[TowerTalk] Shorty Forty Hose Clamp thread

Jim Lux jimlux at earthlink.net
Thu Aug 13 10:11:02 EDT 2015


On 8/11/15 8:50 AM, Doug Renwick wrote:
> My problem with the torque method is so you torque the clamp to specs and
> yet you can still turn the smaller element inside the larger.  So torquing
> in this application in meaningless.  You tighten the clamp until the smaller
> tubing does not turn, let it cycle overnight, check and retighten if
> necessary.  My elements don't fall apart and my temperature swings are from
> minus 40C to plus 35C.  Real world or simple method.
>


I think that's more a design problem than a clamp problem.
That is, the slits in the outer tube aren't right or the material isn't 
right, or something, so that the outer tube can't deform sufficiently 
that the clamp at its rated clamping force won't hold things together.

This kind of joint is one that's really not trivial to analyze, either. 
I suspect that most of the designs are empirical: someone builds one, 
tries it, and iterates until it works.  The problem with this sort of 
design process is that you don't always know what the sensitivity to 
tolerances are.

I've designed and built some tube clamping things and they've not always 
turned out well. The canonical design was a block with a hole bored to 
fit my tube, and a slit, and a bolt to close the slit (reducing the size 
of the hole slightly). I seem to always need to either to shim the hole 
or ream it out: it turns out that this style of clamp is sensitive to 
diameter variations of 0.001" kind of scale, because the bulk of the 
clamp (imagine a 1" x 2" cross section aluminum bar with a 3/8" hole in 
it, the short way) makes it very stiff: you can't put enough force with 
a bolt to deform the metal without stripping the threads out.  That tube 
may *say* 0.375" OD, but in reality it could be anywhere from 0.365 to 
0.385, and that's a huge difference.  (yup, very poor clamp design.. I 
know a lot better now)




Some random googling "aluminum tube dimensional tolerances" turns up a 
page at "tubeweb.com"...

For small tubing, it looks like the diameter tolerance is 0.010 inches 
for the mean, and 0.020 at any point (except for some 5XXX alloys)

For wall thickness small tubes (<1.25" OD) and walls under 0.047" the 
thickness can deviate 0.006" (mean) or +/- 10% of mean thickness with a 
max of 0.060 and a min of 0.010.

So when you're nesting two tubes, they could be a "snug fit" or a "loose 
fit" or even a "very sloppy fit" and still be entirely within spec.

if the original designer bought their tubing from a particular place, 
with particular sizing, they may have found that tightening the clamp to 
the rated torque (not necessarily a good assumption) deformed the outer 
tube the right amount to get a good interference fit with the smaller 
tube.

But a later lot of the antennas, or someone duplicating the design, and 
copying the slit dimensions exactly, might have got tubing that is 
slightly thinner wall, and maybe the inner tube OD is at the minimum OD.
Or even, the tubing is uneven, and the end of the tube happens to be at 
a place where the OD is minimum.

The metal business is price sensitive and manufacturing quality has 
become much better (I would imagine that "run of the mill" tubing you 
buy today is better than "run of the mill" tubing you would have bought 
50 years ago, in terms of dimensional consistency).  A manufacturer 
might decide to run their walls slightly thinner than nominal or 
diameter slightly undersize, but within spec, because they've got good 
enough process controls to guarantee they won't bust the inspection 
spec, and it will save metal.  From saving fractions of an ounce on 
millions of tons of tubing is how you stay in business.










More information about the TowerTalk mailing list