the best, praying that there's not some horrible dynamic problem
that you'll have to cover under warranty. Even if you had a huge budget,
the dynamics analysis is quite sensitive to small changes in assembly and
material properties. A small change in stiffness won't change the failure
load very much, but could radically alter the resonant frequencies and
displacements. It's easy to cast a bell out of almost any metal that will
support its own weight. It's tough to cast one that sounds musical and
holds pitch.
Jim, W6RMK
_______________________________________________
See: http://www.mscomputer.com for "Self Supporting Towers", "Wireless Weather
Stations", and lot's more. Call Toll Free, 1-800-333-9041 with any questions
and ask for Sherman, W2FLA.
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|