US20050280568A1 - Flash adc receiver with reduced errors - Google Patents
Flash adc receiver with reduced errors Download PDFInfo
- Publication number
- US20050280568A1 US20050280568A1 US10/870,054 US87005404A US2005280568A1 US 20050280568 A1 US20050280568 A1 US 20050280568A1 US 87005404 A US87005404 A US 87005404A US 2005280568 A1 US2005280568 A1 US 2005280568A1
- Authority
- US
- United States
- Prior art keywords
- adc
- voltage
- bits
- receiver
- decoding
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 claims description 10
- 230000004044 response Effects 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 4
- 238000012937 correction Methods 0.000 claims description 2
- 239000000872 buffer Substances 0.000 description 10
- 239000011159 matrix material Substances 0.000 description 8
- 238000007476 Maximum Likelihood Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 239000006185 dispersion Substances 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000002441 reversible effect Effects 0.000 description 2
- 230000007480 spreading Effects 0.000 description 2
- 230000002123 temporal effect Effects 0.000 description 2
- 230000005374 Kerr effect Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000033590 base-excision repair Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000010287 polarization Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001902 propagating effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000011144 upstream manufacturing Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M1/00—Analogue/digital conversion; Digital/analogue conversion
- H03M1/06—Continuously compensating for, or preventing, undesired influence of physical parameters
- H03M1/08—Continuously compensating for, or preventing, undesired influence of physical parameters of noise
- H03M1/0809—Continuously compensating for, or preventing, undesired influence of physical parameters of noise of bubble errors, i.e. irregularities in thermometer codes
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M1/00—Analogue/digital conversion; Digital/analogue conversion
- H03M1/12—Analogue/digital converters
- H03M1/18—Automatic control for modifying the range of signals the converter can handle, e.g. gain ranging
- H03M1/181—Automatic control for modifying the range of signals the converter can handle, e.g. gain ranging in feedback mode, i.e. by determining the range to be selected from one or more previous digital output values
- H03M1/182—Automatic control for modifying the range of signals the converter can handle, e.g. gain ranging in feedback mode, i.e. by determining the range to be selected from one or more previous digital output values the feedback signal controlling the reference levels of the analogue/digital converter
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M1/00—Analogue/digital conversion; Digital/analogue conversion
- H03M1/12—Analogue/digital converters
- H03M1/18—Automatic control for modifying the range of signals the converter can handle, e.g. gain ranging
- H03M1/181—Automatic control for modifying the range of signals the converter can handle, e.g. gain ranging in feedback mode, i.e. by determining the range to be selected from one or more previous digital output values
- H03M1/183—Automatic control for modifying the range of signals the converter can handle, e.g. gain ranging in feedback mode, i.e. by determining the range to be selected from one or more previous digital output values the feedback signal controlling the gain of an amplifier or attenuator preceding the analogue/digital converter
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M1/00—Analogue/digital conversion; Digital/analogue conversion
- H03M1/12—Analogue/digital converters
- H03M1/34—Analogue value compared with reference values
- H03M1/36—Analogue value compared with reference values simultaneously only, i.e. parallel type
- H03M1/361—Analogue value compared with reference values simultaneously only, i.e. parallel type having a separate comparator and reference value for each quantisation level, i.e. full flash converter type
Definitions
- This invention relates to the reduction of decoding errors when using a flash analog to digital converter.
- Telecommunications typically involves communicating a bit stream over a channel.
- the bit stream is typically encoded as an analog signal for transmission over the channel.
- the bit stream is decoded from the received analog signal.
- a real-world channel will impart distortions to the signal. It is the function of the receiver to endeavour to accurately recover the bit stream despite these distortions.
- ADC analog to digital decoder
- ADC analog to digital decoder
- One known type of ADC is a flash ADC which uses a set of 2 n ⁇ 1 comparators to directly measure the received analog signal to a resolution of n bits. For example, a three bit flash ADC will have seven comparators, each of which compares an input voltage with a different pre-set reference voltage level in order to “slice” the received analog voltage into one of eight levels.
- the analogue input signal is sampled at a clock rate reflective of the bit rate of the bit stream on the channel. Each sample then represents one bit in the bit stream.
- the voltage level of the sample inputs each comparator.
- the outputs from the comparators will indicate that the voltage of the sample lies between two reference voltages.
- the sample voltage may be digitised to the level of the lower of these two reference voltages.
- the digitised voltage then passes to an equaliser which attempts to address channel distortion in determining, based on the digitised voltage, whether the sample represents a binary 0 or a binary 1.
- Symbol decoding errors at a receiver utilising a flash analog to digital converter can be reduced by adjusting a reference voltage level of the ADC where a decoding error rate at the reference voltage level exceeds a threshold.
- a receiver comprising: a flash analog to digital converter (ADC); a symbol decoder input by said ADC; a link quality indicator input by said symbol decoder for providing link quality indications; a reference voltage adjuster input by said ADC and outputting to a control input of said ADC for, based on said link quality indications, selectively adjusting one or more reference voltages of said ADC.
- ADC flash analog to digital converter
- a receiver comprising: a flash analog to digital converter (ADC); decoding means responsive to said ADC to decode symbols; link quality detecting means responsive to said decoding means to detect symbol decoding errors; means responsive to said ADC and said symbol decoding errors for adjusting reference voltages of said ADC.
- ADC flash analog to digital converter
- a method of improving link quality comprising: where a link quality at a voltage slicing level of said ADC does not meet a threshold, adjusting a reference voltage for said voltage slicing level.
- FIG. 1 is a schematic view of a transmission system including a receiver made in accordance with this invention
- FIG. 2 is a graph of probability density functions (pdf's)
- FIG. 3 illustrates an exemplary look up table
- FIG. 4 illustrates an exemplary trellis in a Viterbi decoder
- FIG. 5 illustrates the trellis of FIG. 4 with probability weights
- FIG. 6 and FIG. 7 are graphs of pdf's overlaid with voltage slicing levels of a flash analog to digital converter (ADC).
- ADC analog to digital converter
- a transmission system 10 comprises a transmitter 12 , a channel 14 , and a receiver 16 .
- the channel may be, for example, an optical channel on an optical fibre.
- the receiver may comprise a serially arranged pre-processing block 18 , adaptive gain controller (AGC) 20 , flash ADC 22 , equaliser 24 , and forward error corrector (FEC) 26 .
- a clock recovery block 28 may recover a clock signal upstream from the ADC and input the recovered clock to the ADC and the equaliser.
- the equaliser may output to an ADC level controller 30 which inputs reference voltages to the ADC 22 .
- the equaliser 24 may also output to an AGC level controller 32 , which in turn outputs to the AGC 20 through a digital to analog converter (DAC) 34 .
- AGC level controller 32 which in turn outputs to the AGC 20 through a digital to analog converter (DAC) 34 .
- DAC digital to analog converter
- the seven comparators also receive one of seven reference voltages from the ADC level control 30 via a DAC 38 .
- the equaliser 24 may be a maximum likelihood sequence estimating (MLSE) equaliser.
- MLSE maximum likelihood sequence estimating
- an MLSE equaliser has a Viterbi Decoder 40 as a symbol decoder and a channel estimation block 42 .
- the channel estimation block has a look-up table 46 .
- the channel estimation block also has a weight update block 44 .
- the weight update block 44 is a processing block; therefore, the weight update block may be, for example, a processor or a field programmable gate array (FPGA).
- the Viterbi Decoder 40 receives the output from the ADC 22 and in turn outputs to the FEC 26 and to the weight update block 44 .
- the weight update block 44 receives the output of the Viterbi Decoder.
- the weight update block also receives the output from the ADC 22 and the FEC 26 .
- the weight update block outputs to the look-up table 46 .
- the look-up table 46 outputs to the Viterbi Decoder 40 and to the DAC Level Controller 30 .
- the channel estimation block 42 can act as an adjuster for the reference voltages of the ADC.
- ISI inter symbol interference
- the look up table 46 may store a plurality of possible channel impulse response models, one of which is designated as the initial impulse response model.
- the weight update block 44 may determine from the initial channel model the probability density functions (pdf's) for each of the eight possible three sequence combinations (from 000 to 111).
- the look up table also stores an initial set of voltage slicing levels which may be loaded into block 44 .
- Block 44 determines a probability for each of the possible three sequence combinations at each of the voltage slicing levels.
- the probability chosen for a sequence at a given voltage slicing level may be the maximum probability for that sequence in the range between the given voltage slicing level and the next higher voltage slicing level. Alternatively, it could be the average probability in that range. Since there are eight voltage slicing levels and eight three bit sequences, this will result in sixty-four probability values (weights) which are loaded into the 8 ⁇ 8 look-up table 44 .
- FIG. 2 is a graph of the eight pdf's that might have been generated from the channel model. As is understood, each pdf indicates the probability, P y , that, given the voltage, y k , of any received signal sample, that voltage is indicative of a given three bit series of bits. As shown in FIG. 2 , the pdf's are presented as the square root of the voltage against the log of the probability a given voltage represents a given bit sequence.
- a table 46 illustrated in FIG. 3 may be generated from the graph of FIG. 2 .
- Each cell in the table has a probability weight, this probability weight being the log of the probability value at a given slicing level for a given bit sequence.
- the Viterbi decoder 40 input with the voltage slicing level of each sample, stores a pre-set number of consecutive (voltage) samples in a buffer.
- the decoder also has a matrix having eight rows, one for each of the eight possible three bit sequences, and one column for each storage location in the buffer.
- the matrix in the decoder will be an 8 ⁇ 7 matrix, as illustrated in FIG. 4 .
- FIG. 4 illustrates all possible paths through the matrix. These paths form what is known as a trellis.
- the decoder can load appropriate probabilities from the table 46 into the matrix. More specifically, the first three columns of the matrix are loaded with the eight probabilities assigned to the eight possible three bit sequences where the sample voltage is the zero slicing voltage. From FIG. 3 this is seen to be 10 ⁇ 3, 10 ⁇ 7, . . . The next column is loaded with the eight probabilities for a 0.2 voltage slicing level, and so on. The result is illustrated in FIG. 5 (where, for ease of illustration, a short-hand notation has been used such that, for example, 10 ⁇ 3 is rendered as ⁇ 3 and 6 * 10 ⁇ 2 is rendered as 6-2).
- the decoder then adds up the probabilities for each path through the trellis and the highest probability path is chosen. This results in an assumed bit sequence in the buffer.
- the decoder then outputs one bit, that being the bit represented by the oldest (first) voltage sample in the buffer. All samples are then shifted one position in the buffer so that the oldest sample is discarded from one end of the buffer and a new sample is received in the other end of the buffer.
- the probability values for the decoder matrix are then updated from the table, the most likely path chosen, and one bit is output. This process then repeats.
- Some of the bits in the decoded (i.e., recovered) sequence may be error correction bits. These may be used by the FEC to correct errors.
- Errors noted by the FEC block may be fed back to the weight update block 44 .
- the weight update block may tune a selected channel model parameter by selecting a slightly different channel model from the look up table to designate as an updated channel model.
- This updated channel model may assume, for example, a slightly different chromatics dispersion, polarization mode dispersion, modal dispersion, or Kerr effect.
- PDF's generated from the updated channel model are then used to update the probability values stored in the look-up table 46 .
- the new set of probability values varies the decisions made by the Viterbi decoder as to the maximum likelihood three bit sequences. This, in turn, affects the BER.
- the weight update block may further tune the selected channel model parameter. If the BER goes down, the weight update block may return the magnitude of the selected parameter to that of its preceding value by returning to a previous channel model. In this way, the BER may be minimized with respect to the selected parameter. If, notwithstanding, the BER remains too high, the weight update block may repeat the process for a further channel model parameter, and so on. In this way, the channel model may be blindly adapted, as the distortion on the channel changes with time.
- An alternate embodiment may not employ FEC 26 in updating the channel model.
- a given short sequence say eight bits long—which is decoded from a sequence of input voltages is assumed to be free of errors. This is a reasonable assumption for a short sequence given the relatively low BER expected in a working channel.
- the weight update block 44 may take the decoded eight bit sequence output from the Viterbi decoder 40 and pass it through the current channel model in reverse direction. More specifically, using the probability density functions of the current channel model, each of the six different three bit sequences in the eight bit output sequence is assumed to have resulted from an input voltage corresponding to the maximum probability of that three bit sequence.
- the resulting (six) regenerated input voltages are then compared with the corresponding actual received input voltages. If there is a sufficiently strong correlation, the current channel model is assumed to be correct. If, however, the correlation is not sufficiently strong, the output sequence is passed through other channel models in reverse direction. If one of these other channel models provides a stronger correlation, that channel model is substituted as the current channel model.
- the reference voltages to the comparators of the flash ADC are fixed at equal increments.
- these increments may not allow discrimination between pdf's that are closely valued over the range of voltages represented by a voltage level.
- the square root of an actual analogue voltage were 0.95, this would be represented as a level 5 voltage, with a nominal square root value of 0.85.
- the maximum likelihood sequence is 011.
- the maximum likelihood sequence is 110 .
- the digitisation of the sample voltage may introduce errors. Even though these errors do not result from channel distortion, it may be possible to reduce these errors by adjusting the channel model (and thereby changing the pdf's). But even with such adjustments, errors resulting from the digitisation of the sample voltage will remain.
- the subject invention may reduce these errors by adjusting the reference voltages dependent upon the channel model. More particularly, the weight update block 44 is input with the voltage level for each sample. With this information and the errors reported by the FEC, the weight update block can determine whether errors at a given voltage level are over represented. In other words, the error rate at a given voltage level may be determined to exceed a pre-defined threshold. In such instance, the weight update block may change the voltage slicing levels and send the new levels to the table 46 . This will result in the table being updated with the new voltage slicing levels for use by the Viterbi decoder and in the table sending a control signal to the ADC level controller 30 with the new levels.
- the weight update block may reduce the voltage bands covered by both level 3 and level 4 voltages. These bands may be reduced incrementally, with the bands continuing to reduce as long as reductions decrease the errors at these voltage levels. The result may be as illustrated in FIG. 7 with the level 3 voltage covering a very narrow band, and the level 4 voltage covering a somewhat narrowed band.
- the adaptations are dynamic. Thus, if the characteristics of the channel change, the weight update block will update the channel model and the reference voltages in order to reduce errors.
- the weight update block may first attempt to minimize the BER by updating the channel model and then, once the channel model has been optimized, adapt the voltage slicing levels to further reduce the BER (at certain voltage slicing levels). This process may then be repeated periodically.
- the weight update block 44 updates the channel model, from the resulting pdf's, it may calculate a set of voltage slicing levels which will minimize the BER at each voltage slicing level. This is done by, for example, reducing the step between voltage slicing levels sufficiently to maximize the discrimination between two or more pdf's that are closely valued between the slicing levels.
- known test sequences may be received and BER statistics may be collected on these known sequences. More specifically, the BER may be collected for each of the sequences from 000 to 111 and these statistics used to update the slicing levels.
- the FEC is not used to determine the BERs, but instead the weight update model determines these, knowing the test sequences and the corresponding decoded sequences.
- the channel model may be static (i.e., the assumed pdf curves may be fixed), but the BER may nevertheless be reduced by adjusting the reference voltage levels of the ADC, as aforedescribed.
- look up table has been described as storing a plurality of channel models, it could alternatively simply store the pdf's that result from each channel model, thereby obviating the need for the weight update block to calculate these.
- weight update block has been described as making adjustments based on the BER, equally, it could make adjustments based on any other parameter which indicates link quality.
- the look up table does not store channel models; instead, the weight update block stores an initial channel impulse response model.
- the weight update block receives the output from the Viterbi decoder. This output is passed through the channel model and the result compared with the original input to the Viterbi decoder (which is also stored in the weight update block). If there is a good match, the channel model is assumed to be correct. This may be assessed by calculating the mean squared error for a number of blindly adapted channel impulse response models that are similar to the current channel model. If one of these blindly adapted models provides a lower mean squared error, it is chosen as the updated channel model. The updated channel model is used to calculate pdf's in order to update the matrix (of FIG. 3 ) in the look up table 46 . This process repeats as new output is received from the Viterbi decoder.
- the voltage slicing levels may be adapted as before described.
- the MLSE equaliser 24 has been described as decoding sequences of three bits, obviously sequences of a different number of bits, such as five bits, could be decoded. A five bit sequence would account for ISI not only for the bits immediately preceding and following a given bit, but also for the bit preceding the preceding bit and the bit following the following bit. While the channel 14 has been described as carrying an encoded bit stream to be decoded, more generally, the channel may carry an encoded symbol stream to be decoded.
- the equaliser 24 has been described as an MLSE equaliser, the equaliser may be any equaliser which can determine that the BER associated with certain voltage slicing levels is unacceptably high and can adjust such voltage slicing levels in an effort to reduce the BER associated with such voltage slicing levels.
- the equaliser could be a Feed Forward Equaliser (FFE), a Decision Feedback Equaliser (DFE), a Fractionally Spaced Equaliser, or a combined FFE/DFE.
Abstract
Description
- This invention relates to the reduction of decoding errors when using a flash analog to digital converter.
- Telecommunications typically involves communicating a bit stream over a channel. At the sending end, the bit stream is typically encoded as an analog signal for transmission over the channel. At a receiver, the bit stream is decoded from the received analog signal. A real-world channel will impart distortions to the signal. It is the function of the receiver to endeavour to accurately recover the bit stream despite these distortions.
- Where a bit stream is encoded as an analog modulated (AM) signal, at the receiver, after removal of any carrier wave, the signal may pass through an analog to digital decoder (ADC). One known type of ADC is a flash ADC which uses a set of 2n−1 comparators to directly measure the received analog signal to a resolution of n bits. For example, a three bit flash ADC will have seven comparators, each of which compares an input voltage with a different pre-set reference voltage level in order to “slice” the received analog voltage into one of eight levels. In use, the analogue input signal is sampled at a clock rate reflective of the bit rate of the bit stream on the channel. Each sample then represents one bit in the bit stream. For each sample, the voltage level of the sample inputs each comparator. The outputs from the comparators will indicate that the voltage of the sample lies between two reference voltages. In this way, the sample voltage may be digitised to the level of the lower of these two reference voltages. The digitised voltage then passes to an equaliser which attempts to address channel distortion in determining, based on the digitised voltage, whether the sample represents a
binary 0 or abinary 1. - One significant cause of channel distortions results from temporal spreading of the signal when propagating over long distances or over nonlinear medium. This phenonenon is not effectively addressed by known equalisers. Therefore, the current invention seeks to provide an improved equalisation approach.
- Symbol decoding errors at a receiver utilising a flash analog to digital converter (ADC) can be reduced by adjusting a reference voltage level of the ADC where a decoding error rate at the reference voltage level exceeds a threshold.
- In accordance to one aspect of the present invention, there is provided a receiver comprising: a flash analog to digital converter (ADC); a symbol decoder input by said ADC; a link quality indicator input by said symbol decoder for providing link quality indications; a reference voltage adjuster input by said ADC and outputting to a control input of said ADC for, based on said link quality indications, selectively adjusting one or more reference voltages of said ADC.
- In accordance to another aspect of the present invention, there is provided a receiver comprising: a flash analog to digital converter (ADC); decoding means responsive to said ADC to decode symbols; link quality detecting means responsive to said decoding means to detect symbol decoding errors; means responsive to said ADC and said symbol decoding errors for adjusting reference voltages of said ADC.
- In accordance to a further aspect of the present invention, there is provided in a receiver utilising a flash analog to digital converter (ADC), a method of improving link quality comprising: where a link quality at a voltage slicing level of said ADC does not meet a threshold, adjusting a reference voltage for said voltage slicing level.
- Other features and advantages will become apparent after referring to the following description in conjunction with the drawings.
- In the figures which illustrate example embodiments of the invention,
-
FIG. 1 is a schematic view of a transmission system including a receiver made in accordance with this invention, -
FIG. 2 is a graph of probability density functions (pdf's), -
FIG. 3 illustrates an exemplary look up table, -
FIG. 4 illustrates an exemplary trellis in a Viterbi decoder, -
FIG. 5 illustrates the trellis ofFIG. 4 with probability weights, and -
FIG. 6 andFIG. 7 are graphs of pdf's overlaid with voltage slicing levels of a flash analog to digital converter (ADC). - Turning to
FIG. 1 , atransmission system 10 comprises atransmitter 12, achannel 14, and areceiver 16. The channel may be, for example, an optical channel on an optical fibre. The receiver may comprise a serially arranged pre-processingblock 18, adaptive gain controller (AGC) 20,flash ADC 22, equaliser 24, and forward error corrector (FEC) 26. Aclock recovery block 28 may recover a clock signal upstream from the ADC and input the recovered clock to the ADC and the equaliser. The equaliser may output to anADC level controller 30 which inputs reference voltages to theADC 22. Theequaliser 24 may also output to anAGC level controller 32, which in turn outputs to theAGC 20 through a digital to analog converter (DAC) 34. Each of thepreprocessor 18,AGC level controller 32, and AGC 20 operate in a conventional manner and so is not further described herein. - As illustrated, the
flash ADC 22 has (22−=1=) sevencomparators 36, each input by any signal at theAGC 20. The seven comparators also receive one of seven reference voltages from theADC level control 30 via aDAC 38. - The
equaliser 24 may be a maximum likelihood sequence estimating (MLSE) equaliser. As is understood by those skilled in the art, an MLSE equaliser has a ViterbiDecoder 40 as a symbol decoder and achannel estimation block 42. The channel estimation block has a look-up table 46. In accordance with this invention, the channel estimation block also has aweight update block 44. Theweight update block 44 is a processing block; therefore, the weight update block may be, for example, a processor or a field programmable gate array (FPGA). - The Viterbi Decoder 40 receives the output from the
ADC 22 and in turn outputs to the FEC 26 and to theweight update block 44. In one embodiment (described hereinafter), theweight update block 44 receives the output of the Viterbi Decoder. The weight update block also receives the output from theADC 22 and the FEC 26. The weight update block outputs to the look-up table 46. The look-up table 46 outputs to the Viterbi Decoder 40 and to theDAC Level Controller 30. For reasons which will become apparent, thechannel estimation block 42 can act as an adjuster for the reference voltages of the ADC. - While any signal sample is supposed to represent only a single bit, in fact, due to temporal spreading, there will be inter symbol interference (ISI). The consequence of ISI is that the state (0 or 1) of the bits preceding and following the bit represented by the sample will affect the voltage level of the current bit. In essence, some of the power in preceding and following bits spreads into to the current bit, changing its voltage level. For this reason, it is common for an MLSE equaliser to operate on three bit sequences (i.e., it is assumed that the ISI results from the bit before and the bit after the current bit). Therefore, in the following example, a three bit sequence is used to adjust for ISI.
- In a first embodiment, the look up table 46 may store a plurality of possible channel impulse response models, one of which is designated as the initial impulse response model. The
weight update block 44 may determine from the initial channel model the probability density functions (pdf's) for each of the eight possible three sequence combinations (from 000 to 111). The look up table also stores an initial set of voltage slicing levels which may be loaded intoblock 44.Block 44 then determines a probability for each of the possible three sequence combinations at each of the voltage slicing levels. The probability chosen for a sequence at a given voltage slicing level may be the maximum probability for that sequence in the range between the given voltage slicing level and the next higher voltage slicing level. Alternatively, it could be the average probability in that range. Since there are eight voltage slicing levels and eight three bit sequences, this will result in sixty-four probability values (weights) which are loaded into the 8×8 look-up table 44. -
FIG. 2 is a graph of the eight pdf's that might have been generated from the channel model. As is understood, each pdf indicates the probability, Py, that, given the voltage, yk, of any received signal sample, that voltage is indicative of a given three bit series of bits. As shown inFIG. 2 , the pdf's are presented as the square root of the voltage against the log of the probability a given voltage represents a given bit sequence. - Assuming that the eight voltage slicing levels are 0.0; 0.2; 0.4; 0.6; 0.8; 1.0; 1.2; and 1.4, a table 46 illustrated in
FIG. 3 may be generated from the graph ofFIG. 2 . Each cell in the table has a probability weight, this probability weight being the log of the probability value at a given slicing level for a given bit sequence. - In a manner well understood by those skilled in the art, the
Viterbi decoder 40, input with the voltage slicing level of each sample, stores a pre-set number of consecutive (voltage) samples in a buffer. The decoder also has a matrix having eight rows, one for each of the eight possible three bit sequences, and one column for each storage location in the buffer. - Say, for example, the buffer is seven samples long (actual buffers are likely to be much longer). Thus, the matrix in the decoder will be an 8×7 matrix, as illustrated in
FIG. 4 . Next assume that the voltage levels of the first four consecutive samples are received into the buffer (and that the voltage level of earlier samples is assumed to be 0). With these assumptions,FIG. 4 illustrates all possible paths through the matrix. These paths form what is known as a trellis. - If the voltage levels of these first four samples are 0.2; 0.8; 0.6; and 0.8, with this (and the fact that earlier voltages are assumed to be zero), the decoder can load appropriate probabilities from the table 46 into the matrix. More specifically, the first three columns of the matrix are loaded with the eight probabilities assigned to the eight possible three bit sequences where the sample voltage is the zero slicing voltage. From
FIG. 3 this is seen to be 10ˆ−3, 10ˆ−7, . . . The next column is loaded with the eight probabilities for a 0.2 voltage slicing level, and so on. The result is illustrated inFIG. 5 (where, for ease of illustration, a short-hand notation has been used such that, for example, 10ˆ−3 is rendered as −3 and 6*10ˆ−2 is rendered as 6-2). - The decoder then adds up the probabilities for each path through the trellis and the highest probability path is chosen. This results in an assumed bit sequence in the buffer. The decoder then outputs one bit, that being the bit represented by the oldest (first) voltage sample in the buffer. All samples are then shifted one position in the buffer so that the oldest sample is discarded from one end of the buffer and a new sample is received in the other end of the buffer. The probability values for the decoder matrix are then updated from the table, the most likely path chosen, and one bit is output. This process then repeats.
- Some of the bits in the decoded (i.e., recovered) sequence may be error correction bits. These may be used by the FEC to correct errors.
- Errors noted by the FEC block may be fed back to the
weight update block 44. If the bit error ratio (BER) rises above an acceptable standard, the weight update block may tune a selected channel model parameter by selecting a slightly different channel model from the look up table to designate as an updated channel model. This updated channel model may assume, for example, a slightly different chromatics dispersion, polarization mode dispersion, modal dispersion, or Kerr effect. Pdf's generated from the updated channel model are then used to update the probability values stored in the look-up table 46. The new set of probability values varies the decisions made by the Viterbi decoder as to the maximum likelihood three bit sequences. This, in turn, affects the BER. If the BER goes down, the weight update block may further tune the selected channel model parameter. If the BER goes up, the weight update block may return the magnitude of the selected parameter to that of its preceding value by returning to a previous channel model. In this way, the BER may be minimized with respect to the selected parameter. If, notwithstanding, the BER remains too high, the weight update block may repeat the process for a further channel model parameter, and so on. In this way, the channel model may be blindly adapted, as the distortion on the channel changes with time. - An alternate embodiment may not employ
FEC 26 in updating the channel model. In such instance, while the channel is operating, a given short sequence—say eight bits long—which is decoded from a sequence of input voltages is assumed to be free of errors. This is a reasonable assumption for a short sequence given the relatively low BER expected in a working channel. Theweight update block 44 may take the decoded eight bit sequence output from theViterbi decoder 40 and pass it through the current channel model in reverse direction. More specifically, using the probability density functions of the current channel model, each of the six different three bit sequences in the eight bit output sequence is assumed to have resulted from an input voltage corresponding to the maximum probability of that three bit sequence. The resulting (six) regenerated input voltages are then compared with the corresponding actual received input voltages. If there is a sufficiently strong correlation, the current channel model is assumed to be correct. If, however, the correlation is not sufficiently strong, the output sequence is passed through other channel models in reverse direction. If one of these other channel models provides a stronger correlation, that channel model is substituted as the current channel model. - As illustrated in
FIG. 6 , in known arrangements, the reference voltages to the comparators of the flash ADC are fixed at equal increments. A drawback with this it that these increments may not allow discrimination between pdf's that are closely valued over the range of voltages represented by a voltage level. For example, with reference toFIG. 6 , if the square root of an actual analogue voltage were 0.95, this would be represented as alevel 5 voltage, with a nominal square root value of 0.85. At 0.85, the maximum likelihood sequence is 011. However, at 0.95, the maximum likelihood sequence is 110. Thus, the digitisation of the sample voltage may introduce errors. Even though these errors do not result from channel distortion, it may be possible to reduce these errors by adjusting the channel model (and thereby changing the pdf's). But even with such adjustments, errors resulting from the digitisation of the sample voltage will remain. - The subject invention may reduce these errors by adjusting the reference voltages dependent upon the channel model. More particularly, the
weight update block 44 is input with the voltage level for each sample. With this information and the errors reported by the FEC, the weight update block can determine whether errors at a given voltage level are over represented. In other words, the error rate at a given voltage level may be determined to exceed a pre-defined threshold. In such instance, the weight update block may change the voltage slicing levels and send the new levels to the table 46. This will result in the table being updated with the new voltage slicing levels for use by the Viterbi decoder and in the table sending a control signal to theADC level controller 30 with the new levels. - For example, with reference to
FIG. 6 , it will be noted that the exemplary pdf's for each of 100, 101, 010 and 011 all cross in the vicinity of a square root of the voltage of 0.65. Thus, sample voltages resolved tolevel 3 orlevel 4 voltages will introduce errors. Receiving indications of these errors at these levels, the weight update block may reduce the voltage bands covered by bothlevel 3 andlevel 4 voltages. These bands may be reduced incrementally, with the bands continuing to reduce as long as reductions decrease the errors at these voltage levels. The result may be as illustrated inFIG. 7 with thelevel 3 voltage covering a very narrow band, and thelevel 4 voltage covering a somewhat narrowed band. - The adaptations are dynamic. Thus, if the characteristics of the channel change, the weight update block will update the channel model and the reference voltages in order to reduce errors.
- In one approach, the weight update block may first attempt to minimize the BER by updating the channel model and then, once the channel model has been optimized, adapt the voltage slicing levels to further reduce the BER (at certain voltage slicing levels). This process may then be repeated periodically. In an alternate embodiment, whenever the
weight update block 44 updates the channel model, from the resulting pdf's, it may calculate a set of voltage slicing levels which will minimize the BER at each voltage slicing level. This is done by, for example, reducing the step between voltage slicing levels sufficiently to maximize the discrimination between two or more pdf's that are closely valued between the slicing levels. In a further embodiment which will more quickly update the slicing level, known test sequences may be received and BER statistics may be collected on these known sequences. More specifically, the BER may be collected for each of the sequences from 000 to 111 and these statistics used to update the slicing levels. In this embodiment, the FEC is not used to determine the BERs, but instead the weight update model determines these, knowing the test sequences and the corresponding decoded sequences. - In another embodiment, the channel model may be static (i.e., the assumed pdf curves may be fixed), but the BER may nevertheless be reduced by adjusting the reference voltage levels of the ADC, as aforedescribed.
- While the look up table has been described as storing a plurality of channel models, it could alternatively simply store the pdf's that result from each channel model, thereby obviating the need for the weight update block to calculate these.
- While the weight update block has been described as making adjustments based on the BER, equally, it could make adjustments based on any other parameter which indicates link quality.
- In another embodiment, the look up table does not store channel models; instead, the weight update block stores an initial channel impulse response model. In this embodiment, the weight update block receives the output from the Viterbi decoder. This output is passed through the channel model and the result compared with the original input to the Viterbi decoder (which is also stored in the weight update block). If there is a good match, the channel model is assumed to be correct. This may be assessed by calculating the mean squared error for a number of blindly adapted channel impulse response models that are similar to the current channel model. If one of these blindly adapted models provides a lower mean squared error, it is chosen as the updated channel model. The updated channel model is used to calculate pdf's in order to update the matrix (of
FIG. 3 ) in the look up table 46. This process repeats as new output is received from the Viterbi decoder. The voltage slicing levels may be adapted as before described. - While the
MLSE equaliser 24 has been described as decoding sequences of three bits, obviously sequences of a different number of bits, such as five bits, could be decoded. A five bit sequence would account for ISI not only for the bits immediately preceding and following a given bit, but also for the bit preceding the preceding bit and the bit following the following bit. While thechannel 14 has been described as carrying an encoded bit stream to be decoded, more generally, the channel may carry an encoded symbol stream to be decoded. - While the
equaliser 24 has been described as an MLSE equaliser, the equaliser may be any equaliser which can determine that the BER associated with certain voltage slicing levels is unacceptably high and can adjust such voltage slicing levels in an effort to reduce the BER associated with such voltage slicing levels. Thus, for example, the equaliser could be a Feed Forward Equaliser (FFE), a Decision Feedback Equaliser (DFE), a Fractionally Spaced Equaliser, or a combined FFE/DFE. - Other modifications will be apparent to those skilled in the art and, therefore, the invention is defined in the claims.
Claims (20)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US10/870,054 US6980140B1 (en) | 2004-06-18 | 2004-06-18 | Flash ADC receiver with reduced errors |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US10/870,054 US6980140B1 (en) | 2004-06-18 | 2004-06-18 | Flash ADC receiver with reduced errors |
Publications (2)
Publication Number | Publication Date |
---|---|
US20050280568A1 true US20050280568A1 (en) | 2005-12-22 |
US6980140B1 US6980140B1 (en) | 2005-12-27 |
Family
ID=35480062
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US10/870,054 Active US6980140B1 (en) | 2004-06-18 | 2004-06-18 | Flash ADC receiver with reduced errors |
Country Status (1)
Country | Link |
---|---|
US (1) | US6980140B1 (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030231722A1 (en) * | 2002-06-18 | 2003-12-18 | Chien-Cheng Tung | Symbol-based decision feedback equalizer (DFE) with maximum likelihood sequence estimation for wireless receivers under multipath channels |
US20060280258A1 (en) * | 2005-06-14 | 2006-12-14 | Ido Bettesh | Modulator and method for producing a modulated signal |
US20100104000A1 (en) * | 2008-10-24 | 2010-04-29 | Stmicroelectronics S.R.L. | Enhancement of transition region equalization in a decision feedback equalizer |
US8362937B2 (en) | 2009-06-12 | 2013-01-29 | Taiwan Semiconductor Manufacturing Company, Ltd. | Integrated circuits for converting analog signals to digital signals, systems, and operating methods thereof |
WO2013037529A1 (en) * | 2011-09-12 | 2013-03-21 | Robert Bosch Gmbh | Method for amplifying an echo signal suitable for vehicle environment detection and device for performing said method |
US20130114766A1 (en) * | 2010-07-30 | 2013-05-09 | Telefonaktiebolaget L M Ericsson (Publ) | Decoding Technique for Tail-Biting Codes |
US20130234755A1 (en) * | 2012-03-07 | 2013-09-12 | Realtek Semiconductor Corp. | Impedance calibration device and method |
CN103312293A (en) * | 2012-03-15 | 2013-09-18 | 瑞昱半导体股份有限公司 | Impedance correction device and impedance correction method |
US20140325319A1 (en) * | 2013-03-20 | 2014-10-30 | Zte (Usa) Inc. | Statistics adaptive soft decision forward error correction in digital communication |
US9329929B2 (en) | 2013-03-20 | 2016-05-03 | Zte (Usa) Inc. | Soft maximum likelihood sequence estimation in digital communication |
US9755763B2 (en) | 2014-07-16 | 2017-09-05 | Zte Corporation | Adaptive post digital filter and inter-symbol interference equalizer for optical communication |
US9819419B2 (en) | 2014-10-07 | 2017-11-14 | Zte Corporation | Maximum likelihood sequence estimation of Quadrature Amplitude Modulated signals |
US10917203B2 (en) * | 2019-05-17 | 2021-02-09 | Oracle International Corporation | Estimate bit error rates of network cables |
Families Citing this family (24)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7805642B1 (en) * | 2006-02-17 | 2010-09-28 | Aquantia Corporation | Low power iterative decoder using input data pipelining and voltage scaling |
US7492292B2 (en) * | 2007-05-29 | 2009-02-17 | Intel Corporation | Calibrating an analog component using digital feedback information |
US7696915B2 (en) * | 2008-04-24 | 2010-04-13 | Agere Systems Inc. | Analog-to-digital converter having reduced number of activated comparators |
US20090287969A1 (en) * | 2008-05-13 | 2009-11-19 | Bpm Microsystems | Electronic apparatus and bit error rate tolerance method for programming non-volatile memory devices |
JP5710475B2 (en) * | 2008-07-01 | 2015-04-30 | エルエスアイ コーポレーション | Method and apparatus for soft demapping and inter-cell interference mitigation in flash memory |
JP5535219B2 (en) * | 2008-09-30 | 2014-07-02 | エルエスアイ コーポレーション | Method and apparatus for soft data generation in a memory device using a reference cell |
WO2010042319A2 (en) * | 2008-10-10 | 2010-04-15 | Ziva Corporation | Techniques and systems for wireless communications |
US9449719B2 (en) * | 2008-12-19 | 2016-09-20 | Seagate Technology Llc | Solid-state storage device including a high resolution analog-to-digital converter |
WO2011029075A2 (en) | 2009-09-03 | 2011-03-10 | Ziva Corporation | Techniques and systems for communications based on time reversal pre-coding |
US9215112B2 (en) | 2010-02-23 | 2015-12-15 | Rambus Inc. | Decision feedback equalizer |
CN102783106B (en) * | 2010-03-16 | 2015-04-29 | 日本电气株式会社 | Digital receiver and optical communication system that uses same |
US9292377B2 (en) | 2011-01-04 | 2016-03-22 | Seagate Technology Llc | Detection and decoding in flash memories using correlation of neighboring bits and probability based reliability values |
US9898361B2 (en) | 2011-01-04 | 2018-02-20 | Seagate Technology Llc | Multi-tier detection and decoding in flash memories |
US9502117B2 (en) | 2011-03-14 | 2016-11-22 | Seagate Technology Llc | Cell-level statistics collection for detection and decoding in flash memories |
US8842026B2 (en) * | 2012-12-05 | 2014-09-23 | Infineon Technologies Ag | Symbol decoder, threshold estimation and correlation systems and methods |
US8981982B2 (en) * | 2013-04-05 | 2015-03-17 | Maxlinear, Inc. | Multi-zone data converters |
US8928506B2 (en) * | 2013-04-09 | 2015-01-06 | Maxlinear, Inc. | Successive approximation analog-to-digital converter (ADC) with dynamic search algorithm |
US9083376B2 (en) | 2013-04-25 | 2015-07-14 | Maxlinear, Inc. | Successive approximation register analog-to-digital converter |
US10135642B2 (en) | 2016-02-29 | 2018-11-20 | Rambus Inc. | Serial link receiver with improved bandwidth and accurate eye monitor |
US10103743B1 (en) * | 2017-11-09 | 2018-10-16 | Huawei Technologies Canada Co., Ltd. | Analog-to-digital converter and control method thereof |
US10594281B1 (en) | 2019-03-04 | 2020-03-17 | Ciena Corporation | Receiver automatic gain control systems and methods for asymmetrical or unbalanced constellations |
US10715169B1 (en) | 2019-05-21 | 2020-07-14 | Ciena Corporation | Coarse-fine gain-tracking loop and method of operating |
US11552703B2 (en) | 2020-12-09 | 2023-01-10 | Ciena Corporation | Detecting power of low-bandwidth and broad-bandwidth optical signals |
US11404596B1 (en) | 2021-04-20 | 2022-08-02 | Ciena Corporation | Balancing a pair of avalanche photodiodes in a coherent receiver |
Citations (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4823360A (en) * | 1988-02-12 | 1989-04-18 | Northern Telecom Limited | Binary data regenerator with adaptive threshold level |
US5606540A (en) * | 1995-01-20 | 1997-02-25 | Pioneer Electronic Corporation | Digital signal reproducing apparatus |
US5663945A (en) * | 1994-12-13 | 1997-09-02 | Pioneer Electronic Corporation | Digital phase locked loop with a digital voltage controlled oscillator in a recording information reproducing apparatus |
US5809071A (en) * | 1993-09-19 | 1998-09-15 | Canon Kabushiki Kaisha | Signal processing apparatus |
US5896391A (en) * | 1996-12-19 | 1999-04-20 | Northern Telecom Limited | Forward error correction assisted receiver optimization |
US5944844A (en) * | 1995-02-23 | 1999-08-31 | Nokia Telecommunications Oy | Method for determining connection quality in a receiver utilizing a Viterbi decoder |
US6097769A (en) * | 1998-02-10 | 2000-08-01 | Lucent Technologies Inc. | Viterbi detector using path memory controlled by best state information |
US6307899B1 (en) * | 1998-06-16 | 2001-10-23 | Ameritech Corporation | Method and system for optimizing coding gain |
US6603416B2 (en) * | 2001-10-01 | 2003-08-05 | International Business Machines Corporation | Method and circuit for dynamic calibration of flash analog to digital converters |
US6618436B2 (en) * | 1998-04-30 | 2003-09-09 | Mysticom Ltd. | Digital base-band receiver |
US20040030965A1 (en) * | 2002-08-12 | 2004-02-12 | Sevgui Hadjihassan | Method and apparatus for adjusting receiver voltage threshold and phase sampling point using FEC counts |
US6735724B1 (en) * | 1999-04-08 | 2004-05-11 | Texas Instruments Incorporated | Detection error estimation and method |
US6735259B1 (en) * | 1999-12-20 | 2004-05-11 | Nortel Networks Limited | Method and apparatus for optimization of a data communications system using sacrificial bits |
US20040091273A1 (en) * | 2001-12-26 | 2004-05-13 | Patrice Brissette | Receiver monitoring and optimization using forward error correction information |
US6877117B1 (en) * | 2001-07-27 | 2005-04-05 | Ciena Corporation | Optical signal receiver and method with decision threshold adjustment based on a relative percentage error indicator |
US6894985B2 (en) * | 2002-08-05 | 2005-05-17 | Harris Corporation | Monitoring link quality in a mobile ad hoc network |
-
2004
- 2004-06-18 US US10/870,054 patent/US6980140B1/en active Active
Patent Citations (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4823360A (en) * | 1988-02-12 | 1989-04-18 | Northern Telecom Limited | Binary data regenerator with adaptive threshold level |
US5809071A (en) * | 1993-09-19 | 1998-09-15 | Canon Kabushiki Kaisha | Signal processing apparatus |
US5663945A (en) * | 1994-12-13 | 1997-09-02 | Pioneer Electronic Corporation | Digital phase locked loop with a digital voltage controlled oscillator in a recording information reproducing apparatus |
US5606540A (en) * | 1995-01-20 | 1997-02-25 | Pioneer Electronic Corporation | Digital signal reproducing apparatus |
US5944844A (en) * | 1995-02-23 | 1999-08-31 | Nokia Telecommunications Oy | Method for determining connection quality in a receiver utilizing a Viterbi decoder |
US5896391A (en) * | 1996-12-19 | 1999-04-20 | Northern Telecom Limited | Forward error correction assisted receiver optimization |
US6097769A (en) * | 1998-02-10 | 2000-08-01 | Lucent Technologies Inc. | Viterbi detector using path memory controlled by best state information |
US6618436B2 (en) * | 1998-04-30 | 2003-09-09 | Mysticom Ltd. | Digital base-band receiver |
US6307899B1 (en) * | 1998-06-16 | 2001-10-23 | Ameritech Corporation | Method and system for optimizing coding gain |
US6735724B1 (en) * | 1999-04-08 | 2004-05-11 | Texas Instruments Incorporated | Detection error estimation and method |
US6735259B1 (en) * | 1999-12-20 | 2004-05-11 | Nortel Networks Limited | Method and apparatus for optimization of a data communications system using sacrificial bits |
US6877117B1 (en) * | 2001-07-27 | 2005-04-05 | Ciena Corporation | Optical signal receiver and method with decision threshold adjustment based on a relative percentage error indicator |
US6603416B2 (en) * | 2001-10-01 | 2003-08-05 | International Business Machines Corporation | Method and circuit for dynamic calibration of flash analog to digital converters |
US20040091273A1 (en) * | 2001-12-26 | 2004-05-13 | Patrice Brissette | Receiver monitoring and optimization using forward error correction information |
US6894985B2 (en) * | 2002-08-05 | 2005-05-17 | Harris Corporation | Monitoring link quality in a mobile ad hoc network |
US20040030965A1 (en) * | 2002-08-12 | 2004-02-12 | Sevgui Hadjihassan | Method and apparatus for adjusting receiver voltage threshold and phase sampling point using FEC counts |
Cited By (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030231722A1 (en) * | 2002-06-18 | 2003-12-18 | Chien-Cheng Tung | Symbol-based decision feedback equalizer (DFE) with maximum likelihood sequence estimation for wireless receivers under multipath channels |
US7197094B2 (en) * | 2002-06-18 | 2007-03-27 | Ralink Technology, Inc. | Symbol-based decision feedback equalizer (DFE) with maximum likelihood sequence estimation for wireless receivers under multipath channels |
US20060280258A1 (en) * | 2005-06-14 | 2006-12-14 | Ido Bettesh | Modulator and method for producing a modulated signal |
US7778356B2 (en) * | 2005-06-14 | 2010-08-17 | Given Imaging Ltd. | Modulator and method for producing a modulated signal |
US20100104000A1 (en) * | 2008-10-24 | 2010-04-29 | Stmicroelectronics S.R.L. | Enhancement of transition region equalization in a decision feedback equalizer |
US8315300B2 (en) * | 2008-10-24 | 2012-11-20 | Stmicroelectronics S.R.L. | Enhancement of transition region equalization in a decision feedback equalizer |
US8362937B2 (en) | 2009-06-12 | 2013-01-29 | Taiwan Semiconductor Manufacturing Company, Ltd. | Integrated circuits for converting analog signals to digital signals, systems, and operating methods thereof |
US20130114766A1 (en) * | 2010-07-30 | 2013-05-09 | Telefonaktiebolaget L M Ericsson (Publ) | Decoding Technique for Tail-Biting Codes |
US8995583B2 (en) * | 2010-07-30 | 2015-03-31 | Telefonaktiebolaget L M Ericsson (Publ) | Decoding technique for tail-biting codes |
WO2013037529A1 (en) * | 2011-09-12 | 2013-03-21 | Robert Bosch Gmbh | Method for amplifying an echo signal suitable for vehicle environment detection and device for performing said method |
KR101885410B1 (en) | 2011-09-12 | 2018-08-03 | 로베르트 보쉬 게엠베하 | Method for amplifying an echo signal for vehicle environment detection and device for performing said method |
US9698802B2 (en) | 2011-09-12 | 2017-07-04 | Robert Bosch Gmbh | Method for amplifying an echo signal suitable for vehicle surroundings detection and device for carrying out the method |
KR20140065408A (en) * | 2011-09-12 | 2014-05-29 | 로베르트 보쉬 게엠베하 | Method for amplifying an echo signal suitable for vehicle environment detection and device for performing said method |
TWI500261B (en) * | 2012-03-07 | 2015-09-11 | Realtek Semiconductor Corp | Impedance calibration device and method therefor |
US9018974B2 (en) * | 2012-03-07 | 2015-04-28 | Realtek Semiconductor Corp. | Impedance calibration device and method |
US20130234755A1 (en) * | 2012-03-07 | 2013-09-12 | Realtek Semiconductor Corp. | Impedance calibration device and method |
CN103312293A (en) * | 2012-03-15 | 2013-09-18 | 瑞昱半导体股份有限公司 | Impedance correction device and impedance correction method |
US20140325319A1 (en) * | 2013-03-20 | 2014-10-30 | Zte (Usa) Inc. | Statistics adaptive soft decision forward error correction in digital communication |
US9329929B2 (en) | 2013-03-20 | 2016-05-03 | Zte (Usa) Inc. | Soft maximum likelihood sequence estimation in digital communication |
US9762351B2 (en) * | 2013-03-20 | 2017-09-12 | Zte (Usa) Inc. | Statistics adaptive soft decision forward error correction in digital communication |
US9755763B2 (en) | 2014-07-16 | 2017-09-05 | Zte Corporation | Adaptive post digital filter and inter-symbol interference equalizer for optical communication |
US9819419B2 (en) | 2014-10-07 | 2017-11-14 | Zte Corporation | Maximum likelihood sequence estimation of Quadrature Amplitude Modulated signals |
US10917203B2 (en) * | 2019-05-17 | 2021-02-09 | Oracle International Corporation | Estimate bit error rates of network cables |
Also Published As
Publication number | Publication date |
---|---|
US6980140B1 (en) | 2005-12-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US6980140B1 (en) | Flash ADC receiver with reduced errors | |
US7460619B2 (en) | Method and system for optimizing coding gain | |
JP4741254B2 (en) | Decision feedback equalizer and feedback filter coefficient update method | |
US5621764A (en) | Soft decision signal outputting receiver | |
JP4063677B2 (en) | Two-stage equalizer for trellis coded systems | |
US6504868B1 (en) | Adaptive equalizer | |
US6178209B1 (en) | Method of estimating trellis encoded symbols utilizing simplified trellis decoding | |
US20040085917A1 (en) | Channel estimation apparatus and methods | |
EP1612968B1 (en) | Reduced bitstream candidate based receiver and received signal processing method | |
US20070104264A1 (en) | Apparatus and method of decision feedback equalization in terrestrial digital broadcasting receiver | |
US7184477B2 (en) | Decision feedback structure with selective sampling phase control | |
EP1645093A1 (en) | Channel estimation and sequence estimation for the reception of optical signal | |
WO2005004459A2 (en) | Method and apparatus for delayed recursion decoder | |
US5307374A (en) | Digital receiver with reduced memory requirement for Viterbi equalization | |
KR100988225B1 (en) | Apparatus for and method of providing trellis decoded data, and equalizer/trellis decoder system | |
US8462037B2 (en) | Method and system having adjustable analog-to-digital conversion levels | |
US20220103404A1 (en) | Direct digital sequence detection and equalization | |
JP4443874B2 (en) | Equalization system | |
EP2398175A1 (en) | Method and circuit for BER estimation | |
US11456900B2 (en) | Data receiving device and method | |
US20070183489A1 (en) | Apparatus for decoding a signal and method thereof and a trellis coded modulation decoder and method thereof | |
US10848298B1 (en) | Method and apparatus for a selective pulse amplitude modulation signal phase detector | |
US6782046B1 (en) | Decision-directed adaptation for coded modulation | |
EP1633050B1 (en) | Analog/digital conversion with adjustable thresholds | |
US8644369B1 (en) | Equalizer adaptation for heavily compressed or clipped communications signals |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: NORTEL NETWORKS LIMITED, CANADA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:ROWLAND, ANDY;LUK, TOM;HADJIHASSAN, SEVGUI;REEL/FRAME:015492/0229 Effective date: 20040608 |
|
FEPP | Fee payment procedure |
Free format text: PAYOR NUMBER ASSIGNED (ORIGINAL EVENT CODE: ASPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
STCF | Information on status: patent grant |
Free format text: PATENTED CASE |
|
AS | Assignment |
Owner name: DYMO, BELGIUM Free format text: CORRECTIVE ASSIGNMENT, REEL 017636, FRAME 0935;ASSIGNOR:ESSELTE;REEL/FRAME:017706/0321 Effective date: 20051108 Owner name: DYMO,BELGIUM Free format text: CORRECTIVE ASSIGNMENT, REEL 017636, FRAME 0935, RE;ASSIGNOR:ESSELTE;REEL/FRAME:017706/0321 Effective date: 20051108 |
|
FPAY | Fee payment |
Year of fee payment: 4 |
|
AS | Assignment |
Owner name: CIENA LUXEMBOURG S.A.R.L.,LUXEMBOURG Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:NORTEL NETWORKS LIMITED;REEL/FRAME:024213/0653 Effective date: 20100319 Owner name: CIENA LUXEMBOURG S.A.R.L., LUXEMBOURG Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:NORTEL NETWORKS LIMITED;REEL/FRAME:024213/0653 Effective date: 20100319 |
|
AS | Assignment |
Owner name: CIENA CORPORATION,MARYLAND Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:CIENA LUXEMBOURG S.A.R.L.;REEL/FRAME:024252/0060 Effective date: 20100319 Owner name: CIENA CORPORATION, MARYLAND Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:CIENA LUXEMBOURG S.A.R.L.;REEL/FRAME:024252/0060 Effective date: 20100319 |
|
FPAY | Fee payment |
Year of fee payment: 8 |
|
AS | Assignment |
Owner name: DEUTSCHE BANK AG NEW YORK BRANCH, NEW YORK Free format text: SECURITY INTEREST;ASSIGNOR:CIENA CORPORATION;REEL/FRAME:033329/0417 Effective date: 20140715 |
|
AS | Assignment |
Owner name: BANK OF AMERICA, N.A., AS ADMINISTRATIVE AGENT, NO Free format text: PATENT SECURITY AGREEMENT;ASSIGNOR:CIENA CORPORATION;REEL/FRAME:033347/0260 Effective date: 20140715 |
|
FEPP | Fee payment procedure |
Free format text: PAYOR NUMBER ASSIGNED (ORIGINAL EVENT CODE: ASPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY Free format text: PAYER NUMBER DE-ASSIGNED (ORIGINAL EVENT CODE: RMPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
FPAY | Fee payment |
Year of fee payment: 12 |
|
AS | Assignment |
Owner name: CIENA CORPORATION, MARYLAND Free format text: RELEASE BY SECURED PARTY;ASSIGNOR:DEUTSCHE BANK AG NEW YORK BRANCH;REEL/FRAME:050938/0389 Effective date: 20191028 |
|
AS | Assignment |
Owner name: BANK OF AMERICA, N.A., AS COLLATERAL AGENT, ILLINO Free format text: PATENT SECURITY AGREEMENT;ASSIGNOR:CIENA CORPORATION;REEL/FRAME:050969/0001 Effective date: 20191028 |
|
AS | Assignment |
Owner name: CIENA CORPORATION, MARYLAND Free format text: RELEASE BY SECURED PARTY;ASSIGNOR:BANK OF AMERICA, N.A.;REEL/FRAME:065630/0232 Effective date: 20231024 |