Hi lee
The 1dB gain, linear polar and the 25MHz
(about 2.5%) bandwidth are all achievable with a
variety of designs, such as normal mode helical
quarter wave etc.
It was just you used the term
"efficiency" and find that 38% is unacceptable.
I wondered what your reference model is for 100%
so I can see what basis you are using for acceptable
performance.
When you say you are using a chamber,
that does not describe the actual method. Are you
plotting the E and H plane far-field patterns and
integrating them in software as part of a suite,
like Scientific Atlanta or similar, then accepting
the given answer? Or are you comparing the radiated
power with the input power to the antenna and computing
the answer as a percentage?
As a rule of thumb,
the ground-plane dimensions (critical for compact
antennas) should be about lambda/3 by lambda/3 at
least in order to achieve high efficiency. Anything
smaller will greatly reduce the efficiency.
Have you checked the reflection coefficient and/or
S parameters?