LMH2100 Datasheet, PDF (22/32 Pages) National Semiconductor (TI) – 50 MHz to 4 GHz 40 dB Logarithmic Power Detector for CDMA and WCDMA

sured. This can be accomplished by using two LMH2100 RF
power detectors according to Figure 4. A directional coupler
is used to separate the forward and reflected power waves on
the transmission line between the PA and the antenna. One
secondary output of the coupler provides a signal proportional
to the forward power wave, the other secondary output pro-
vides a signal proportional to the reflected power wave. The
outputs of both RF detectors that measure these signals are
connected to a micro-controller or baseband that calculates
the VSWR from the detector output signals.
A sketch of this conceptual configuration is depicted in
Figure 5 .
FIGURE 4. VSWR Application
30014094
2.0 ACCURATE POWER MEASUREMENT
The power measurement accuracy achieved with a power
detector is not only determined by the accuracy of the detector
itself, but also by the way it is integrated into the application.
In many applications some form of calibration is employed to
improve the accuracy of the overall system beyond the intrin-
sic accuracy provided by the power detector. For example, for
LOG-detectors calibration can be used to eliminate part to
part spread of the LOG-slope and LOG-intercept from the
overall power measurement system, thereby improving its
power measurement accuracy.
This section shows how calibration techniques can be used
to improve the accuracy of a power measurement system be-
yond the intrinsic accuracy of the power detector itself. The
main focus of the section is on power measurement systems
using LOG-detectors, specifically the LMH2100, but the more
generic concepts can also be applied to other power detec-
tors. Other factors influencing the power measurement accu-
racy, such as the resolution of the ADC reading the detector
output signal will not be considered here since they are not
fundamentally due to the power detector.
2.1 Concept of Power Measurements
Power measurement systems generally consists of two clear-
ly distinguishable parts with different functions:
1. A power detector device, that generates a DC output
signal (voltage) in response to the power level of the (RF)
signal applied to its input.
2. An “estimator” that converts the measured detector
output signal into a (digital) numeric value representing
the power level of the signal at the detector input.
30014079
FIGURE 5. Generic Concept of a Power Measurement
System
The core of the estimator is usually implemented as a soft-
ware algorithm, receiving a digitized version of the detector
output voltage. Its transfer F_EST from detector output voltage
to a numerical output should be equal to the inverse of the
detector transfer F_DET from (RF) input power to DC output
voltage. If the power measurement system is ideal, i.e. if no
errors are introduced into the measurement result by the de-
tector or the estimator, the measured power P_EST - the output
of the estimator - and the actual input power P_IN should be
identical. In that case, the measurement error E, the differ-
ence between the two, should be identically zero:
From the expression above it follows that one would design
the F_EST transfer function to be the inverse of the F_DET transfer
function.
In practice the power measurement error will not be zero, due
to the following effects:
• The detector transfer function is subject to various kinds
of random errors that result in uncertainty in the detector
output voltage; the detector transfer function is not exactly
known.
• The detector transfer function might be too complicated to
be implemented in a practical estimator.
The function of the estimator is then to estimate the input
power P_IN, i.e. to produce an output P_EST such that the power
measurement error is - on average - minimized, based on the
following information:
1. Measurement of the not completely accurate detector
output voltage V_OUT
2. Knowledge about the detector transfer function F_DET, for
example the shape of the transfer function, the types of
errors present (part-to-part spread, temperature drift) etc.
Obviously the total measurement accuracy can be optimized
by minimizing the uncertainty in the detector output signal (i.e.
select an accurate power detector), and by incorporating as
much accurate information about the detector transfer func-
tion into the estimator as possible.
The knowledge about the detector transfer function is con-
densed into a mathematical model for the detector transfer
function, consisting of:
• A formula for the detector transfer function.
www.national.com
22