Selective sampled peak detector and method

Details Selective sampled peak detector and method Selective sampled peak detector and method

1999-05-28

2001-05-15

Tran, Toan (Department: 2816)

Miscellaneous active electrical nonlinear devices, circuits, and

Specific signal discriminating without subsequent control

By amplitude

C327S058000

Reexamination Certificate

active

06232802

ABSTRACT:

FIELD OF THE INVENTION
The present invention relates generally to the field of peak detection, and more particularly to a peak detector that samples selected portions of an input data signal.
BACKGROUND OF THE INVENTION
Peak detection techniques are required in various applications that require the information of the signal strength or power level. For example, in physical layer designs, particularly in designs with 100 Base TX receivers, peak detection plays an important role in the process of recovering data that has propagated across a medium such as a CAT-5 cable. In some architectures, an average peak voltage level of an input signal is used to determine the length of the cable, and is also used to define the level of equalization required to compensate for amplitude loss and phase distortion incurred by the signal after transmitting along the cable line. The design of peak detectors with good noise immunity becomes more important for systems with smaller incoming signal levels (due to, for example, long cable lengths and/or reduced supply voltage ranges), and for systems with a higher level of chip integration (analog and digital).
FIG. 1A
illustrates a block diagram of a conventional peak detector
100
which includes a comparator
105
, charge pump
110
, and a capacitor
115
. The output of the comparator
105
is coupled to charge pump
110
including a controlled pull-up current source
112
for generating current I
1
, a controlled pull-down current source
114
for generating current I
2
. The output of the charge pump
110
is coupled to the capacitor
115
as well as the negative input terminal “−” of the comparator
105
, thus forming a unity gain feedback configuration.
The charge pump
110
and the capacitor
115
generate an average peak voltage signal V
0
across capacitor
115
. To detect the peak voltage signal for positive pulses in a data signal, the pull-up current source
112
is much greater than the pull-down current source
114
(i.e., I
1
>>I
2
). On the other hand, for negative pulses in a data signal, the pull-down current source
114
is much greater than the pull-up current source
112
(i.e., I
1
<<I
2
). As depicted by
FIG. 1B
, the principle of this conventional peak detector
100
is as follows. The positive pulses peak detection is used as an example. After the average peak level V
0
is achieved, the total area of data signal
150
which is above the level V
0
is denoted as A
1
. The total area of data signal
150
which is below the level V
0
is denoted as A
2
. The average peak level V
0
is derived to include A
1
( x )=A
2
, wherein x=I
1
/I
2
. The ratio of the pump up current I
1
over the pump down current I
2
(or x) is much great than one (1). Similarly, for a negative pulses peak detection, the following is satisfied: A
1
=A
2
( x ), wherein x=I
2
/I
1
. The ratio of the pump down current I
2
over the pump up current I
1
(or x) is much great than one (1).
Conventional peak detectors suffer from various problems and drawbacks such as, for example, data dependency, high sensitivity to noise, and level fluctuation, as discussed below. The data dependent nature of conventional peak detectors is shown in the example of FIG.
1
B. Assume that an input data signal
150
is received at the positive input terminal “+” of the comparator
105
. Since the data input signal
150
has a dense pulse pattern (i.e., logic high occurs more frequently than logic low), the level of the average peak voltage signal V
0
will be close to the peak level
155
of the input data signal
150
pulses. In contrast, for a data input signal
160
with a sparse pulse pattern (i.e., logic low occurs more frequently than logic high), the level of the average peak voltage signal V
0
is significantly less than the peak level
165
of the input data signal
160
pulses. The average peak voltage signal V
0
tends to drift downward toward the logic low level due to the sparse pulse pattern, and, as a result, may not provide a correct measurement of the peak level
165
of the input data signal
160
. To reduce the data-dependent nature of conventional peak detectors, the current ratio provided by current sources
112
and
114
(
FIG. 1A
) must be adjusted. For example, to detect the peak of positive pulses in a data signal, the ratio of the pull-up current source
112
over its pull-down current source
114
is set at a much higher value (i.e., x=I
1
/I
2
>>1). Thus, even if a sparse pulse pattern signal occurs, the average peak voltage signal V
0
will quickly pull-up to the pulse peak in the data signal.
However, the much higher ratio between the current sources
112
and
114
causes a conventional peak detector to be more sensitive to noise induced at the peak detector input. For example, in
FIG. 1C
noise
170
may occur at a pulse peak of an input data signal
175
. The average peak voltage signal V
0
will quickly rise to at least the noise
170
level. Since the charge rate of current source
112
is much higher than the discharge rate of current source
114
, the average peak voltage signal V
0
requires significant time before decreasing to the correct pulse peak level
180
. This characteristic makes the average peak voltage signal V
0
very sensitive to the induced noise.
Conventional peak detectors also suffer from a level fluctuation problem that occurs when the peak detector tries to overcome a change in the pulse peak level, as described below with reference to
FIGS. 1D and 1E
. Conventional peak detectors typically use a drooping mechanism for tracking pulses as the pulses gradually decrease in amplitude. In the case of detecting the peak of a positive pulse, the droop rate is controlled by pull-down current
114
and the capacitor
115
. An average peak voltage signal V
0
generated by a conventional peak detector may “droop” so that the decreasing peak levels
180
of an input data signal
185
are properly tracked.
FIG. 1D
illustrates how this drooping condition permits the tracking of the decreasing peak amplitude. However,
FIG. 1E
illustrates the drawback caused by the drooping condition of the average peak voltage signal V
0
. The average peak voltage signal V
0
will fluctuate if the peak amplitude of the pulse
185
does not decrease its level in a subsequent pulse. In particular, the average peak voltage signal V
0
will droop between pulse occurrences and then suddenly increase by an amount
190
to the peak level
180
during a subsequent pulse occurrence. This condition results in an undesired signal fluctuation.
Therefore, there is a need for an improved peak detector that overcomes the problems of data dependency, high sensitivity to noise, and undesired level fluctuation.
SUMMARY OF THE INVENTION
The apparatus and method of the present invention operates to track a peak level of an input signal. The apparatus includes a comparator for comparing the peak level of the input signal with a reference peak voltage signal. A sample and block circuit is coupled to the output of the comparator and is capable of sampling a portion of the input signal. The sampled portion of the input signal is defined by a smart window (timing window) which is received by the sample and block circuit. The sample and block circuit controls a charge pump that adjusts the level of the reference peak voltage signal.
The use of a “smart window” in accordance with the present invention provides a peak detector that has a high immunity to noise. In addition, the use of a “smart widow” reduces the level of fluctuation in the reference peak voltage signal generated by the peak detector. Furthermore, the present invention provides a peak detector with an improved peak detection performance that is not negatively affected by the pulse pattern of the input signal.
The apparatus and method of the present invention is useful in many applications that require signal peak detection. Thus, the present invention can improve the performance of transceivers, sensors, cellular phones transmit output lev

Affiliated with

Also associated with

Rating

Selective sampled peak detector and method does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Selective sampled peak detector and method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Selective sampled peak detector and method will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2504611