ZigBee signal through a metal pipe (as waveguide) experiënce?

Well I did have to try it myself, so I bought a eZ430-RF2480 demo-kit to try it out.

The first attempts were with a piece of pipe with a length of 1,5 meter, I fitted one of the battery board modules at the closed side of the pipe.
-It looks like the pipe is acting like a complete antenna when it’s free in the air, because I can receive the signal over the complete length of the pipe a few cm. away from it.
-But If I put the pipe on a table or on the floor the signal attenuates dramatically, so that could be a problem.

Can changing the position of the antenna have a positive effect on the attenuation?
Now the antenna is in the middle of the pipe, where the edge of the target board is facing towards the open end of the pipe. So the top side of the antenna is facing towards the sidewall of the pipe.
Should I turn the target board 90 degrees so the top of the antenna is facing towards the open end of the pipe?

Any idea's are welcome.

I've had a lot of "hands on" time with waveguide from .5GHz to 183GHz. It's great stuff the way it cuts off very hard on the low end of it's bandwidth at wavelength/2. The way you feed it (E or H field) does not change the cutoff frequency more than a  percent. Where you are well below cutoff you are not going to get anything through it. You have three choices: Get a bigger pipe, move up to the 5.8GHz band, send the signal on it, not in it.

1.) Bigger pipe: The lowest loss guide is 1/4 wave by 1/2 wave. Squish your pipe to approach this.
2.) 5.8GHz: This would be well above cutoff and there are standard products in this ISM band.
3.) On it, not in it: This approach is not ideal but can work. Drive the RF onto the pipe's external surface taking care to have a ground plane for the unit near to the drive point. You will find hot spots all along the lenght of the pipe.

Here are the numbers for circular waveguide:

     dia (I.D.)                      Cutoff TE11              Cutoff TM01

    31.7mm (1 1/4")       5.5 GHz                     7.2 GHz  (great single mode performance at 5.8GHz)

     40mm                        4.4  GHz                   5.7 GHz (OK, may require a little more tuning of probe)

     60mm                        2.9 GHz                   3.8 GHz ( 2.4 GHz below cutoff, multimode at 5.8GHz)

    75mm                        2.3 GHz                    3 GHz  (great single mode performance at 2.4GHz)

Ideally for single mode performance you want to be above TE11 and below TM01. So 31.7 mm at 5.8 or 75mm at 2.4.

40mm at 5.8 GHz  would be OK but may require a little additional playing with the probe to launch the desired mode.

Buried pipe: It does not matter what is on the outside provided it stays dry inside.
Looks like 5.8 GHz and  1-1/2" O.D. tubing is a fit.

Just a few thoughts here.

a -- if you can put multiple foreign objects into the pipe (antennas or even whole RF transceivers) then you must have pretty good access to the interior and exterior of the pipe.

Consider just running a traditional underground or plenum rated cable either wire or fiber optic alongside the pipe or inside the pipe and communicate over that means and not via free air RF.

b -- at 2.4 GHz real waveguides are ususally made from pretty conductive stuff such as copper, brass, or especially high conductivity type silver plated (on the inside!) layers with some other material like metal being the mechanical outer support.  You'll get a fair bit of loss due to surface resistivity of the pipe even if the pipe were the perfect dimensions to be a waveguide at your transmission frequency in the case where you're using something like wet / dirty / corroded steel / galvanized / aluminium or even copper pipe, though of those copper would work the best excepting questions of corrosion and lifetime issues specific to your application.

c -- You haven't mentioned (that I recall) one of the most important factors -- if it is a pipe, what's IN the pipe?  If it is just free air or a similar gas then your exploration can be based just upon the dimensional / propagation mode / propagation distance issues that you're already looking into.  If it is filled with something else such as water or whatever, the dielectric loss factor as well as the dielectric constant of the contained material will possibly have a profound effect upon the functionality of your system even if the pipe had the ideal dimensions to be a waveguide.  The material on the RF carrying surface of the pipe and inside the RF carrying volume of space (inside or outside the pipe as the case may be in your choice of method) must not be too lossy at 2.4GHz, and the dielectric constant must be accounted for when you're calculating the propagation.

d -- there is a "simple" way to get a 2.4GHz RF signal down a pipe of a diameter much less than quarter-wavelength... you put a transmission line inside the pipe such as a coaxial cable has a very narrow pipe/shield diameter, then some insulation, then an inner wire that carries the signal as well.  Corrosion and resistivity in the pipe would still affect you if you use the pipe walls as a return path, so maybe it would be better to just put a real piece of coax into the pipe, though if you do that, you might as well use ethernet cable or anything else rated for such burial as opposed to 2.4GHz RF.

e -- not that it is very practical but if you can stuff the pipe interior with the right higher permittivity / permeability contents you could reduce its dimensions and still have proper waveguide action since the wavelength of propagation in a material scales down as the permeability and/or permittivity increases relative to free air values, though of course the material cannot be too lossy. 

I'd just run a proper communications cable alongside or inside and call it a day as my first choice.

Well, first of all thank you for all the information.

The fact is that I can’t give you the details of the purpose of the wireless link, because it’s a confidential research I’m doing. But so far you did give me very useful information. At this moment I’m doing some practical test with various pipe sizes and the result are pretty good. At the end the pipe size has to be 40[mm] again, but in this test stage I also use bigger sizes. The pipes are empty (air) so that won’t be a problem, but it’s impossible to run cable or something else in it. (for some reason).

I think this will be my last message about this topic, simply because I can’t give anymore details so I just have to try it out myself.
But the information you guys gave was very useful to me, and probably other people can use it as well.