[Top] [All Lists]

[TowerTalk] how vertical is vertical

To: <>
Subject: [TowerTalk] how vertical is vertical
From: (Jim Lux)
Date: Tue Mar 4 23:21:08 2003
----- Original Message -----
From: <>
To: <>; <>
Cc: <>
Sent: Tuesday, March 04, 2003 6:43 PM
Subject: Re: [TowerTalk] how vertical is vertical

> In a message dated 3/4/2003 5:25:06 PM US Mountain Standard Time,
> writes:
> > I'd be interested to know the basis of the "verticality" requirement...
> > part in 100 (about 1/2 degree) wouldn't appreciably change the loads at
> >
> Hi Jim;
> My EIA-222-D (yes, its old but its the most current I have here and this
> section probably hasnt changed) says;
> Section 6.1 Standard
> Section - Plumb - For guyed structures, the maximum deviation from
> the true vertical shall be one part in 400. For self-supporting
> the maximum deviation from true vertical shall be one part in 250.

Indeed, it looks like a standard industry spec.  But I was wondering why
they selected that?  And, is it aimed at 50-150 foot towers or 1000 ft
behemoths..   For all we know, the EIA standard said that because it "seemed
like a good number" and "it was in some other spec we had that seemed to
work". Or, they may have had some independent analysis to back it up.

Some background.... In my job at Jet Propulsion Lab, I deal with all manner
of specifications, and, in the context of the ever pressing schedule,
sometimes specs get in because they just got copied from somewhere else,
without any backup for why that spec existed in the first place.  The scary
thing is that the original spec (7 or 8 generations back) may have been
written as a "let's see what the prospective bidders will say" kind of
requirement; or, perhaps, in that previous project, "cost is no object"; or,
maybe they wound up writing a waiver against an impossible requirement (but
never went back and modified the spec).

Therefore, when I see a specification that is cast in nice round numbers (1
in 250, 1 in 400), and that appears to be reasonably achievable in a
commercial environment (i.e. cranes, trucks, lots of people being paid to do
their job), I wonder if the spec was set as "well, it shouldn't cause any
real cost impact to meet, and it would cost us a lot to figure out a real
number that is looser, but still safe."  Those kinds of commercial specs may
not be appropriate for a different kind of operation (i.e. less labor, less
capital intensive) where a perfectly safe(!) installation might result, even
though it doesn't meet the commercial spec. It also properly embodies a just
conservatism.  After voting to approve that spec, you could go home and not
lose sleep worrying about whether a tower meeting the spec would fall over
from not being plumb.

Let's be realistic.. relatively few amateurs do all aspects of their station
to full commercial standard (Hmm.. surely you all comply with all aspects of
the 2002 NEC and the NESC, and the IEEE grounding standard, and you use CCS
rated transformers in your power supplies, you wouldn't think of shooting a
AWG 20 wire over a tree with a slingshot, you use only fully type accepted
hardware,  etc.etc.etc.)..  Commercial and industry standards exist for
reasons other than pure safety.  Commercial broadcasters, for example, need
substantially better reliability than the average amateur. Industrial
installations often have to account for people who may not have a technical
understanding, and may not appreciate, certain fundamental limitations.

So, to come back to the original question.... Is a tower that is 1 foot out
of plumb in a 150 foot column unsafe?  Sure, it doesn't meet EIA-222, so
it's suspect at first glance, but, is it unsafe?  I don't know that you
could actually see it be off-vertical without an outside reference (a
transit or plumb line), so it probably meets the "pretty enough" test.
Again, pretty doesn't imply safe, nor does safe imply pretty.

So, what's the origin of the original 1 in 400 spec?   (I don't actually
expect to find out.. it's one of those things that is probably lost in the
mists of time, or is loosely derived on some rule of thumb for eccentric
column loading/buckling)

<Prev in Thread] Current Thread [Next in Thread>