Patent Grant
11233523
This disclosure is related to the technical field of generating analog signals and digital-to-analog conversion.
“Ultra-Wideband Digital-to-Analog Conversion Technologies for Tbit/s channel transmission” by Yamazaki et al, presented at ECOC 2017, notes that “high-speed electronic digital-to-analog converters (DACs) are of key importance in modern optical transmission systems” and “in multilevel optical transmitters, the analog bandwidth of the DACs is one of the factors limiting the transmitter's bandwidth”. Yamazaki et al. describes a digital-preprocessed analog-multiplexed DAC (DP-AM-DAC) that uses a digital pre-processor, two sub-DACs, and an analog multiplexer (AMUX). “With sub-DACs with a bandwidth of ˜½fB, we can generate signals with a bandwidth of ˜fB as the output from the AMUX.” The AMUX is a heterojunction bipolar transistor (HBT) analog multiplexer (mux). FIG. 2 of Yamazaki et al. shows an interleaving method (type I) and a preprocessed spectrum method (type II) that reduces the switching frequency of the analog mux by a factor of two. However, type II is very sensitive to imperfections of the matching of the analog characteristics of the two inputs of the analog mux, as very large signal components need to be almost-perfectly cancelled.
“An 8-bit 100-GS/s Distributed DAC in 28-nm CMOS for Optical Communications” by Huang et al., IEEE Transactions on Microwave Theory and Techniques, vol. 63, no. 4 (April 2015), discloses a distributed structure to interleave together the outputs from two DACs. Huang et al. uses two interleaved NRZ (non-return-to-zero) DACs sampled at 90 degrees out of phase with respect to each other and summed up at the output stage. The interleaving is structured to invert one of the image spectra so that they are cancelled when summed. Again, there are very strong interference terms that are suppressed only with precise matching of the two halves of the analog circuit.
U.S. Pat. No. 8,693,876 to Krause et al. discloses the combining of two half-band signals from two DACs into a full-band signal, by shifting up the frequency of one of the half-band signals with a bipolar mixer. It is desirable to have the circuit implemented using lower energy technologies such as complementary metal-oxide-semiconductor (CMOS).
Packet and burst switches are known, where typically 1500 bytes received from one tributary are sent in sequence out of one optical or electrical output.
According to a broad aspect, an apparatus comprises circuitry configured to generate, from a stream of periodic digital samples, a first sub-stream of periodic analog samples and a second sub-stream of periodic analog samples, each sub-stream comprising substantially stable time intervals. The apparatus further comprises circuitry configured to generate a sign-modulated sub-stream by applying to the second sub-stream a sign modulation operation effecting a sign transition during a stable time interval of the second sub-stream. The apparatus further comprises circuitry configured to generate an output stream of periodic analog samples based on a sum of the first sub-stream and the sign-modulated sub-stream, wherein a period of the output stream is shorter than periods of the first and second sub-streams.
According to some examples, the sign modulation operation effects a plurality of sign transitions during a respective plurality of stable time intervals of the second sub-stream, each sign transition occurring at approximately the middle of a respective stable time interval.
According to some examples, the sign modulation operation effects a plurality of sign transitions during a respective plurality of stable time intervals of the second sub-stream, the plurality of sign transitions occurring at regular offset times within the plurality of stable time intervals.
According to some examples, each sample of the first sub-stream is generated from a sum of two of the digital samples, and each sample of the second sub-stream is generated from a difference between two of the digital samples.
According to some examples, the periods of the first and second sub-streams are equal.
According to some examples, the apparatus further comprises circuitry configured to generate one or more additional sub-streams of periodic analog samples from the stream of periodic digital samples, each additional sub-stream comprising substantially stable time intervals, circuitry configured to generate at least one additional sign-modulated sub-stream from the one or more additional sub-streams by effecting sign transitions during stable time intervals of at least one of the one or more additional sub-streams, and circuitry configured to generate the output stream as a function of the at least one additional sign-modulated sub-stream.
According to some examples, the apparatus further comprises circuitry configured to generate a plurality of input streams from the first sub-stream, the sign-modulated sub-stream, and the one or more additional sign-modulated sub-streams, and circuitry configured to generate the output stream by applying a multiplexing operation to the plurality of input streams.
According to some examples, the apparatus further comprises circuitry configured to generate the output stream from a sum of the first sub-stream, the sign-modulated sub-stream, and the at least one additional sign-modulated sub-stream.
According to some examples, the sign modulation operation comprises a multiplexing operation driven by a clock signal.
According to some examples, the sign modulation operation comprises multiplication by a substantially square clock signal.
A symbol source 124 is operative to generate a stream of symbols representing data to be transmitted in the optical signal 122. A digital signal processor (DSP) 126 is operative to process the symbols output from the symbol source 124, for example, performing one or more of pulse shaping, subcarrier multiplexing, chromatic dispersion pre-compensation, and distortion pre-compensation on the symbols. The DSP 126 is operative to generate I and Q digital drive signals 128, 129 for the X-polarization to be converted by DACs 130, 131, respectively, into I and Q analog drive signals 132, 133 for the X-polarization that, after amplification by respective amplifiers 134, 135, are used to drive the electrical-to-optical modulator 112. The DSP 126 is operative to generate I and Q digital drive signals 136, 137 for the Y-polarization to be converted by DACs 138, 139, respectively, into I and Q analog drive signals 140, 141 for the Y-polarization that, after amplification by respective amplifiers 142, 143, are used to drive the electrical-to-optical modulator 114.
Each of the DACs 130, 131, 138, 139 is operative to produce a high-bandwidth analog signal at a sample rate of FS. For example, the I and Q analog drive signals 132, 133 for the X-polarization may comprise an output stream of analog samples A={A0, A1, A2, A3, . . . } that contains one analog sample in each time period of duration ˜TS. The symbol “˜” is shorthand for the word “approximately”. The index i of each analog sample Ai represents an order of the analog samples in the stream A. The output stream of analog samples A={A0, A1, A2, A3, . . . } may be generated from a respective input stream of digital samples D={D0, D1, D2, D3, . . . }, where the index i of each digital sample Di represents an order of the digital samples in the stream D.
In some implementations, the DSP 126 and the DACs 130, 131, 138, 139 are comprised in a CMOS module, and the amplifiers 134, 135, 142, 143 are comprised in a BiCMOS module.
A clock signal 216 is provided to each of the sub-DAC components 212, 214, which are configured to sample the digital sub-streams 208, 210 at rising edges and falling edges of the clock signal 216, thus converting the digital sub-stream 208, 210 into respective analog sub-streams 218, 220. The analog sub-streams 218, 220 are input to the AMUX component 222, which is controlled by a clock signal 224.
A DAC with a traditional 2:1 multiplexer architecture, such as the DAC 200, requires a minimum clock frequency of ˜FS/2 in order to convert digital samples to analog samples having a sample period ˜TS. Although this 2:1 multiplexer architecture allows the required clock frequency to be reduced by half, it is still a challenge to generate a high-speed clock approaching hundreds of GHz and beyond. Furthermore, an analog 2:1 multiplexer component controlled by a high-speed clock consumes power that is proportional to the clock speed. There remains a need for DAC architectures and methods that are capable of achieving high-bandwidth and low-distortion digital-to-analog conversion using clock frequencies lower than FS/2.
The example DACs described in this document could be used as any of the DACs 130, 131, 138, 139 in the transmitter 100 or in some other electronic apparatus operative to perform high-bandwidth digital-to-analog conversion.
The DAC 300 comprises a pre-processing module 306, a first sub-DAC component 312, a second sub-DAC component 314, a sign modulation component 322, and a summation component 328.
The pre-processing module 306, which may be implemented in hardware, is operative to calculate, from terms of the input stream 302 of digital samples D, a first processed sub-stream 308 of digital samples DP1 and a second processed sub-stream 310 of digital samples DP2. In this example, each term of the first sub-stream 308 is equal to the sum of a pair of adjacent digital samples of the input stream 302, while each term of the second sub-stream 310 is equal to a difference between a pair of adjacent digital samples of the input stream 302. That is, as illustrated in the timing diagram of
A clock signal CLK1 316 operating at ¼FS (and therefore having a period of ˜4TS) is provided to each of the sub-DAC components 312, 314. The first sub-DAC component 312 samples the first sub-stream 308 at rising edges and falling edges of the clock signal 316, thus converting the first sub-stream 308 of digital samples DP1 into a first sub-stream 318 of analog samples AP1={(A0+A1), (A2+A3), (A4+A5), (A6+A7), . . . }. The second sub-DAC component 314 samples the second sub-stream 310 at rising edges and falling edges of the clock signal 316, thus converting the second sub-stream 310 of digital samples DP2 into a second sub-stream 320 of analog samples AP2={(A0−A1), (−A2+A3), (A4−A5), (−A6+A7), . . . }. The conversion of the processed sub-streams 308, 310 of digital samples to the respective sub-streams 318, 320 of analog samples is represented by the arrows 404. The clock signal CLK1 316 may alternate between amplitudes of +1 and −1, as illustrated in
The sign modulation component 322 applies a sign modulation operation to the second sub-stream 320 of analog samples AP2, thereby generating a sign-modulated sub-stream 326 of analog samples ASM2. The sign modulation component 322 is controlled by a clock signal CLK2 324 which operates at the same frequency as the clock signal CLK1 316 (i.e., ˜¼FS). However, the clock signal CLK2 324 operates at a phase offset of three quarters of a period (i.e., 270° or 3π/2 or 3TS) relative to the clock signal CLK1 316. The clock signal CLK2 324 may be configured to have an inherent non-zero phase offset relative to the clock signal CLK1 316. Alternatively, the effect of the non-zero phase offset may be achieved by applying an appropriate time delay to the second sub-stream 320 of analog samples AP2.
Different techniques are contemplated for implementing the sign modulation operations described throughout this document. According to some examples, sign modulation may be implemented using a multiplexing operation driven by a clock signal. According to other examples, where the clock signal is substantially square and alternates between positive and negative values, sign modulation may be achieved by multiplying a sub-stream of analog signals directly by the clock signal CLK2 324. For example, as illustrated in
Table 1 is helpful for understanding the operation of the DAC 300 and the timing diagram illustrated in
The effect of applying the sign modulation operation 322 to the second sub-stream 320 will now be considered. Referring to the time interval [0, TS), the second sub-stream 320 has a value of (A0−A1) and the clock signal CLK2 324 has an amplitude of +1. Therefore, the sign-modulated sub-stream 326 is equal to the product of (A0−A1) and (+1), such that the value of ASM2 during the time interval [0, TS) is (A0−A1). Referring to the next time interval [TS, 2TS), the second sub-stream 320 still has a value of (A0−A1), but the clock signal CLK2 324 has transitioned to an amplitude of −1. Therefore, the sign-modulated sub-stream 326 is equal to the product of (A0−A1) and (−1), such that the value of ASM2 during the time interval [TS, 2TS) is (−A0+A1). Referring to the time interval [2TS, 3TS), the second sub-stream 320 has transitioned to a value of (−A2+A3), while the clock signal CLK2 324 still has the amplitude of −1. Therefore, the sign-modulated sub-stream 326 is equal to the product of (−A2+A3) and (−1), such that the value of ASM2 during the time interval [2TS, 3TS) is (A2−A3).
Referring again to the time interval [0, TS), the first sub-stream 318 has a value of (A0+A1). When this value is added to the value of the sign-modulated sub-stream 326 during the same time interval, the result is (A0+A1)+(A0−A1)=2A0. During the next time interval [TS, 2TS), the first sub-stream 318 still has a value of (A0+A1). When this value is added to the value of the sign-modulated sub-stream 326 during the same time interval, the result is (A0+A1)+(−A0+A1)=2A1. During the next time interval [2TS, 3TS), the first sub-stream 318 has transitioned to a value of (A2+A3). When this value is added to the value of the sign-modulated sub-stream 326 during the same time interval, the result is (A2+A3)+(A2−A3)=2A2.
In general, it may be shown that the sum of the first sub-stream 318 of analog samples AP1 and the sign-modulated sub-stream 326 of analog samples ASM2 is equivalent to a stream of analog samples {2A0, 2A1, 2A2, 2A3, . . . } having a frequency ˜FS. Therefore, by using the summation component 328 to add the sub-streams 318 and 326 together, while accounting for the factor of two, it is possible to generate the output stream 304 of analog samples A={A0, A1, A2, A3, . . . }. According to some examples, the factor of two is accounted for in the digital hardware pre-processing of the signals 308 and 310. In other words, the samples of the output stream 304 may be substantially determined from a sum of the sub-streams 318 and 326. The term “substantially determined” is used to reflect the fact that the samples of the output stream 304 may be dependent on other factors such as distortion and/or noise due to imperfect circuits and inter-symbol interference (ISI) caused by analog filtering, peaking, hysteresis, reconstruction filtering, and parasitic circuit elements. The generation of the output stream 304 from the sub-streams 318 and 326 is represented by the arrows 406.
It is notable that the output stream 304 of analog samples A has a frequency of ˜FS, even though each one of the clock signals CLK1 316 and CLK2 324 only has a frequency of ˜¼ FS. The combination of the pre-processing 306, the sign modulation 322, and the summation 328 provides for a DAC technique that uses clock signals with lower frequencies than the analog signal that is ultimately output by the DAC. Since each value in the digital sub-streams 308, 310 is equal to the sum of (or difference between) two samples D of the digital stream 302, the dynamic range of sub-streams 308, 310 is reduced by 50% relative to a standard DAC that relies on a clock signal of frequency FS. Therefore, the reduction in clock frequency achievable with the DAC 300 may be balanced against signal quality requirements.
The DAC 500 comprises a pre-processing module 506, a first sub-DAC component 516, a second sub-DAC component 522, a third sub-DAC component 526, a fourth sub-DAC component 532, two sign modulation components 536, 540, two summation components 544, 546, and an analog multiplexer (AMUX) component 552.
The pre-processing module 506 is operative to calculate, from selected terms of the input stream 502 of digital samples D, a first processed sub-stream 508 of digital samples DP1, a second processed sub-stream 510 of digital samples DP2, a third processed sub-stream 512 of digital samples DP3, and a fourth processed sub-stream 514 of digital samples DP4. In this example, each term of the first sub-stream 508 and the third sub-stream 512 is calculated from a sum of a different respective pair of adjacent digital samples of the input stream 502, while each term of the second sub-stream 510 and the fourth sub-stream 514 is calculated from a difference between a different respective pair of adjacent digital samples of the input stream 502. For clarity, the input stream 502 of digital samples D is not shown in the timing diagram of
A clock signal CLK1 518 operating at ˜¼FS (and therefore having a period of ˜8TS) is provided to each of the sub-DAC components 516, 522. The clock signal CLK1 518 may alternate between positive and negative amplitudes, such as +1 and −1, as illustrated in
A clock signal CLK2 528 is provided to each of the sub-DAC components 526, 532. As illustrated in
The conversion of the processed sub-streams 508, 510, 512, 514 of digital samples to the respective sub-streams 520, 524, 530, 534 of analog samples is represented by the arrows 602.
The sign modulation component 536 applies a sign modulation operation to the second sub-stream 524 of analog samples AP2, thereby generating a sign-modulated sub-stream 538 of analog samples ASM2. The sign modulation component 536 is controlled by the clock signal CLK2 528. Where the clock signal CLK2 528 comprises a substantially square clock signal which alternates between positive and negative amplitudes, as illustrated in
The sign modulation component 540 applies a sign modulation operation to the fourth sub-stream 534 of analog samples AP4, thereby generating a sign-modulated sub-stream 542 of analog samples ASM4. As illustrated in
The generation of the sign-modulated sub-streams 538, 542 from the respective sub-streams 524, 534 is represented by the arrows 604.
As illustrated in
The streams 548, 550 may be provided to the AMUX component 552 which is controlled by the rising and falling edges of a clock signal CLK3 554 operating at ˜¼FS (and therefore having a period of ˜4TS). According to some examples, the clock signal CLK3 554 may alternate between positive and negative amplitudes, such as +1 and −1. As illustrated in
Table 2 is helpful for understanding the operation of the DAC 500 and the timing diagram illustrated in
A0 − A−1
A0 + A−1
Generation of the stream 548 of analog samples AJ will be considered first. Referring to the time interval [0, 2TS), the first sub-stream 520 of analog samples AP1 has a value of (A−1+A0) and the second sub-stream 524 of analog samples AP2 has a value of (A0−A1). During this time interval, the clock signal CLK2 528 has an amplitude of +1, so it follows that the sign-modulated sub-stream 538 is equal to the product of (A0−A−1) and (+1), such that the value of ASM2 during the time interval [0, 2TS) is (A0−A−1). Referring to the time interval [2TS, 4TS), the first sub-stream 520 of analog samples AP1 has transitioned to a value of (A3+A4) and the second sub-stream 524 of analog samples AP2 has transitioned to a value of (A3−A4). During this time interval, the clock signal CLK2 528 still has to the amplitude of +1, so it follows that the sign-modulated sub-stream 538 is equal to the product of (A3−A4) and (+1), such that the value of ASM2 during the time interval [2TS, 4TS) is (A3−A4). Referring to the time interval [4TS, 6TS), the first sub-stream 520 of analog samples AP2 still has a value of (A3+A4) and the second sub-stream AP2 still has the value of (A3−A4). However, during this time interval, the clock signal CLK2 528 has transitioned to an amplitude of −1, so it follows that the sign-modulated sub-stream 538 is equal to the product of (A3−A4) and (−1), such that the value of ASM2 during the time interval [4TS, 6TS) is (−A3+A4). Referring to the time interval [6TS, 8TS), the first sub-stream 520 of analog samples AP1 has transitioned to a value of (A7+A8) and the second sub-stream AP2 has transitioned to value of (A8−A7). During this time interval, the clock signal CLK2 528 still has the amplitude of −1, so it follows that the sign-modulated sub-stream 538 is equal to the product of (A8−A7) and (−1), such that the value of ASM2 during the time interval [6TS, 8TS) is (−A8+A7).
As illustrated in
Generation of the stream 550 of analog samples AK will now be considered. Referring to the time interval [0, 2TS), the third sub-stream 530 of analog samples AP3 has a value of (A1+A2) and the fourth sub-stream 534 of analog samples AP4 has a value of (A1−A2). During this time interval, the clock signal CLK1 518 has an amplitude of +1, so it follows that the sign-modulated sub-stream 542 is equal to the product of (A1−A2) and (+1), such that the value of ASM4 during the time interval [0, 2TS) is (A1−A2). Referring to the time interval [2TS, 4TS), the third sub-stream 530 of analog samples AP3 still has the value of (A1+A2) and the fourth sub-stream 534 of analog samples AP4 still has the value of (A1−A2). However, during this time interval, the clock signal CLK1 518 has transitioned to an amplitude of −1, so it follows that the sign-modulated sub-stream 542 is equal to the product of (A1−A2) and (−1), such that the value of ASM4 during the time interval [2TS, 4TS) is (−A1+A2). Referring to the time interval [4TS, 6TS), the third sub-stream 530 of analog samples AP3 has transitioned to a value of (A5+A6) and the fourth sub-stream AP4 has transitioned to a value of (A6−A5). During this time interval, the clock signal CLK1 518 still has the amplitude of −1, so it follows that the sign-modulated sub-stream 542 is equal to the product of (A6−A5) and (−1), such that the value of ASM4 during the time interval [4TS, 6TS) is (−A6+A5). Referring to the time interval [6TS, 8TS), the third sub-stream 530 of analog samples AP3 still has the value of (A5+A6) and the fourth sub-stream AP4 still has the value of (A6−A5). However, during this time interval, the clock signal CLK1 518 has transitioned to the amplitude of +1, so it follows that the sign-modulated sub-stream 542 is equal to the product of (A6−A5) and (+1), such that the value of ASM4 during the time interval [6TS, 8TS) is (A6−A5).
As illustrated in
The streams 548, 550 are provided to the AMUX component 552 which is controlled by the clock signal CLK3 554. As previously noted, each transition of the clock signal CLK3 554 (i.e., from +1 to −1, or from −1 to +1) occurs at approximately the middle or center of a respective stable time interval of the streams 548, 550. In this example, the AMUX component 552 is operative to output the stream 548 when the clock signal CLK3 554 has a value of +1 and to output the stream 550 when the clock signal CLK3 554 has a value of −1. As provided in Table 2 and as illustrated in
Notably, the output stream 504 of analog samples A has a frequency of s, even though each one of the clock signals CLK1 518 and CLK2 528 only has a frequency of ˜⅛FS, and the clock signal CLK3 554 only has a frequency of ˜¼FS. The combination of the pre-processing 506, the sign modulation operations 536, 540, the summation operations 544, 546, and the AMUX component 552 provides for a DAC technique that uses clock signals with lower frequencies than the analog signal that is ultimately output by the DAC. Since each value in the digital sub-streams 508, 510, 512, 514 is equal to the sum of (or difference between) two samples D of the digital stream 502, the dynamic range of sub-streams 508, 510, 512, 514 is reduced by 50% relative to a standard DAC that relies on a clock signal of frequency FS. Therefore, the reduction in clock frequency achievable with the DAC 500 must be balanced against signal quality requirements.
The DAC 700 comprises a pre-processing module 706, a first sub-DAC component 716, a second sub-DAC component 722, a third sub-DAC component 726, a fourth sub-DAC component 732, three sign modulation components 734, 740, 746, and a summation component 752.
The pre-processing module 706 is operative to calculate, from selected terms of the input stream 702 of digital samples D, a first processed sub-stream 708 of digital samples DP1, a second processed sub-stream 710 of digital samples DP2, a third processed sub-stream 712 of digital samples DP3, and a fourth processed sub-stream 714 of digital samples DP4. In this example, each term of the first sub-stream 708 is calculated from a sum of a different respective pair of adjacent digital samples of the input stream 702, while each term of the second sub-stream 710, the third sub-stream 712, and the fourth sub-stream 714 is calculated from a difference between a different respective pair of adjacent digital samples of the input stream 702. Specifically, for an input stream 702 of digital samples D={D−1, D0, D1, D2, D3, D4, . . . }, the first processed sub-stream 708 comprises the digital samples DP1={(D0+D3), (D4+D7), (D8+D11), . . . }; the second processed sub-stream 710 comprises the digital samples DP2={(D1−D2), (−D5+D6), (D7−D8), . . . }; the third processed sub-stream 712 comprises the digital samples DP3={(−D0+D1), (D4−D5), (−D8+D9), . . . }; and the fourth processed sub-stream 714 comprises the digital samples DP4={(D2−D3), (−D6+D7), (D10−D11), . . . }. Each term of D has a substantially stable time interval of ˜TS, while each term of DP1, DP2, DP3, and DP4 has a substantially stable time interval of ˜4TS. Thus, each processed sub-stream 708, 710, 712, 714 has a frequency that is one quarter of the frequency of the input stream 702.
A clock signal CLK1 718 operating at ˜⅛FS (and therefore having a period of ˜8TS) is provided to each of the sub-DAC components 716, 722, 726, 730. The clock signal CLK1 718 may alternate between positive and negative amplitudes, such as +1 and −1, as illustrated in
The sign modulation component 734 applies a sign modulation operation to the second sub-stream 724 of analog samples AP2, thereby generating a sign-modulated sub-stream 738 of analog samples ASM2. The sign modulation component 734 is controlled by a clock signal CLK2 736. As illustrated in
The sign modulation component 740 applies a sign modulation operation to the third sub-stream 728 of analog samples AP3, thereby generating a sign-modulated sub-stream 744 of analog samples ASM3. The sign modulation component 740 is controlled by a clock signal CLK3 742. As illustrated in
The sign modulation component 746 applies a sign modulation operation to the fourth sub-stream 732 of analog samples AP4, thereby generating a sign-modulated sub-stream 750 of analog samples ASM4. The sign modulation component 746 is controlled by a clock signal CLK4 748. As illustrated in
Table 3 is helpful for understanding the operation of the DAC 700 and the timing diagram illustrated in
Referring to the time interval [0, TS), the first sub-stream 720 of analog samples AP1 has a value of (A0+A3), the second sub-stream 724 of analog samples AP2 has a value of (A1−A2), the third sub-stream 728 of analog samples AP3 has a value of (−A0+A1), and the fourth sub-stream 732 of analog samples AP4 has a value of (A2−A3). During this time interval, the clock signal CLK2 736 has an amplitude of +1, so it follows that the sign-modulated sub-stream 738 is equal to the product of (A1−A2) and (+1), such that the value of ASM2 during the time interval [0, TS) is (A1−A2). During this time interval, the clock signal CLK3 740 has an amplitude of −1, so it follows that the sign-modulated sub-stream 742 is equal to the product of (−A0+A1) and (−1), such that the value of ASM3 during the time interval [0, TS) is (A0−A1). During this time interval, the clock signal CLK4 744 has an amplitude of +1, so it follows that the sign-modulated sub-stream 746 is equal to the product of (A2−A3) and (+1), such that the value of ASM4 during the time interval [0, TS) is (A2−A3). Adding together the values of the sub-streams 720, 738, 744, 750 during the time interval [0, TS) is equivalent to (A0+A3)+(A1−A2)+(A0−A1)+(A2−A3)=2A0.
Referring to the time interval [TS, 2TS), all of the sub-streams 720, 724, 728, 732 still have the same values as they did during the time interval [0, TS). Furthermore, the clock signal CLK2 736 and the clock signal CLK4 744 still have the same amplitudes as they did during the time interval [0, TS). Accordingly, the values of AP1, ASM2, and ASM4 remain the same as they were during the previous time interval [0, TS), namely, (A0+A3), (A1−A2), and (A2−A3), respectively. However, the clock signal CLK3 740 has transitioned from the amplitude of −1 to an amplitude of +1. Accordingly, the sign-modulated sub-stream 742 is equal to the product of (−A0+A1) and (+1), such that the value of ASM3 during the time interval [TS, 2TS) is (−A0+A1). Therefore, when the values of the sub-streams 720, 738, 744, 750 during the time interval [TS, 2TS) are added together, the result is equivalent to (A0+A3)+(A1−A2)+(−A0+A1)+(A2−A3)=2A1.
This same exercise may be repeated for each time interval of duration TS. In general, it may be shown that the sum of the sub-streams 720, 724, 728, 732 is equivalent to a stream of analog samples {2A0, 2A1, 2A2, 2A3, . . . } having a frequency ˜FS. Therefore, by using the summation component 752 to add the sub-streams 720, 724, 728, 732 together, while accounting for the factor of two, it is possible to generate the output stream 704 of analog samples A={A0, A1, A2, A3, . . . }. The generation of the output stream 704 from the sub-streams 720, 738, 744, 750 is represented by the arrows 804.
As in the previous examples, it is notable that the output stream 704 of analog samples A has a frequency of ˜FS, even though each one of the clock signals CLK1 718, CLK2 736, CLK3 742, and CLK4 748 only has a frequency of ˜⅛FS. The combination of the pre-processing 706, the sign modulation operations 734, 740, 746, and the summation operation 752 provides for a DAC technique that uses clock signals with lower frequencies than the analog signal that is ultimately output by the DAC. Since each value in the digital sub-streams 708, 710, 712, 714 is equal to the sum of (or difference between) two samples D of the digital stream 702, the dynamic range of sub-streams 708, 710, 712, 714 is reduced by 50% relative to a standard DAC that relies on a clock signal of frequency F. Therefore, the reduction in clock frequency achievable with the DAC 700 must be balanced against signal quality requirements.
The DAC 900 comprises a pre-processing module 906, eight sub-DAC components 916, 917, 918, 919, 920, 921, 922, 923, six sign modulation components 929, 930, 931, 943, 944, 945, two summation components 952, 954, and an AMUX component 956.
The pre-processing module 906 is operative to calculate, from selected terms of the input stream 902 of digital samples D, eight processed sub-streams 908, 909, 910, 911, 912, 913, 914, 915 of digital samples DP1, DP2, DP3, DP4, DP5, DP6, DP7, DP8, respectively. In this example, each term of the first sub-stream 908 and the fifth sub-stream 912 is calculated from a sum of a different respective pair of digital samples of the input stream 902, while each term of the remaining sub-streams 909, 910, 911 and 913, 914, 915 is calculated from a difference between a different respective pair of adjacent digital samples of the input stream 902. Specifically, for an input stream 902 of digital samples D={D−1, D0, D1, D2, D3, D4, . . . }, the first processed sub-stream 908 comprises the digital samples DP1={(D2+D5), (D10+D13), . . . }; the second processed sub-stream 909 comprises the digital samples DP2={(−D2+D3), (D10−D11), . . . }; the third processed sub-stream 910 comprises the digital samples DP3={(−D3+D4), (D11−D12), . . . }; the fourth processed sub-stream 911 comprises the digital samples DP4={(−D4+D5), (D12−D13), . . . }; the fifth processed sub-stream 912 comprises the digital samples DP5={(D−2+D0, (D6+D9), . . . }; the sixth processed sub-stream 913 comprises the digital samples DP6={D−2−D−1), (−D7+D8), . . . }; the seventh processed sub-stream 914 comprises the digital samples DP7={(D−1−D0), (−D7+D8), . . . }; and the eighth processed sub-stream 915 comprises the digital samples DP8={(D0−D0, (−D8+D9), . . . }. Each term of D has a substantially stable time interval of ˜TS, while each term of DP1 to DP8 has a substantially stable time interval of ˜8TS. Thus, each processed sub-stream 908, 909, 910, 911, 912, 913, 914, 915 has a frequency that is one eighth of the frequency of the input stream 902.
A clock signal CLK1 924 operating at ˜ 1/16FS (and therefore having a period of ˜16TS) is provided to each of the sub-DAC components 916, 917, 918, 919. The clock signal CLK1 924 may alternate between positive and negative amplitudes, such as +1 and −1. The first sub-DAC component 916 samples the first sub-stream 908 at rising edges and falling edges of the clock signal CLK1 924, thus converting the first sub-stream 908 of digital samples DP1 into a first sub-stream 925 of analog samples AP1={(A2+A5), (A10±A13), . . . }. In a similar manner, the sub-DAC components 917, 918, 919 sample the respective sub-streams 909, 910, 911 of digital samples at rising edges and falling edges of the clock signal CLK1 924, thus resulting in respective sub-streams 926, 927, 928 of analog samples, namely AP2={(−A2+A3), (A10−A11), . . . }, AP3={(−A3+A4), (A11−A12), . . . }, and AP4={(−A4+A5), (A12−A13), . . . }. Each of the sub-streams 925, 926, 927, 928 has a frequency of ˜ 1/16FS.
Another clock signal CLK5 938 is provided to each of the sub-DAC components 920, 921, 922, 923. The clock signal CLK5 938 operates at the same frequency as the clock signal CLK1 924 (i.e., ˜ 1/16FS), but with a relative phase offset of one quarter of a period (i.e., 90° or π/2 or 4TS). The clock signal CLK5 938 may alternate between positive and negative amplitudes, such as +1 and −1. The fifth sub-DAC component 920 samples the fifth sub-stream 912 at rising edges and falling edges of the clock signal CLK5 938, thus converting the fifth sub-stream 912 of digital samples DP5 into a fifth sub-stream 939 of analog samples AP5={(A−2+A1), (A6+A9), . . . }. In a similar manner, the sub-DAC components 921, 922, 923 sample the respective sub-streams 913, 914, 915 of digital samples at rising edges and falling edges of the clock signal CLK5 938, thus resulting in respective sub-streams 940, 941, 942 of analog samples, namely AP6={(A−2−A−1), (−A7+A8), . . . }, AP7={(A−1−A0), (−A7+A8), . . . }, and AP8={(A0−A1), (−A8+A9), . . . }. Each of the sub-streams 939-942 has a frequency of ˜ 1/16FS.
Table 4 is helpful for understanding the operation of the DAC 900.
The sign modulation component 929 applies a sign modulation operation to the second sub-stream 926 of analog samples AP2, thereby generating a sign-modulated sub-stream 935 of analog samples ASM2. The sign modulation component 929 is controlled by a clock signal CLK2 932. The clock signal CLK2 932 operates at the same frequency as the clock signal CLK1 924 (i.e., ˜ 1/16FS), but with a relative phase offset of 3/16 of a period (i.e., 67.5° or 3π/8 or 3TS). The transitions of the clock signal CLK2 932 (i.e., from +1 to −1, or from −1 to +1) occur at regular positions within the stable time intervals of the second sub-stream 926 of analog samples AP2. Specifically, in this example, each transition of the clock signal CLK2 932 occurs at approximately three eighths of the duration of a respective time interval of the second sub-stream 926.
The sign modulation component 930 applies a sign modulation operation to the third sub-stream 927 of analog samples AP3, thereby generating a sign-modulated sub-stream 936 of analog samples ASM3. The sign modulation component 930 is controlled by a clock signal CLK3 933. The clock signal CLK3 933 operates at the same frequency as the clock signal CLK1 924 (i.e., ˜ 1/16FS), but with a relative phase offset of one quarter of a period (i.e., 90° or π/2 or 4TS). Each transition of the clock signal CLK3 933 (i.e., from +1 to −1, or from −1 to +1) occurs at approximately the middle or center of a respective stable time interval of the third sub-stream 927 of analog samples AP3. As a result, the sign-modulated sub-stream 936 of analog samples ASM3 has double the frequency of the third sub-stream 927 of analog samples AP3, that is ˜⅛FS.
The sign modulation component 931 applies a sign modulation operation to the fourth sub-stream 928 of analog samples AP4, thereby generating a sign-modulated sub-stream 937 of analog samples ASM4. The sign modulation component 931 is controlled by a clock signal CLK4 934. The clock signal CLK4 934 operates at the same frequency as the clock signal CLK1 924 (i.e., ˜ 1/16FS), but with a relative phase offset of 5/16 of a period (i.e., 112.5° or 5π/8 or 5TS). The transitions of the clock signal CLK4 934 (i.e., from +1 to −1, or from −1 to +1) occur at regular positions within the stable time intervals of the fourth sub-stream 928 of analog samples AP4. Specifically, in this example, each transition of the clock signal CLK4 934 occurs at approximately five eighths of the duration of a respective time interval of the fourth sub-stream 928.
The sign modulation components 943, 944, 945 operate in a similar manner on the sub-streams 940, 941, 942, respectively, thereby generating the sign-modulated sub-streams 949, 950, 951. The sign modulation components 943, 944, 945 are controlled by clock signals 946, 947, 948, respectively. Each of the clock signals 946, 947, 948 operates at the same frequency as the clock signal CLK5 938 (i.e., ˜ 1/16FS), but with different respective phase shifts relative to the clock signal CLK5 938. Specifically, the clock signal CLK6 946 has a relative phase shift of 3π/8, such that each transition of the clock signal CLK6 946 occurs at approximately three eighths of the duration of a respective time interval of the sixth sub-stream 940 of analog samples AP6; the clock signal CLK7 947 has a relative phase shift of π/2, such that each transition of the clock signal CLK7 947 occurs at approximately the middle or center of a respective time interval of the seventh sub-stream 941 of analog samples AP7; and the clock signal CLK8 948 has a relative phase shift of 5π/8, such that each transition of the clock signal CLK8 948 occurs at approximately five eighths of the duration of a respective time interval of the eighth sub-stream 942 of analog samples AP8.
As illustrated in
In this example, the AMUX component 956 is operative to output the stream 953 when the clock signal CLK9 957 has a value of (+1) and to output the stream 955 when the clock signal CLK9 957 has a value of −1. As provided in Table 4, during the time interval [0, TS), the clock signal CLK9 957 has a value of −1, so the AMUX component 956 outputs the current value of the stream 955 during that time interval. The stream 955 of analog samples AK is determined from the summation of the sub-stream 939 of analog samples AP5, the sign-modulated sub-stream 949 of analog samples ASM6, the sign-modulated sub-stream 950 of analog samples ASM7, and the sign-modulated sub-stream 951 of analog samples ASM8. According to Table 4, during the time interval [0, TS), this summation is calculated as (A−2+A1)+(−A−2+A−1)+(−A−1+A0)+(A0−A1)=2A0. During the next time interval [TS, 2TS), the clock signal CLK9 957 still has a value of −1, so the AMUX component 956 outputs the current value of the stream 955. According to Table 4, during the time interval [TS, 2TS), the summation of AP5, ASM6, ASM7, ASM8 is calculated as (A−2+A1)+(−A−2+A−1)+(−A−1+A0)+(−A0+A1)=2A1. During the next time interval [2TS, 3TS), the clock signal CLK9 957 transitions to a value of (+1), so the AMUX component 956 outputs the current value of the stream 953 during that time interval. The stream 953 of analog samples AJ is determined by the summation of the sub-stream 925 of analog samples AP1, the sign-modulated sub-stream 935 of analog samples ASM2, the sign-modulated sub-stream 936 of analog samples ASM3, and the sign-modulated sub-stream 937 of analog samples ASM4. According to Table 4, during the time interval [2TS, 3TS), this summation is calculated as (A2+A5)+(A2−A3)+(A3−A4)+(A4−A5)=2A2. During the next time interval [3TS, 4TS), the clock signal CLK9 957 still has a value of (+1), so the AMUX component 956 outputs the current value of the stream 953. According to Table 4, during the time interval [3TS, 4TS), the summation of AP2, ASM2, ASM3, ASM4 is calculated as (A2+A5)+(−A2+A3)+(A3−A4)+(A4−A5)=2A3. Thus, when the factor of two is accounted for, the AMUX component 956 is configured to generate the stream 904 of analog samples A={A0, A1, A2, A3, . . . }.
The output stream 904 of analog samples A has a frequency of ˜FS, even though each one of the clock signals 924, 932, 933, 934, 938, 943, 944, 945 only has a frequency of ˜ 1/16FS, and the clock signal 957 only has a frequency of ˜⅛FS. The combination of the pre-processing 906, the sign modulation operations 929, 930, 931, 943, 944, 945, the summation operations 952, 954, and the AMUX component 956 provides for a DAC technique that uses clock signals with lower frequencies than the analog signal that is ultimately output by the DAC. Since each value in the digital sub-streams 908, 909, 910, 911, 912, 913, 914, 915 is equal to the sum of (or difference between) two samples D of the digital stream 902, the dynamic range of sub-streams 908, 909, 910, 911, 912, 913, 914, 915 is reduced by 50% relative to a standard DAC that relies on a clock signal of frequency FS. Therefore, the reduction in clock frequency achievable with the DAC 900 must be balanced against signal quality requirements.
The DAC 1000 comprises a pre-processing module 1006, eight sub-DAC components 1016, 1017, 1018, 1019, 1020, 1021, 1022, 1023, seven sign modulation components 1033, 1034, 1035, 1036, 1037, 1038, 1039, and a summation component 1054.
The pre-processing module 1006 is operative to calculate, from selected terms of the input stream 1002 of digital samples D, eight processed sub-streams 1008, 1009, 1010, 1011, 1012, 1013, 1014, 1015 of digital samples DP1, DP2, DP3, DP4, DP5, DP6, DP7, DP8, respectively. In this example, each term of the first sub-stream 1008 is calculated from a sum of a different respective pair of digital samples of the input stream 802, while each term of the remaining sub-streams 1009, 1010, 1011, 1012, 1013, 1014, 1015 is calculated from a difference between a different respective pair of adjacent digital samples of the input stream 1002. Specifically, for an input stream 1002 of digital samples D={D0, D1, D2, D3, D4, . . . }, the first processed sub-stream 1008 comprises the digital samples DP1={(D0+D7), (D8+D15), . . . }; the second processed sub-stream 1009 comprises the digital samples DP2={(−D0+D1), (D8−D9), . . . }; the third processed sub-stream 1010 comprises the digital samples DP3={(−D1+D2), (D9−D10), . . . }; the fourth processed sub-stream 1011 comprises the digital samples DP4={(−D2+D3), (D10−D11), . . . }; the fifth processed sub-stream 1012 comprises the digital samples DP5={(−D3+D4), (D11−D12), . . . }; the sixth processed sub-stream 1013 comprises the digital samples DP6={(−D4+D5), (D12−D13), . . . }; the seventh processed sub-stream 1014 comprises the digital samples DP7={(−D5+D6), (D13−D14), . . . }; and the eighth processed sub-stream 1015 comprises the digital samples DP8={(−D6+D7), (D14−D15), . . . }. Each term of D has a substantially stable time interval of ˜TS, while each term of DP1 to DP8 has a substantially stable time interval of −8TS. Thus, each processed sub-stream 1008, 1009, 1010, 1011, 1012, 1013, 1014, 1015 has a frequency that is one eighth of the frequency of the input stream 1002.
A clock signal CLK1 1024 operating at ˜ 1/16FS (and therefore having a period of ˜16TS) is provided to each of the sub-DAC components 1016, 1017, 1018, 1019, 1020, 1021, 1022, 1023. The clock signal CLK1 1024 may alternate between positive and negative amplitudes, such as +1 and −1. The first sub-DAC component 1016 samples the first sub-stream 1008 at rising edges and falling edges of the clock signal CLK1 1024, thus converting the first sub-stream 1008 of digital samples DP1 into a first sub-stream 1025 of analog samples AP1={(A0+A7), (A8+A15), . . . }. In a similar manner, the sub-DAC components 1017, 1018, 1019, 1020, 1021, 1022, 1023 sample the respective sub-streams 1009, 1010, 1011, 1012, 1013, 1014, 1015 of digital samples at rising edges and falling edges of the clock signal CLK1 1024, thus resulting in respective sub-streams 1026, 1027, 1028, 1029, 1030, 1031, 1032 of analog samples, namely AP2={(−A0+A1), (A8−A9), . . . }, AP3={(−A1±A2), (A9−A10), . . . }, AP4={(−A2+A3), (A10−A11), . . . }, AP5={(−A3+A4), (A11−A12), . . . }, AP6={(−A4+A5), (A12−A13), . . . }, AP7={(−A5+A6), (A13−A14), . . . }, and AP8={(−A6+A7), (A14−A15), . . . }. Each of the sub-streams 1025, 1026, 1027, 1028, 1029, 1030, 1031, 1032 has a frequency of ˜ 1/16FS.
Table 5 is helpful for understanding the operation of the DAC 1000.
A8 + A15
A9 − A10
A9 − A10
The sign modulation component 1033 applies a sign modulation operation to the second sub-stream 1026 of analog samples AP2, thereby generating a sign-modulated sub-stream 1047 of analog samples ASM2. The sign modulation component 1033 is controlled by a clock signal CLK2 1040. The clock signal CLK2 1040 operates at the same frequency as the clock signal CLK1 1024 (i.e., ˜ 1/16FS), but with a relative phase shift of 1/16 of a period (i.e., 22.5° π/8 or TS). The transitions of the clock signal CLK2 1040 (i.e., from +1 to −1, or from −1 to +1) occur at regular positions within the stable time intervals of the second sub-stream 1026 of analog samples AP2. Specifically, in this example, each transition of the clock signal CLK2 1040 occurs at approximately one eighth of the duration of a respective time interval of the second sub-stream 1026.
The sign modulation components 1034, 1035, 1036, 1037, 1038, 1039 operate in a similar manner on the sub-streams 1027, 1028, 1029, 1030, 1031, 1032, respectively, thereby generating the sign-modulated sub-streams 1048, 1049, 1050, 1051, 1052, 1053. The sign modulation components 1034, 1035, 1036, 1037, 1038, 1039 are controlled by clock signals 1041, 1042, 1043, 1044, 1045, 1046, respectively. Each of the clock signals 1041, 1042, 1043, 1044, 1045, 1046 operates at the same frequency as the clock signal CLK1 1024 (i.e., ˜ 1/16FS), but with different respective phase offsets relative to the clock signal CLK1 1024. Specifically, the clock signal CLK3 1041 has a relative phase offset of ⅛ of a period (i.e., 45° or π/8 or 2TS), such that each transition of the clock signal CLK3 1041 occurs at approximately one quarter of the duration of a respective time interval of the third sub-stream 1027 of analog samples AP3; the clock signal CLK4 1042 has a positive phase offset of 3/16 of a period (i.e., 67.5° or 3π/8 or 3TS), such that each transition of the clock signal CLK4 1042 occurs at approximately three eighths of the duration of a respective time interval of the fourth sub-stream 1028 of analog samples AP4; the clock signal CLK5 1043 has a relative phase offset of one quarter of a period (i.e., 90° or π/4 or 4TS), such that each transition of the clock signal CLK5 1043 occurs at approximately the middle or center of the duration of a respective time interval of the fifth sub-stream 1029 of analog samples AP5; the clock signal CLK6 1044 has a relative phase offset of 5/16 of a period (i.e., 112.5° or 5π/8 or 5TS), such that each transition of the clock signal CLK6 1044 occurs at approximately five eighths of the duration of a respective time interval of the sixth sub-stream 1030 of analog samples AP6; the clock signal CLK7 1045 has a relative phase offset of ⅜ of a period (i.e., 135° or 3π/4 or 6TS), such that each transition of the clock signal CLK7 1045 occurs at approximately three quarters of the duration of a respective time interval of the seventh sub-stream 1031 of analog samples AP7; and the clock signal CLK8 1046 has a relative phase offset of 7/16 of a period (i.e., 157.5° or 7π/8 or 7TS), such that each transition of the clock signal CLK8 1046 occurs at approximately seven eighths of the duration of a respective time interval of the eighth sub-stream 1032 of analog samples AP8.
In general, it may be shown that a summation of the sub-stream 1025 and the sign-modulated sub-streams 1047, 1048, 1049, 1050, 1052, 1053 results in a stream of analog samples {2A0, 2A1, 2A2, 2A3, . . . } having a frequency ˜FS. Therefore, by using the summation component 1054 to add the sub-streams 1025, 1047, 1048, 1049, 1050, 1051, 1052, 1053 together, while accounting for the factor of two, it is possible to generate the output stream 1004 of analog samples A={A0, A1, A2, A3, . . . }.
At 1102, a first sub-stream of periodic analog samples and a second sub-stream of periodic analog samples are generated from a stream of periodic digital samples. Each sub-stream comprises substantially stable time intervals. The term “periodic” is used to denote the uniform spacing of the samples in time. In other words, a stream (or sub-stream) of periodic analog (or digital) samples comprises samples that are uniformly spaced in time, such that the stream (or sub-stream) may be characterized by a period (or frequency).
According to some examples, each sample of the first sub-stream may be generated from a sum of two of the digital samples, which may be adjacent or non-adjacent. According to some examples, each sample of the second sub-stream may be generated from a difference between two of the digital samples, which may be adjacent or non-adjacent.
In one example, the stream of periodic digital samples comprises the stream 302 of digital samples D, the first sub-stream of periodic analog samples comprises the sub-stream 318 of analog samples AP1, and the second sub-stream of periodic analog samples comprises the sub-stream 320 of analog samples AP2. In this example, the generation at 1102 comprises applying the pre-processing 306 to the stream 302 to generate the sub-streams 308, 310, respectively, followed by the sampling of the sub-streams 308, 310 using the sub-DACs 312, 314, respectively.
In another example, the stream of periodic digital samples comprises the stream 502 of digital samples D, the first sub-stream of periodic analog samples comprises the sub-stream 520 of analog samples AP1, and the second sub-stream of periodic analog samples comprises the sub-stream 524 of analog samples AP2. Alternatively, where the stream of periodic digital samples comprises the stream 502 of digital samples D, the first sub-stream of periodic analog samples may comprise the sub-stream 530 of analog samples AP3, and the second sub-stream of periodic analog samples may comprise the sub-stream 534 of analog samples AP4. In these alternative examples, the generation at 1102 comprises applying the pre-processing 506 to the stream 502 to generate the sub-streams 508, 510 (or 512, 514), followed by the sampling of the sub-streams 508, 510 (or 512, 514) using the respective sub-DACs (i.e., either 516, 522 or 526, 532).
In yet another example, the stream of periodic digital samples comprises the stream 702 of digital samples D, the first sub-stream of periodic analog samples comprises the sub-stream 720 of analog samples AP2, and the second sub-stream of periodic analog samples comprises the sub-stream 724 of analog samples AP2, or the sub-stream 728 of analog samples AP3, or the sub-stream 732 of analog samples AP4. In this example, the generation at 1102 comprises applying the pre-processing 706 to the stream 702 to generate the sub-stream 708 and the sub-stream 710 (or 712 or 714), followed by the sampling of the sub-streams using the respective sub-DACs for those sub-streams.
In a further example, the stream of periodic digital samples comprises the stream 902 of digital samples D, the first sub-stream of periodic analog samples comprises the sub-stream 925 of analog samples AP1, and the second sub-stream of periodic analog samples comprises the sub-stream 926 of analog samples AP2, or the sub-stream 927 of analog samples AP3, or the sub-stream 928 of analog samples AP4. Alternatively, where the stream of periodic digital samples comprises the stream 902 of digital samples D, the first sub-stream of periodic analog samples may comprise the sub-stream 939 of analog samples AP5, and the second sub-stream of periodic analog samples may comprise the sub-stream 940 of analog samples AP6, or the sub-stream 941 of analog samples AP7, or the sub-stream 942 of analog samples AP8. In these alternative examples, the generation at 1102 comprises applying the pre-processing 906 to the stream 902 to generate the two sub-streams (e.g., sub-streams 908, 909), followed by the sampling of the sub-streams using the respective sub-DACs for those sub-streams.
In yet a further example, the stream of periodic digital samples comprises the stream 1002 of digital samples D, the first sub-stream of periodic analog samples comprises the sub-stream 1025 of analog samples AP1, and the second sub-stream of periodic analog samples comprises any one of the sub-streams 1026, 1027, 1028, 1029, 1030, 1031, 1032 of analog samples AP2-AP8, respectively. In this example, the generation at 1202 comprises applying the pre-processing 1006 to the stream 1002 to generate the sub-stream 1008 and any one of the sub-streams 1009, 1010, 1011, 1012, 1013, 1014, 1015, followed by the sampling of the sub-streams using the sub-DAC 1016 and any one of the sub-DACs 1017, 1018, 1019, 1020, 1021, 1022, 1023, respectively.
According to each of these examples, the periods of the first and second sub-streams generated at 1102 are equal (i.e., ˜2TS for DAC 300; ˜4TS for DACs 500 and 700; ˜8TS for DACs 900 and 1000). However, other examples are contemplated wherein the period of the first sub-stream may differ from the period of the second sub-stream.
At 1104, a sign-modulated sub-stream is generated by applying to the second sub-stream of periodic analog samples a sign modulation operation that effects or generates or causes or results in a sign transition during a stable time interval of the second sub-stream. As described previously, the sign modulation operation may comprise, for example, a multiplexing operation driven by a clock signal, or multiplication by a substantially square clock signal. Again, the term sign transition is used for clarity and in reference to a potentially advantageous implementation employing a +1/−1 clock. However, other clocks may be used to induce some other form of distinct transition.
In the example where the second sub-stream comprises the sub-stream 320, the sign-modulation operation comprises the operation 322 that results in the sign-modulated sub-stream 326 of analog samples ASM2. In the example where the second sub-stream comprises the sub-stream 524 (or 534), the sign-modulation operation comprises the operation 536 (or 540) that results in the sign-modulated sub-stream 538 (or 542). In the example where the second sub-stream comprises the sub-stream 724 (or 728 or 732), the sign-modulation operation comprises the operation 734 (or 740 or 746) that results in the sign-modulated sub-stream 738 (or 744 or 750). In the example where the second sub-stream comprises one of the sub-streams 926, 927, 928, 940, 941, 942, the sign-modulation operation comprises a respective one of the operations 929, 930, 931, 943, 944, 945 that results in a respective one of the sign-modulated sub-streams 935, 936, 937, 949, 950, 951. In the example where the second sub-stream comprises one of the sub-streams 1026, 1027, 1028, 1029, 1030, 1031, 1032, the sign-modulation operation comprises a respective one of the operations 1033, 1034, 1035, 1036, 1037, 1038, 1039 that results in a respective one of the sign-modulated sub-streams 1047, 1048, 1049, 1050, 1051, 1052, 1053.
As illustrated in the timing diagrams 400, 600, 800 and in Tables 1 through 5, the sign modulation operation applied at 1104 may effect (or generate or cause or result in) a plurality of sign transitions during a plurality of stable time intervals of the second sub-stream.
In some examples, each sign transition may occur at approximately the middle of a respective stable time interval. For example, the sign modulation operation 322 as controlled by the clock signal CLK2 324 causes a sign transition to occur at approximately the middle or center of each period of the sub-stream 320, such that the resulting sign-modulated sub-stream 326 has one half the period of the sub-stream 320 (or twice the frequency), as shown in the timing diagram 400 and in Table 1. In another example, the sign modulation operation 536 as controlled by the clock signal CLK2 528 causes a sign transition to occur at approximately the middle or center of each period of the sub-stream 524, such that the resulting sign-modulated sub-stream 538 has one half the period of the sub-stream 524 (or twice the frequency), as shown in the timing diagram 600 and in Table 2. In yet another example, the sign modulation operation 734 as controlled by the clock signal CLK2 736 causes a sign transition to occur at approximately the middle or center of each period of the sub-stream 724, such that the resulting sign-modulated sub-stream 738 has one half the period of the sub-stream 724 (or twice the frequency), as shown in the timing diagram 800 and in Table 3. Referring to Tables 4 and 5, it is apparent that the sign modulation operations 930, 944, and 1036 also cause sign transitions at approximately the middle of each period of the respective sub-streams to which they are applied.
In other examples, the sign modulation operations may cause sign transitions to occur at regular offset times within the plurality of stable time intervals. For example, as shown in the timing diagram 800 and in Table 3, the sign modulation operation 740 as controlled by the clock signal CLK3 724 causes a sign transition to occur within each stable time interval of the sub-stream 728 at a regular or consistent offset time of approximately TS relative to the beginning of the interval, while the sign modulation operation 746 as controlled by the clock signal CLK4 748 causes a sign transition to occur within each stable time interval of the sub-stream 732 at a regular or consistent offset time of approximately 3TS relative to the beginning of the interval. In another example, as shown in Table 4, the sign modulation operation 929 as controlled by the clock signal CLK2 932 causes a sign transition to occur within each stable time interval of the sub-stream 926 at a regular offset time of approximately 3TS relative to the beginning of the interval; the sign modulation operation 931 as controlled by the clock signal CLK4 934 causes a sign transition to occur within each stable time interval of the sub-stream 928 at a regular or consistent offset time of approximately 5TS relative to the beginning of the interval; the sign modulation operation 943 as controlled by the clock signal CLK6 946 causes a sign transition to occur within each stable time interval of the sub-stream 940 at a regular or consistent offset time of approximately 3TS relative to the beginning of the interval; and the sign modulation operation 945 as controlled by the clock signal CLK8 948 causes a sign transition to occur within each stable time interval of the sub-stream 942 at a regular offset time of approximately 5TS relative to the beginning of the interval. In yet another example, referring to Table 5, it is apparent that the sign modulation operations 1033, 1034, 1035, 1036, 1037, 1038, 1039 cause sign transitions to occur at regular offset times of approximately TS, 2TS, 3TS, 4TS, 5TS, 6TS, 7TS, respectively, relative to the beginning of each stable time interval of the respective sub-streams to which the operations are applied. In the case of the sign modulation operation 1036, the offset time of 4TS, coincides with the center of the time interval.
At 1106, an output stream of periodic analog samples is generated based on a sum of the first sub-stream and the sign-modulated sub-stream, wherein a period of the output stream is shorter than periods of the first and second sub-streams. For example, the stream 304 of analog samples A is generated based on a sum of the sub-stream 318 of analog samples AP1 and the sign-modulated sub-stream 326 of analog samples ASM2. The period of the stream 304 is TS, which is shorter than the period 2TS of the sub-stream 318 of analog samples AP2 and also shorter than the period 2TS of the sub-stream 320 of analog samples AP2. In another example, the stream 504 of analog samples A is equivalent to either (i) the stream 548 of samples AJ, which is based on a sum of the sub-stream 520 of analog samples AP1 and the sign-modulated sub-stream 538 of analog samples ASM2, or (ii) the stream 550 of samples AK, which is based on a sum of the sub-stream 530 of analog samples AP3 and the sign-modulated sub-stream 542 of analog samples ASM4. The period of the stream 504 is TS, which is shorter than the period 4TS of the sub-streams 520, 524, 530, 534 of analog samples AP2, AP2, AP3, AP4, respectively. In yet another example, the stream 704 of analog samples A is generated based on a sum of the sub-stream 318 of analog samples AP1 and the sign-modulated sub-stream 738 of analog samples ASM2 (as well as the other sign-modulated sub-streams 744 and 750). The period of the stream 704 is TS, which is shorter than the period 8TS of the sub-stream 720 of analog samples
AP1 and also shorter than the period 8TS of the sub-stream 724 of analog samples AP2. It may similarly be shown that the period of the output stream 904 is shorter than the periods of the sub-streams 925, 926, 927, 928, 939, 940, 941, 942, and also that the period of the stream 1004 is shorter than the periods of the sub-streams 1025, 1026, 1027, 1028, 1029, 1030, 1031, 1032.
According to some examples, the method 1100 may further comprise generating one or more additional sub-streams of periodic analog samples from the stream of periodic digital samples, each additional sub-stream comprising substantially stable time intervals; generating at least one additional sign-modulated sub-stream from the one or more additional sub-streams by effecting sign transitions during stable time intervals of at least one of the one or more additional sub-streams; and generating the output stream as a function of the at least one additional sign-modulated sub-stream.
According to one example, a plurality of input streams may be generated from at least the first sub-stream, the sign-modulated sub-stream, and the one or more additional sign-modulated sub-streams, and the output stream may be generated by applying a multiplexing operation to the plurality of input streams. For example, where the plurality of input streams comprise the streams 953 and 955 (which are generated, respectively, from the sub-streams 925, 935, 936, 937 and 939, 949, 950, 951), the output stream 904 is generated by applying the analog multiplexing operation 956 to the streams 953 and 955.
According to another example, the output stream may be generated from a sum of the first sub-stream, the sign-modulated sub-stream, and the at least one additional sign-modulated sub-stream. For example, where the first and second sub-streams generated at 1102 comprise the sub-streams 720 and 724, respectively, and where the sign-modulated sub-stream generated at 1104 comprises the sub-stream 738, the additional sign-modulated sub-streams may comprise the sub-streams 744 and 750. In this case, the output stream 704 is generated from a sum of first sub-stream 720, the sign-modulated sub-stream 738, and the additional sign-modulated sub-streams 744 and 750.
By employing the mechanisms described above with respect to
A potential advantage of the example DACs 300, 700, and 1000 is that the sub-DACs and sign modulation components are controlled by clock signals operating at a single frequency. The example DACs 500 and 900 rely on multiple clock signals operating at different frequencies, namely, one frequency to control the sub-DACs and sign modulation components, and another higher frequency to control the AMUX. However, this increased complexity may increase resilience to distortions as well as time mismatches between sub-DACs.
Alternative DAC architectures are contemplated. For example, an alternative implementation of DAC 500 may use a first AMUX component on the sub-streams 520, 530 and a second AMUX component on the sub-streams 524, 534, each AMUX component controlled by a clock signal of frequency ˜FS/4. A sign modulation operation, controlled by another clock signal of frequency ˜FS/4, may be applied to the sub-stream output by the second AMUX, thereby resulting in a sign-modulated sub-stream. A summation operation may then be used to combine the sign-modulated sub-stream with an output of the first AMUX component. It is noted that this alternative implementation of the DAC 500 would require the sign modulation operation to be controlled by a higher-frequency clock signal (i.e., ˜FS/4 rather than frequency ˜FS/8).
In a similar manner, an alternative implementation of DAC 900 may use four parallel AMUX components applied to the sub-streams 925, 939 and 926, 940 and 927, 941 and 928, 942, respectively, where each AMUX component is controlled by a clock signal of frequency ˜FS/8. In this case, three sign modulation operations, controlled by three other clocks of frequency ˜FS/8, may be applied to the sub-streams output by the latter three AMUX components, thereby resulting in three sign-modulated sub-streams. A summation operation may then be used to combine the three sign-modulated sub-streams with the output of the first AMUX component. It is noted that this alternative implementation of the DAC 900 would require the sign modulation operations to be controlled by higher-frequency clock signals (i.e., ˜FS/8 rather than frequency ˜FS/16).
In general, the DAC architectures and methods described throughout this document are merely examples, and should not be considered necessarily limiting. Labels such as “first”, “second”, “third”, etc., as well as subscripts such as “1”, “2”, “3”, etc., are used merely for ease of explanation. It is also noted that the sub-streams generated by the pre-processing operations 306, 506, 706, 906, 1006 are merely examples of the various sub-streams that might be generated from the digital samples of the streams 302, 502, 702, 902, 1002, respectively. Any given sample of a pre-processed sub-stream of digital samples may be generated from any two (or more) samples of the original digital signal provided that, in combination with the various sub-DACs, sign-modulation components, summation operations and/or AMUX components, the DAC is operative to convert the original digital signal to an output analog signal using clock frequencies that are lower than the frequency of the analog signal.
Linear digital filtering of the sub-streams of digital samples may adapt the signals in each sub-stream to obtain a cleaner output from the DAC. This may become more important when the DAC is physically further away from the sources of the sub-streams. The linear digital filtering may be calibrated in the factory. Alternatively, local or remote feedback may be used to dynamically control the linear digital filtering.
Nonlinear compensation may be included in the generation of the sub-streams, for example, as described in U.S. Pat. No. 6,781,537 to Taraschuk et al., without memory, or with memory (time delays) in the response. This nonlinear compensation may compensate for nonlinearity in the DACs or in downstream elements.
The techniques described in this document may be used to convert integer sub-streams into a voltage stream, in CMOS. However, other instantiations may be used. For example, current sub-streams may be converted to an optical E-Field stream, as was described in U.S. Pat. No. 7,277,603 to Roberts et al. An integer sub-stream may be combined with a voltage sub-stream to produce a voltage stream. The analog characteristic of the stream that is being created may be an optical or electrical phase, or other modulation of an input analog signal.
The scope of the claims should not be limited by the details set forth in the examples, but should be given the broadest interpretation consistent with the description as a whole.
| Number | Name | Date | Kind |
|---|---|---|---|
| 6781537 | Taraschuk et al. | Aug 2004 | B1 |
| 7277603 | Roberts et al. | Oct 2007 | B1 |
| 8237595 | Petrovic | Aug 2012 | B2 |
| 8693876 | Krause et al. | Apr 2014 | B2 |
| 10374623 | Oveis Gharan et al. | Aug 2019 | B1 |
| 10666285 | Mehdizad Taleie | May 2020 | B1 |
| Entry |
|---|
| Huang., et al., “An 8-bit 100-GS/s Distributed DAC in 28-nm CMOS for Optical Communications”, IEEE Transactions an Microwave Theory and Techniques, Apr. 2015, vol. 63, issue. 4, 8 pages. |
| Nagatani et al., “A 50-GHz-Bandwidth InP-HBT Analog-MUX Module for High-Symbol-Rate Optical Communications Systems”, 2016 IEEE MTT-S International Microwave Symposium (IMS). |
| Yamazaki et al., “160-Gbps Nyquist PAM4 Transmitter Using a Digital-Preprocessed Analog-Multiplexed DAC”, ECOC 2015. |
| Yamazaki H., et al., “Ultra-Wideband Digital-to-Analog Conversion Technologies for Tbit/s channel transmission”, European Conference on Optical Communication (ECOC), Sep. 2017, 3 pages. |
