In the late eighties, the first handheld mobile phone from Nokia called
CityMan 900 had a quarter length antenna stub. As an option, you could by a
longer antenna which I believe was a 1,5 WL coaxial dipole. It was fitted
with a SMA connector at the end. Anyhow, this antenna worked fine also at
the end of a 52 ohm coaxial cable. I used it on a cheap four meter carbon
fibre telescopic fishing rod stuck in the ground for hiking when base
stations were still rare in the early days of 900 MHz mobile telephony in
the Nordic countries. Signals strengths used to be much better with the
fishing rod antenna. Results were also better than using a six element 900
MHz yagi hand held 2 meters up.

Recently, a surveyor told me of the problem he and his collegues have in
getting a decent signal for the GPRS connection used to get RTK (real time
kinematics) correction signals. Field work is often done in difficult
terrain where cellular wave propagation is bad close to the ground. Getting
up to four meters using my old fishing rod design could help. Only, I have
to reinvent the coaxial antenna. Since the cable is ancored at the top of
the rod, it doesn't matter that the coax is pliable. A design that gives
additional field strength in the horisontal plane over a quarter wave would
be preferable (1,25 or1,5 lambda?)

BTW, Sony-Ericsson phones i.a. have the external antenna connector needed.

I'd use a low loss 52 ohm cable with a FME connector at one end. If we
assume that the TX band for european GSM is at 914 MHz, what should be the
length of the "unbraided" part of the cable - and should the insulator be
stripped off?

Regards

Jonas Lonnroth
ex SM5ACM ( in the sixties)