Pulse or digital communications – Equalizers – Automatic
Reexamination Certificate
1999-02-03
2001-11-06
Le, Amanda T. (Department: 2734)
Pulse or digital communications
Equalizers
Automatic
C375S355000
Reexamination Certificate
active
06314133
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to an automatic equalizer and a method for generating a sampling clock used therein, and a recording medium in which a control program for controlling the automatic equalizer with a computer is stored, particularly to an automatic equalizer for automatically equalizing a signal which has been distorted due to an intersymbolic interference, and a method for generating a sampling clock used therein, and a recording medium in which a control program for controlling the automatic equalizer with a computer is stored.
In a method of selecting a sample timing in an automatic equalizer, the timing is selected by which the impulse response having the maximum peak value is obtained. Hereinafter, a case of applying this method to an automatic equalizer described in Japanese Unexamined Patent Publication (JP-A) No. Hei 10-65580 by the present applicant will be described as an earlier technology with reference to FIG.
1
.
The automatic equalizer shown in
FIG. 1
comprises a sampler
101
for sampling a received signal Sr with a sampling clock SCLK to output a sampled received-signal Ssr, a subtracter
102
having inputs of the sampled received-signal Ssr and N estimated received-signals Ser, where N represents a positive integer, for producing N estimated error signals Seerr, and a detector
109
having an input of the N estimated error signals Seerr for producing a part of the maximum likelihood transmission signal sequence to the outside as a detected output-signal Sd.
The automatic equalizer further comprises an impulse-response calculation circuit
103
having an input of the sampled received-signal Ssr for obtaining impulse responses of the sampled received-signal Ssr to output impulse-response signals Sir, a sampling-clock output circuit
111
having inputs of the impulse-response signals Sir and a demodulation-point setting-up signal Sdp for producing a sampling clock SCLK and a tap-coefficient selection signal Scsel, and a received-signal estimation circuit
112
having inputs of the impulse-response signals Sir and the tap-coefficient selection signal Scsel for producing a demodulation-point setting-up signal Sdp and N estimated received-signals Ser.
Before selecting a sample timing, the sampling-clock output circuit
111
having inputs of the impulse-response signals Sir and the demodulation-point setting-up signal Sdp outputs a sampling clock at the rate of L times of that after selecting. After selecting the sample timing, the sampling-clock output circuit
111
outputs a tap-coefficient selection signal Scsel according to the selected sample timing, and a sampling clock SCLK at the rate of 1/L times of that before selecting, according to the selected sample timing.
The received-signal estimation circuit
112
having inputs of the impulse-response signals Sir and the tap-coefficient selection signal Scsel outputs N estimated received-signals Ser and a demodulation-point setting-up signal Sdp corresponding to the demodulation component by the use of a tap coefficient of one among the impulse-response signals Sir corresponding to one selected from L sample timings, where L represents a positive integer, with the tap-coefficient selection signal Scsel. This received-signal estimation circuit
112
comprises a demodulation-point setting-up circuit
104
for producing a demodulation-point setting-up signal Sdp, a filter-coefficient output circuit
105
for producing filter-coefficient groups Tc
1
and Tc
2
, a counter
106
for producing a transmission signal sequence Scs, a precursor estimation circuit
107
for producing a precursor estimation signal Spr, an adder
108
for producing N estimated received-signals Ser, and a transversal filter
110
for producing a postcursor estimation signal Spo. The distortion components generated in a signal include precursor components, which are generated before the peak in the signal, and postcursor components, which are generated after the peak in the signal.
In this construction, the sampler
101
samples a received signal Sr with a sampling clock SCLK and outputs a sampled received-signal Ssr. The subtracter
102
subtracts each of N estimated received-signals Ser from the sampled received-signal Ssr to output each of N estimated error signals Seerr. The detector
109
having an input of the N estimated error signals Seerr detects the least significant bit of the sequence corresponding to one of the estimated error signals Seerr in which signal the minimum absolute value is obtained, as a value that the distortion components are removed from the received signal Sr, and outputs it to the outside as a detected output-signal Sd.
The impulse-response calculation circuit
103
having an input of a received signal Sr, for example, as shown in FIG.
8
(
a
), obtains impulse responses as shown in FIG.
8
(
b
) and outputs them as impulse-response signals Sir. The demodulation-point setting-up circuit
104
having inputs of the impulse-response signals Sir as shown in FIG.
8
(
b
) outputs a demodulation-point setting-up signal Sdp corresponding to the impulse response having the maximum absolute value with respect to each of L sample timings.
The filter-coefficient output circuit
105
has inputs of the impulse-response signals Sir, the demodulation-point setting-up signal Sdp and a tap-coefficient selection signal Scsel. If the m-th impulse response among the n impulse responses, where m and n represent positive integers, respectively, in the impulse-response signal Sir corresponding to a sample timing j selected from L sample timings, where j represents a positive integer, with the tap-coefficient selection signal Scsel is the maximum, this filter-coefficient output circuit
105
outputs the (m+1)th to n-th impulse responses as a filter-coefficient group Tc
1
, and the 1st to m-th impulse responses as a filter-coefficient group Tc
2
.
The counter
106
outputs a transmission signal sequence Scs that represents 0 to N-1 by binary number. The precursor estimation circuit
107
having inputs of the filter-coefficient group Tc
2
and the transmission signal sequence Scs estimates the precursor components of the received signal Sr and outputs N precursor estimation signals Spr. The adder
108
adds each of the N precursor estimation signals Spr to a postcursor estimation signal Spo to output each of N estimated received-signals Ser.
The transversal filter
110
having inputs of a detected output-signal Sd and the filter-coefficient group Tc
1
outputs postcursor estimation signals Spo corresponding to the postcursor components of the distortion. Before selecting a sample timing, the sampling-clock output circuit
111
having inputs of the impulse-response signals Sir and the demodulation-point setting-up signal Sdp outputs a sampling clock SCLK at the rate of L times of that after selecting. After selecting the sample timing by which the impulse response having the maximum peak value is obtained, the sampling-clock output circuit
111
outputs a tap-coefficient selection signal Scsel according to the selected sample timing, and a sampling clock SCLK at the rate of 1/L times of that before selecting, according to the selected sample timing.
As described above, in the prior art automatic equalizer, the timing is selected by which the impulse response having the maximum peak value is obtained. That is, in the sampling-clock output circuit
111
, as shown in
FIG. 9
, after obtaining the absolute values I of impulse responses (step
201
), the peak components IP
1
to IPL in the absolute values I are respectively obtained with respect to L sample timings (step
701
). The sample timing corresponding to the maximum one among the peak components IP
1
to IPL is then selected (step
702
).
In an automatic equalizer of feedback type (for example, an automatic equalizer described in Japanese Patent Unexamined Publication No. Hei 5-14126), in which the postcursor components are highly important in estimation, there is a case that its characteristics are rather good even in case of the peak value
Kakura Yoshikazu
Ohsawa Tomoki
Le Amanda T.
NEC Corporation
Scully Scott Murphy & Presser
LandOfFree
Automatic equalizer capable of surely selecting a suitable... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Automatic equalizer capable of surely selecting a suitable..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Automatic equalizer capable of surely selecting a suitable... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2615929