3" is rather too small to work as a waveguide on 2m.
If the sniffer antenna is something like a rubber duckie then it has a
null off the end. If that is the part that is pointing towards the transmitter
then it will tend to null the signal at about the same time as the tubing
points to it, making it difficult to get a reliable bearing.
You don't really get effective slot / horn / waveguide antennas until
the dimensions are on the order of a half to one wavelength. (For
example, a waveguide would be about 48 x 24 inches.)
That isn't to say you can't try it, of course, but I doubt that you will see
the effect that you expect. You could try building a small dipole probe
and sliding it near the mouth of the tube to see if you can detect a
pattern to the antenna. I suspect that fringing around the end of the
tube will couple signals picked up on the outside even when the tube
is not pointed at the transmitter, but you might get it to work.
Such antennas are often built for 2.4 GHz reception, but the optimum
diameter for that frequency is around 4 to 5 inches IIRC, making a 2m
version rather large for hand-held use.
Where you can use a conductive tube, however, is as a "waveguide
below cutoff attenuator". I use a cardboard tube covered with aluminum
foil, though aluminum stovepipe or other material will work if it is
large enough. I clip a string to the strap on my HT and lower the whole
thing down inside the tube. At some point the signal goes away - I pull
it back up a bit and use the combination for body shielding. It is quite
obvious when the receiver is shielded by your body when the height
of the HT in the tube is adjusted for optimum sensitivity. This has the
advantage the the HT itself is shielded, since the limit of how strong of
a signal you can hunt is when the HT picks up enough signal through
the case to pin the meter even without an antenna connected. Once
you get to that point, no amount of attenuation in the coax or extra
feedline shielding (such as hardline) will do a bit of good.