Patent Grant
11592786
The instant patent application is related to and claims priority from the co-pending India provisional patent application entitled, “Time-to-Digital Converter (TDC) Architecture for Measuring Phase Differences Among Multiple Clocks”, Serial No.: 202141044473, Filed: 30 Sep. 2021, which is incorporated in its entirety herewith to the extent not inconsistent with the description herein.
The instant application is related to co-pending US Patent Application No: UNASSIGNED, Entitled, “Counter Design for a Time-To-Digital Converter (TDC)”, inventors Debasish Behera, et al, Filed: On even date herewith; Ser. No. 17/662,662, which is incorporated in its entirety herewith.
The instant application is related to co-pending US Patent Application No: UNASSIGNED, Entitled, “Time-to-digital Converter (TDC) Measuring Phase difference between Periodic Inputs with Jitter”, inventors Debasish Behera, et al, Filed: On even date herewith; Ser. No. 17/662,669, which is incorporated in its entirety herewith.
Embodiments of the present disclosure relate generally to time-to-digital converters (TDCs), and more specifically to TDCs measuring phase difference between periodic inputs.
TDCs are generally used to measure a time interval between a pair of events. The measurement is often quantified as a corresponding digital value. The events are typically voltage or logic-level changes of a corresponding pair of signals. TDCs find applications in electronic devices such as phase locked loops (PLL), timing cards, ToF (time-of-flight) measuring devices such as in radiation detectors, diagnostic equipment, etc.
TDCs often receive periodic inputs and are required to generate successive digital values representing the phase difference at corresponding time instances. Each digital value represents the time difference between corresponding edges (of the same direction) of the two periodic inputs. Examples of periodic signals include clocks, etc.
Aspects of the present disclosure are directed to such TDCs.
Example embodiments of the present disclosure will be described with reference to the accompanying drawings briefly described below.
In the drawings, like reference numbers generally indicate identical, functionally similar, and/or structurally similar elements. The drawing in which an element first appears is indicated by the leftmost digit(s) in the corresponding reference number.
A time-to-digital converter (TDC) provided according to an aspect of the present disclosure includes a count logic and a digital core. The count logic generates a first sequence of counts and a second sequence of counts respectively for a first periodic signal and a second periodic signal received as input signals. The first sequence of counts represents respective time instances on a time scale at which a first sequence of edges with a first direction of the first periodic signal occur, and the second sequence of counts represents respective time instances on the same time scale at which a second sequence of edges with the first direction of the second periodic signal occur.
The digital core generates a sequence of outputs representing the phase differences between the first periodic signal and the second periodic signal from the first sequence of counts and the second sequence of counts. Each output of the sequence of outputs is generated from a pair of successive edges of the first direction of one of the periodic signals and an individual one of the other periodic signal occurring between the pair, and the output is set equal to the minimum of difference of the individual one with the first value of the pair and the individual one with the second value of the pair.
By setting each output to such a minimum difference, the TDC may be particularly suited for environments which seek to obtain a steady state quickly.
According to an aspect, the TDC is designed to operate with the first periodic signal and the second periodic signal in durations when the signals have the same frequency, in addition to durations when the first periodic signal has a frequency of N times the frequency of the second periodic signal, where N is greater than or equal to 2 (integer relationship).
When such integer relationship is present, the pair of successive edges noted above are of the second periodic signal and the individual one is of the first periodic signal. Multiple clock cycles of the first periodic signal are present between the pair of successive edges, wherein the individual one is the earliest or last edge of the multiple clock cycles which is closer in time to either of the pair of successive edges.
Several aspects of the present disclosure are described below with reference to examples for illustration. However, one skilled in the relevant art will recognize that the disclosure can be practiced without one or more of the specific details or with other methods, components, materials and so forth. In other instances, well known structures, materials, or operations are not shown in detail to avoid obscuring the features of the disclosure. Furthermore, the features/aspects described can be practiced in various combinations, though only some of the combinations are described herein for conciseness.
Count logic 105 internally contains counters for each of input clocks clk_in_1 (101-1) and clk_in_2 (101-2). Count logic 105 is connected to receive input clocks clk_in_1 (101-1) and clk_in_2 (101-2). Clk_meas (103) represents a measurement clock applied to the counters in count logic 105. In an embodiment, each of the counters is designed to increment by one count at each cycle (e.g., at each rising edge) of clk_meas (103). However, in alternative embodiments, other conventions for counting, such as decrementing the counters by 1 or some other fixed value for each clock cycle of clk_meas (103), may be employed, as will be apparent to a skilled practitioner. The frequency of clk_meas (103) at least partially or substantially determines the resolution with which the time difference between two events (such as, for example, rising edges of the input clocks) can be generated.
Count logic 105 is shown as generating count_1 (104-1) and count_2 (104-2) respectively for input clocks clk_in_1 (101-1) and clk_in_2 (101-2). Each sequence of counts in count_1 (104-1) and count_2 (104-2) represents respective time instances on a same time scale at which edges of a pre-specified direction (rising or falling) of corresponding input clocks occur. The term ‘same time scale’ implies that each count represents a respective relative time duration from a common start time. For example, counters for each input clock start counting (e.g., from zero) at the same time (e.g., upon reset) and with identical frequency. Count logic 105 can be implemented in a known way. In an embodiment of the present disclosure, counts generated by counters in count logic 105 are 36-bits wide, and the time resolution of the counters is 62.5 pico-seconds (ps).
In an embodiment, count logic 105 is implemented as described in U.S. application Ser. No. UNASSIGNED, filed: on even date herewith, entitled “Counter Design for a Time-To-Digital Converter (TDC)”, and assigned to the assignee of the present application, the contents of which are incorporated herein by reference in its entirety.
Digital core 110 is connected to receive count_1 (104-1) and count_2 (104-2), and generates output digital values representing the phase difference between input clock signals (101-1 and 101-2). The output digital values are provided in sequence on path 115.
Several features of the present disclosure would be better understood and appreciated when the manner in which a TDC measures phase difference for a pair of periodic inputs is illustrated. Accordingly, the manner in which a TDC measures phase difference is briefly described next with respect to
The rising edges of clk_in_1 (101-1) are shown to be occurring at time instants t10, t12, t14 and t16, while the rising edges of clk_in_2 (101-2) are shown to be occurring at time instants t11, t13, t15 and t18. The time-period of clk_in_1 (101-1) is depicted as T1 (time interval between two consecutive rising edges).
Since digital core 110 receives a sequence of counts corresponding to the rising edges of clk_in_1 (101-1) and clk_in_2 (101-2), digital core 110 would generate two possible differences between the clock edges of clk_in_1 and clk_in_2, namely Pmin (representing for example, the smaller time interval t10-t11) and Pmax (representing, for example, the larger time interval t11-t12).
According to an aspect of the present disclosure, Pmin (rather than Pmax), i.e., the smaller of the two differences, is provided as the (correct) phase difference. By providing Pmin as the phase difference, circuits (e.g., a PLL) that rely on processing of the phase difference may attain a steady state sooner, in addition to reducing jittery output.
According to another aspect, digital core 110 computes both Pmin and Pmax, and then identifies the smaller of Pmin and Pmax as the phase difference between a pair of input clocks. Such an approach avoids the need for the knowledge of the exact frequency of the input clocks, which can otherwise be used to determine Pmin from Pmax alone (or Pmax from Pmin alone).
Although the timing diagram of
Further, although the timing diagram of
TDC 100 is noted above as operating on only a pair of periodic inputs. However, a TDC can be implemented to operate simultaneously on more than one pair of periodic inputs. In such cases, multiple phase differences (between a respective pair of input clocks) may be measured in parallel and provided as corresponding outputs by the TDC. Accordingly, the implementation details of a digital core contained in such a TDC in an embodiment of the present disclosure are briefly provided next.
As will be clear from the description, the phase difference may be measured between any one or multiple pairs of input clock signals. The output digital values are provided in sequence on combination of output paths 165-1 (accum_valid_data) and 165-3 (running_acc_data) in case of accumulation modes, and on path 115 (single_shot_data) in case of single-shot modes, as described below in further detail.
In an embodiment of the present disclosure, each count values (counts) 104-1 through 104-N is 36-bits wide, and the respective blocks in digital core 110 are designed with internal paths and blocks (e.g., registers) having widths that can handle the bit-width of the counts.
Clk_dig_tdc (107) is generated internal to count logic 105, and is provided as a master clock to the corresponding components of digital core 110, as also described below. In an embodiment of the present disclosure, the frequency of clk_dig_tdc (107) is greater than the maximum frequency allowed for input clocks clk_in_1 (101-1) through clk_in_N (101-N) by a factor of at least two. All sequential blocks of digital core 110 are designed to operate synchronously based on a single master clock clk_dig_tdc (107). Clk_dig_tdc_sel (108) is generated by digital core 110, and is received by count logic 105 (not shown). Clk_dig_tdc_sel (108) is used to select the frequency of clk_dig_tdc (107) for the digital core.
Referring to
The specific blocks and interconnections of
Register block 109 receives and stores count_1 (104-1) through count_N (104-N) in respective internal registers, and generates P copies of count_1 (104-1) through count_N (104-N) with the timing of storing controlled by clk_dig_tdc 107. Register block 109 forwards the P copies of count_1 (104-1) through count_N (104-N) according to clk_dig_tdc (107) as respective outputs.
Each of MUX 125-1 through 125-2P is an N:1 multiplexer, receives input counts count_1 (104-1) through count_N (104-N), and a respective select signal sel_1a (120-1), sel_1b (120-2) through sel Pa (120-2P-1), sel_Pb (120-2P). Thus, each of MUX 125 forwards one of the N counts as respective outputs (MUX outputs/selected counts) on path 128-1 through 128-2P based on the logic value of select signal 120. Select signals sel_1a (120-1) through sel_Pb (120-2P) may be set by a user (via corresponding means not shown) to indicate a pair of count-1 to count-N to be used as input clocks to respective calculation unit 150.
For example, MUX 125-1 may forward one of count_1 through count_N on path 128-1 to calculation unit 1 (150-1) based on logic value of select signal sel_1a (120-1), and MUX 125-2 may forward one of count_1 through count_N on path 128-2 to calculation unit 1 (150-1) based on logic value of select signal sel_1b (120-2). Thus, a corresponding pair of counts may be provided as inputs to each calculation unit 150 for determining the phase difference between the corresponding inputs.
Each calculation unit 150 receives a corresponding pair of counts as inputs and generates respective outputs based on mode of operation of TDC 100. Thus, calculation unit 150-1 is shown to be generating outputs accum_valid_data_1 (155-1-1), data_valid_1 (155-1-2), running_acc_data_1 (155-1-3), running_nsample_1 (155-1-4) and single_shot_data_1 (155-1-5), while calculation unit 150-P is shown to be generating outputs accum_valid_data_P (155-1-P), data_valid_P (155-P-2), running_acc_data_P (155-P-3), running_nsample_P (155-P-4) and single_shot_data_P (155-P-5). Each of single_shot_data_1 through single_shot_data_P provides a corresponding sequence of phase differences between the corresponding pairs of input clocks to the calculation units, and corresponds to single_shot_data (115) in the case of only a pair of input clocks as shown in
MUX 160 receives outputs from calculation units 150 and a select signal pair_sel (159). Pair_sel (159) may be set by a user (via corresponding means not shown) to indicate one of calculation units 150-1 through 150-P whose outputs may be provided as outputs of TDC 100. Outputs accum_valid_data (155-1), data_valid (155-2), running_acc_data (155-3), running_nsample (155-4) and single_shot_data (115) thus respectively correspond to the outputs of calculation unit 150 that is selected based on pair_sel (159). Pair_sel (159) can cycle through P values so as to enable the outputs of each calculation unit to be provided as outputs of TDC 100 substantially in parallel. The various modes of operation noted above are now described.
In an embodiment, TDC 100 is designed to operate in the following modes:
1. Accumulator mode, with input clocks clk_in_1 (101-1) through clk_in_N (101-N) having the same frequency.
2. Accumulator mode, with frequency of input clocks clk_in_1 (101-1) through clk_in_N (101-N) being integer-related.
3. Continuous single-shot mode
4. One-time single-shot mode
In modes (1) and (2), calculation unit 150 operates on periodic inputs (such as clk_in_1 (101-1) through clk_in_N (101-N)) and receives (from user via corresponding means not shown) a pre-determined value indicating the number of phase difference samples to be accumulated. In these two modes, TDC 100 generates outputs on paths 155-1 (accum_valid_data), 155-2 (data_valid), 155-3 (running_acc_data) and 155-4 (running_nsample). At the end of sampling every pre-determined number of samples, calculation unit 150 asserts data_valid (155-2) and provides the accumulated phase difference value on path 155-1 (accum_valid_data). At the end of accumulating each phase difference sample, calculation unit 150 provides the phase difference value accumulated thus far on path 155-3 (running_acc_data). Running_nsample (155-4) indicates the number of samples accumulated in running_acc_data (155-3). An average of the accumulated values (total accumulated value divided by the number of samples accumulated) may be determined by a component external to TDC 100.
In mode (3), calculation unit 150 operates on periodic inputs (such as clk_in_1 (101-1) through clk_in_N (101-N)) and generates phase difference between each pair of the periodic inputs (without accumulation of phase difference samples as in mode 1 and 2), and provides the sequence of phase differences on path 115 (single_shot_data).
In mode (4), calculation unit 150 operates on pulse inputs, generates a one-time phase difference between the pulses, and provides the phase difference on path 115 (single_shot_data).
A calculation unit in
The manner in which digital core 110 of
Input analyzer block 250 is shown as receiving inputs count_1_1 on path 128-1, count_1_2 on path 128-2 and clk_dig_tdc on path 107. Input analyzer block 250 generates signals window_identifier on path 232, count1_sub on path 242-1 and count2_sub on path 242-2. Input analyzer block 250 receives count_1_1 (128-1) and count_1_2 (128-2) at every rising edge of the corresponding input clock. Input analyzer block 250 may store count_1_1 (128-1) and count_1_2 (128-2) values, and additionally perform operations such as handling rollover of counters generating count_1_1 (128-1) and count_1_2 (128-2), as is described below in detail. In the absence of such additional operations, count1_sub (128-1) and count2_sub (128-2) are functionally equivalent to count_1_1 (128-1) and count_1_2 (128-2) respectively.
Input analyzer 250, among other operations, operates to determine the boundaries of windows (respective time instances) corresponding to (or defined by) a pair of successive edges of either one of the two input clocks clk_in_1 (101-1) and clk_in_2 (101-2) when both clocks have the same frequency, and of the slower clock when the clocks have frequencies that are integer multiples greater than 1. Input analyzer 250 indicates the boundaries (the left and right edges) of a window using signal window_identifier (232), as will be described in detail below.
Although the description is continued with respect to rising edges of input clocks, aspects of the present disclosure are equally applicable to falling edges of input clocks, as will be apparent to a skilled practitioner by reading the disclosure herein.
Output generator block 290 is shown as receiving window_identifier on path 232, count1_sub on path 242-1, count2_sub on path 242-2 and clk_dig_tdc on path 107. Output generator block 290 generates a sequence of phase differences on path 115 (single_shot_data) upon the end of each window defined by window_identifer (232). The phase differences in the sequence may all be substantially equal unless the phase difference between the two input clocks were to change.
The details of the operations performed by input analyzer block 250 and Output generator block 290 of digital core 110 (or calculation unit 150 in general) in determining the phase difference between a pair of input clocks are described in detail next.
As noted above, the frequency of clk_dig_tdc (107) is greater than the frequency of clk_in_1 (101-1) and clk_in_2 (101-2), and the values of count_1_1 (128-1) and count_1_2 (128-2) are received at every rising edge of respective clocks. Thus, the values of count_1_1 (128-1) and count_1_2 (128-2) can remain unchanged for two or more cycles of clk_dig_tdc (107). Accordingly, a change in respective values of count_1_1 (128-1) and count_1_2 (128-2) may be inspected for determining phase difference between clk_in_1 (101-1) and clk_in_2 (101-2). Blocks count_1_q-generator 305-1, count_2_q-generator 305-2, unique_1-generator 310-1, unique_2-generator 310-2 and window-identifier-block 320 operate to provide such functionality.
Count_1_q-generator 305-1 receives count_1_1 on path 128-1 and generates count1_q on path 306-1. In an embodiment, count1_q generator is implemented as a positive edge triggered flip-flop. Count1_q generator 305-1 is clocked by clk_dig_tdc (107). Count_1_q (306-1) represents a time-delayed version of count_1_1 (128-1). Accordingly, signal count_1_1 (128-1) is delayed by one clock cycle of clk_dig_tdc (107) before being output on path 306-1.
Count_2_q-generator 305-2 receives count_1_2 on path 128-2 and generates count2_q on path 306-2. In an embodiment, count2_q generator is implemented as a positive edge triggered flip-flop. Count2_q generator 305-2 is clocked by clk_dig_tdc (107). Count_2_q (306-2) represents a time-delayed version of count_1_2 (128-2). Accordingly, signal count_1_2 (128-2) is delayed by one clock cycle of clk_dig_tdc (107) before being output on path 306-2.
Unique_1-generator 310-1 receives count_1_1 (128-1), count1_q (306-1) and window_identifier (232), and generates unique_count1_q on path 311-1. As may be observed from
Unique_1-generator 310-1 generates a logic LOW on path 311-1 when count_1_1 (128-1) is equal to count1_q (306-1) and window_identifier (232) is a logic HIGH, and a logic HIGH when count_1_1 (128-1) is not equal to count1_q (306-1). In other words, unique_1-generator 310-1 operates to determine that the count value of count_1_1 (128-1) has undergone a change (indicating the occurrence of a rising edge of clk_in_1, 101-1) from its value in a previous clock cycle of clk_dig_tdc (107), and also additionally takes into account the value of window_identifier (232).
Unique_2-generator 310-2 receives count_1_2 (128-2), count2_q (306-2) and window_identifier (232), and generates unique_count2_q on path 311-2. As may be observed from
Unique_2-generator 310-2 generates a logic LOW on path 311-2 when count_1_2 (128-2) is equal to count2_q (306-2) and window_identifier (232) is a logic HIGH, and a logic HIGH when count_1_2 (128-2) is not equal to count2_q (306-2). In other words, unique_2-generator 310-2 operates to determine that the count value of count_1_2 (128-2) has undergone a change (indicating the occurrence of a rising edge of clk_in_2, 101-2) from its value in a previous clock cycle of clk_dig_tdc (107), and additionally takes into account the value of window_identifier (232).
Digital core 110 receives a continuous stream of counts representing corresponding edges of input clocks (clk_in_1, 101-1 and clk_in_2, 101-2). Digital core 110 generates each output of the sequence of outputs (noted above) from a pair of successive edges (of a same direction, rising/falling) of one input clock (e.g., clk_in_1, 101-1) and an individual one of the other input clock (e.g., clk_in_2, 101-2). Digital core 110, therefore, needs to identify a respective pair of successive edges for generating each output. Window-identifier-block 320, implemented in input analyzer block 250, performs such an operation.
Window-identifier-block 320 operates to identify the window boundaries noted above. That is, window-identifier-block 320 operates to identify time instants of occurrence of a pair of successive edges of the same direction (rising or falling) of the slower input clock when the frequencies of the two input clocks are integer multiples greater than one, or a pair of successive edges of the same direction (rising or falling) of either of the two input clocks when the frequencies of the two input clocks are equal. The duration between the time instants is referred to herein as a window. Each window may have a respective stream of differences that need to be examined in order to generate an output (representing the phase difference between input clocks, clk_in_1, 101-1 and clk_in_2, 101-2) upon the end of each window.
Window-identifier-block 320 receives unique_count1_q (311-1) and unique_count2_q (311-2), and generates signal window_identifier on path 232. Window-identifier-block 320 generates a logic HIGH on path 232 when both unique_count1_q (311-1) and unique_count2_q (311-2) are logic HIGH, and a logic LOW otherwise. As will become apparent in the description below, a pair of successive logic HIGHs represents the window boundaries noted above.
Rollover1-generator 315-1 receives count_1_1 (128-1), count1_q (306-1), window_identifier (232) and generates signal rollover_1 on path 316-1. In an embodiment of the present disclosure, rollover1-generator 315-1 is implemented as a sequential block, clocked by clk_dig_tdc (107). Rollover1-generator 315-1 generates a logic HIGH on path 316-1 when a count value of count_1_1 (128-1) has exceeded the full-scale value (rolled over) of the counter. As is well known in the relevant arts, a rollover is said to occur when the value of a counter reaches its full-scale value, and the counter starts counting from zero. For example, for a 10-bit counter, a rollover is said to occur when the value of count reaches 1024.
Rollover2-generator 315-2 receives count_1_2 (128-2), count2_q (306-2), window_identifier (232) and generates signal rollover_2 on path 316-2. In an embodiment of the present disclosure, rollover2-generator 315-2 is implemented as a sequential block, clocked by clk_dig_tdc (107). Rollover2-generator 315-2 generates a logic HIGH on path 316-2 when a count value of count_1_2 (128-2) has exceeded the full-scale value (rolled over) of the counter.
Rollover block 325 receives count_1_1 (128-1), count_1_2 (128-2), rollover_1 (316-1) and rollover_2 (316-2), and generates count1_sub (242-1) and count2_sub (242-2). In an embodiment of the present disclosure, rollover block 325 is implemented as a combinational block. Rollover block 325 operates to ensure that a respective correct value of input counts is forwarded to Output generator block 290 in case of a rollover of either input count. The details of rollover handling are described below in detail with respect to
The implementation details of output generator block 290 in an embodiment of the present disclosure are provided next.
The specific blocks and interconnections of
Output generator block 290 receives signals window_identifier (232), count1_sub (242-1) and count2_sub (242-2). In an embodiment, signals window_identifier (232), count1_sub (242-1) and count2_sub (242-2) are buffered in a pipeline stage (not shown) that is one-sample deep to enable digital core 110 to meet circuit timing requirements. In an embodiment, the pipeline delay stage is implemented using registers, and the signals are delayed/buffered by one clock cycle of clk_dig_tdc (107).
Subtraction-unit 330 receives count1_sub on path 242-1 and count2_sub on path 242-2. Subtraction-unit 330 generates running_count_diff on path 333. In an embodiment of the present disclosure, subtraction-unit 330 is implemented as a signed subtraction block. Thus, subtraction-unit 330 generates a signed difference between count1_sub (242-1) and count2_sub (242-2). Running_count_diff (333) represents the stream of time differences between edges or edge-pairs of input clocks, clk_in_1 (101-1) and clk_in_2 (101-2) within a window.
Positive-min-generator 335 operates to determine the smallest positive value among/between a set of difference values (running_count_diff (333)) generated by subtraction-unit 330 in each window identified by window-identifier-block 320, i.e., between a pair of successive positive (rising) edges of the slower of the two input clocks. Positive-min-generator 335 receives running_count_diff on path 333 and window_identifier on path 232, and generates pos_min on path 336. In an embodiment of the present disclosure, positive-min-generator 335 is implemented as a sequential block clocked by clk_dig_tdc (107), and operates according to the following logic:
If window_identifier signal (232) is a logic HIGH and running_count_diff (333) is zero or a positive value, then pos_min (336) is recorded as running_count_diff (333). Else, if window_identifier signal (232) is a logic HIGH and running_count_diff (333) is non-zero and a negative value, then pos_min (336) is recorded as full-scale positive value. In other word, if count_1_1 (128-1) and count_1_2 (128-2) consist of P bits each, then full-scale positive value is +(2P−1). Else, if running_count_diff (333) is a positive value and less than pos_min (336), then pos_min (336) is updated to running_count_diff (333). Thus, positive-min-generator 335 operates to determine the minimum positive value of running_count_diff (333) between two successive window_identifier (232) pulses.
Negative-max-generator 345 operates to determine the greatest negative value among/between a set of difference values (running_count_diff (333)) generated by subtraction-unit 330 in each window identified by window-identifier-block 320, i.e., between a pair of successive positive (rising) edges of the slower of the two input clocks. Negative-max-generator 345 receives running_count_diff on path 333 and window_identifier on path 232, and generates neg_max on path 346. In an embodiment of the present disclosure, negative-max-generator 345 is implemented as a sequential block clocked by clk_dig_tdc (107), and operates on the following logic:
If window_identifier signal (232) is a logic HIGH and running_count_diff (333) is zero or a negative value, then neg_max (346) is recorded as running_count_diff (333). Else, if window_identifier signal (232) is a logic HIGH and running_count_diff (333) is non-zero and a positive value, then neg_max (346) is recorded as full-scale negative value. In other word, if count_1_1 (128-1) and count_1_2 (128-2) consist of P bits each, then full-scale negative value is −(2P-1). Else, if running_count_diff (333) is a negative value and greater than neg_max (346) value, then neg_max (346) is updated to running_count_diff (333). Thus, negative-max-generator 345 operates to determine the maximum negative value of running_count_diff (3330 between two window_identifier (232) pulses.
nSample-counter 344 receives pos_min (336) and neg_max (346), and generates signal nsample_done on path 347. In an embodiment, nSample-counter 344 increments a value of a counter upon the end of each window, and asserts nsample_done (347) when the counter value equals the pre-determined number of phase difference samples provided as user input.
Accumulator-positive 340 receives pos_min on path 336, window_identifier on path 232, nsample_done on path 347, and generates acc_pos_q on path 342. Accumulator-positive 340 accumulates (adds) the received pos_min (336) values.
Accumulator-negative 350 receives neg_max on path 346, window_identifier on path 232, nsample_done on path 347, and generates acc_neg_q on path 352. Accumulator-negative 350 accumulates (adds) received neg_max (346) values.
Acc_pos_q (342) and acc_neg_q (352) are examined by running-accumulator-data-generator 365 and accumulator-valid-data-generator 375 to determine outputs (running_acc_data, 155-1-3 and accum_valid_data, 155-1-1) in modes (1) and (2) noted above. Blocks 340, 350, 365 and 375 together comprise accumulation logic operable to generate corresponding outputs in accumulation modes of operation of TDC 100.
Single-shot-data-generator 355 receives pos_min on path 336, neg_max on path 346 and running_count_diff on path 333, and generates single_shot_data on path 155-1-5. In an embodiment of the present disclosure, single-shot-data-generator 355 is implemented as a sequential block, clocked by clk_dig_tdc (107). In continuous single-shot mode of operation, single-shot-data-generator 355 determines the absolute minimum value of pos_min (336) and neg_max (346) and provides it on path 155-1-5. In one-time single-shot mode, single-shot-data-generator 355 forwards running_count_diff(333) on path 155-1-5.
Running-accumulator-data-generator 365 receives acc_pos_q (342) and acc_neg_q (352), and generates running_acc_data on path 155-1-3 and running_nsample on path 155-1-4. Running-accumulator-data-generator 365 determines the absolute minimum value of acc_pos_q (342) and acc_neg_q (352) and provides it on path 155-1-3. Running_nsample (155-1-4) indicates the number of samples accumulated in running_acc_data (155-1-3).
Accumulator-valid-data-generator 375 receives acc_pos_q (342), acc_neg_q (352) and nsample_done (347), and generates signal data_valid on path 155-1-2 and accum_valid_data on path 155-1-1. At the end of sampling every pre-determined number of samples (indicated by nsample_done, 347), accumulator-valid-data-generator 375 asserts data_valid (155-1-2) and provides the absolute minimum of acc_pos_q (342) and acc_neg_q (352) on path 155-1-1.
The implementation details of unique 1 generator and unique_2 generator in an embodiment of the present disclosure are provided next.
Logic_block 405-1 receives count_1_1 (128-1) and count1_q (306-1) and generates unique_count1 (401-1). In an embodiment, logic_block 405-1 generates a logic HIGH on path 401-1 if count_1_1 (128-1) and count1_q (306-1) are unequal, and a logic LOW otherwise. Thus, logic_block 405-1 implements the logic: unique_count1=(count_1 !=count1_q).
MUX 407-1 receives a logic HIGH on path 403-1 and unique_count1_q (311-1), and forwards one of logic HIGH and unique_count1_q on path 408-1 as an output (MUX selected output) based on the logic value of select signal 401-1 (unique_count1). Accordingly, when unique_count1_q (401-1) is a logic HIGH, MUX 407-1 forwards logic HIGH (403-1) on path 408-1. When unique_count1_q (401-1) is a logic LOW, MUX 407-1 forwards unique_count1_q (311-1) on path 408-1.
MUX 409-1 receives a logic LOW on path 404-1 and output of MUX 407-1 on path 408-1, and forwards one of logic LOW and MUX output 408-1 on path 410-1 as an output (MUX selected output) based on the logic value of select signal 419-1. Accordingly, when the value on path 419-1 is a logic HIGH, MUX 407-1 forwards logic LOW (404-1) on path 410-1. When the value on path 419-1 is a logic LOW, MUX 409-1 forwards value 408-1 on path 410-1.
MUX 411-1 receives a logic LOW on path 406-1 and output of MUX 409-1 on path 410-1, and forwards one of logic LOW and MUX output 410-1 on path 412-1 as an output (MUX selected output) based on the logic value of select signal special_case_re (417). Signal special_case_re (417) is used to indicate jitter in input clocks that have zero phase difference in steady state. Accordingly, when special_case_re (417) is a logic HIGH, MUX 411-1 forwards logic LOW (406-1) on path 412-1. When special_case_re (417) is a logic LOW, MUX 411-1 forwards value 410-1 on path 412-1.
MUX 413-1 receives a logic LOW on path 416-1 and output of MUX 411-1 on path 412-1, and forwards one of logic LOW and MUX output 412-1 on path 414-1 as an output (MUX selected output) based on the logic value of select signal rstn_i (419). Rstn_i (419) represents chip reset or power-ON reset. Accordingly, when rstn_i (419) is a logic HIGH, MUX 413-1 forwards logic LOW (416-1) on path 414-1. When rstn_i (419) is a logic LOW, MUX 413-1 forwards value 412-1 on path 414-1.
Flip-flop 415-1 is clocked by clk_dig_tdc (107). Flip-flop 415-1 receives input signal on path 414-1 at its D input and generates output (Q), unique_count1_q, on path 311-1. In an embodiment, flip-flop 415-1 is implemented as a positive edge triggered flip-flop.
AND gate 418-2 receives output of inverter 420-1 on path 421-1 and signal window_identifier on path 232, and generates result of AND operation on path 419-1.
Window-identifier-block 320 receives unique_count1_q (311-1) and unique_count2_q (311-2), and generates signal window_identifier on path 232. Window-identifier-block 320 is implemented as an AND gate in an embodiment of the present disclosure. Thus, window-identifier-block 320 implements the logic: window_identifier=(unique_count1_q && unique_count2_q). In the illustrative embodiment, window_identifier (232) signal is de-asserted after one clock cycle of clk_dig_tdc (107). Accordingly, the pulse width of window_identifier (232) signal is equal to one clock cycle of clk_dig_tdc (107) and the frequency of window_identifier (232) signal is equal to the slower of the two selected input clock (clk_in_1, 101-1 and clk_in_2, 101-2) frequencies (in case of clocks whose frequencies are integer-related).
Components 405-1, 407-1, 409-1, 411-1 and 413-1 together implement the following logic:
if (window_identifier && !unique_count1)
unique_count1_q=0
The expression (window_identifier && !unique_count1) above contains the operators && and ! which respectively represent AND operator and NOT operator. Accordingly, expression (window_identifier && !unique_count1) would evaluate to a 1 when the value of window_identifier is a logic HIGH and the value of unique_count1 is a logic LOW, and to 0 otherwise.
Components 405-2, 420-2, 418-2, 407-2, 409-2, 411-2, 413-2 and 415-2 operate in a manner similar to components 405-1, 420-1, 418-1, 407-1, 409-1, 411-1, 413-1 and 415-1 respectively and the description is not repeated here in the interest of conciseness.
The manner in which unique-count generators and window-identifier-block operate to determine a window is illustrated next with example waveforms.
The value of ‘289’ (count_1_1, 128-1) received at t510 corresponds to the digital representation of the time of occurrence of rising edge (RE11) of clk_in_1 (101-1), while the value ‘310’ (count_1_2, 128-2) received at t512 corresponds to the digital representation of the time of occurrence of rising edge (RE21) of clk_in_2 (101-2). Counters (in count logic, 105, not shown) that are sampled by respective input clocks clk_in_1 and clk_in_2 provide the corresponding count values count_1_1 (128-1) and count_1_2 (128-2) at the respective rising edges of the input clocks clk_in_1 and clk_in_2. As noted above, the counters corresponding to each pair of input clocks are synchronized, i.e., start counting from the same initial value (e.g., zero) and increment at the same rate. Thus, the counters represent time instances on a same time scale.
Counts corresponding to rising edges of clk_in_1 (101-1) are shown as being received at time instants t510, t518 and t526. Thus, count_1_1 (128-1) is shown as having values 289 between t510-t518, 373 between t518-t526 and 457 between t526-t532. Count1_q (306-1) is the value of count_1 (128-1) but delayed by one clock cycle of clk_dig_tdc (107) by flip-flop 305-1. Thus, while count_1 (128-1) has the value 289 starting at t510, count1_q (306-1) is shown as having value 289 starting at t512 (after one cycle of clk_dig_tdc (107)). Accordingly, count1_q (306-1) is shown as having values 289 between t512-t520, 373 between t520-t528 and 457 between t528-t532.
Signal unique_count1 (401-1) is asserted when the value of count_1_1 (128-1) is not equal to the value of count1_q (306-1). In other words, unique_count1 (401-1) is set to a logic HIGH when count of clk_in_1 (101-1) has undergone a change from its value in a previous clock cycle of clk_dig_tdc (107). Accordingly, unique_count1 (401-1) is shown as being asserted only in time intervals t510-t512, t518-t520 and t526-t528, when the value of count_1_1 (128-1) is not equal to the value of count1_q (306-1).
As noted above, unique_count1_q (311-1) is determined based on the following logic:
if (window_identifier && !unique_count1)
unique_count1_q=0
else if (unique_count1)
unique_count1_q=1 (Logic 1.1)
The above logic is evaluated based on values of window_identifier (232) and unique_count1 (401-1) one clock cycle previous to the current time instant. Thus, at t512, value of window_identifier (232) considered for evaluation is a logic LOW (value in time interval t510-t512) and the value of unique_count1 (401-1) considered for evaluation is a logic HIGH. Accordingly, unique_1-generator 310-1 evaluates Logic 1.1 and sets unique_count1_q (311-1) to a logic HIGH starting at t512.
Counts corresponding to rising edges of clk_in_2 (101-2) are shown as being received at time instants t512, t520 and t528. Thus, count_1_2 (128-2) is shown as having values 310 between t512-t520, 394 between t520-t528 and 478 between t528-t532. Count2_q (306-2) is the value of count_2 (128-2) but delayed by one clock cycle of clk_dig_tdc (107) by flip-flop 305-2. Thus, while count_1_2 (128-2) has the value 310 starting at t512, count2_q (306-2) is shown as having value 310 starting at t514 (after one cycle of clk_dig_tdc (107)). Accordingly, count2_q (306-2) is shown as having values 310 between t514-t522, 394 between t522-t530 and 478 between t530-t532.
Signal unique_count2 (401-2) is asserted when the value of count_1_2 (128-2) is not equal to the value of count2_q (306-2). In other words, unique_count2 (401-2) is set to a logic HIGH when count of clk_in_2 (101-2) has undergone a change from its value in a previous clock cycle of clk_dig_tdc (107). Accordingly, unique_count2 (401-2) is shown as being asserted only in time intervals t512-t514, t520-t522 and t528-t530, when the value of count_1_2 (128-2) is not equal to the value of count2_q (306-2).
As noted above with respect to
if (window_identifier && !unique_count2)
unique_count2_q=0
else if (unique_count2)
unique_count2_q=1 (Logic 1.2)
As noted above with respect to unique_count1_q (311-1), the above logic is evaluated based on values of window_identifier (232) and unique_count2 (401-2) one clock cycle previous to the current time instant. Thus, at t512, value of window_identifier (232) considered for evaluation is a logic LOW (value in time interval t510-t512) and the value of unique_count2 (401-2) considered for evaluation is a logic LOW. Accordingly, unique_2-generator 310-2 evaluates Logic 1.2 and unique_count2_q (311-2) remains at a logic LOW starting at t512.
Similarly, at t514, for the evaluation of Logic 1.2, the value of window_identifier (232) is a logic LOW and the value of unique_count2 (401-2) is a logic HIGH. Accordingly, unique_2-generator 310-2 evaluates the above Logic 1.2 and sets unique_count2_q (311-2) to a logic HIGH starting at t514.
When unique_count2_q (311-2) is set to a logic HIGH at t514, window_identifier (232) also goes to logic HIGH as both unique_count1_q (311-1) and unique_count2_q (311-2) are at logic HIGH. As noted above, window_identifier (232) gets de-asserted at t516, i.e., after one clock cycle of clk_dig_tdc (107).
At t516, for the determination of unique_count1_q (311-1), the value of window_identifier (232) is a logic HIGH and the value of unique_count1 (401-1) is a logic LOW. Accordingly, unique_count1_q (311-1) is set to a logic LOW starting at t516.
At t516, for the determination of unique_count2_q (311-2), the value of window_identifier (232) is a logic HIGH and the value of unique_count2 (401-2) is a logic LOW. Accordingly, unique_count2_q (311-2) is set to a logic LOW starting at t516.
The values for unique_count1_q (311-1), unique_count2_q (311-2) and window_identifier (232) are generated in a similar manner for respective time instants up to t532, as depicted in
Thus, window_identifier (232) is shown as identifying the boundaries of windows, with each window corresponding to a pair of successive edges of an input clock (clk_in_1, 101-1 in this illustration).
The manner in which unique count signals (410, 311) and signal window_identifier (232) are generated for a pair of input clocks having frequencies that are integer-related multiples is described next.
In
It is noted here that the rising edges of the input clocks are not synchronous with respect to the measurement clock clk_meas (103) or used in count logic 105 or clk_dig_tdc (107). Accordingly, an uncertainty of up to 3 cycles of clk_dig_tdc (107) may be present between the occurrence of a rising edge and the arrival of the corresponding count. For example, count_1_1 (128-1) corresponding to rising edge E11 of clk_in_1 (101-2) is shown as being received at t610, slightly later than the actual occurrence of the rising edge.
Counts corresponding to rising edges of clk_in_1 (101-1) are shown as being received at time instants t610, t616, t618, t620, t622, t624, t626, t628 and t630. Thus, count_1_1 (128-1) is shown as having values 297 between t610-t616, 381 between t616-618, 465 between t618-t620, 549 between t620-t622, 633 between t622-t624, 717 between t624-t626, 801 between t626-t628, 885 between t628-t630 and 969 between t630-t636. Count1_q (306-1) is the value of count_1_1 (128-1) but delayed by one clock cycle of clk_dig_tdc (107) by flip-flop 305-1. Thus, while count_1_1 (128-1) has the value 297 starting at t610, count1_q (306-1) is shown as having value 297 starting at t612 (after one cycle of clk_dig_tdc (107)). Accordingly, each value of count1_q (306-1) is shown to be delayed by one cycle of clk_dig_tdc (107) with respect to a respective value of count_1_1 (128-1) in
Signal unique_count1 (401-1) is asserted when the value of count_1_1 (128-1) is not equal to the value of count1_q (306-1). In time interval t610-t612, count_1_1 (128-1) has a value 297 while count1_q (306-1) has a value 214, and therefore unique_count1 (401-1) is shown as being asserted in time interval t610-t612. Each respective value of unique_count1 (401-1) is set to a logic HIGH in a time interval when the corresponding value of count_1_1 (128-1) is not equal to the value of count1_q (306-1) in that time interval.
Prior to t610, unique_count1_q (311-1) is shown as having a logic HIGH value. Unique_count1_q (311-1) is determined according to Logic 1.1 noted above. Logic 1.1 is evaluated based on values of window_identifier (232) and unique_count1 (401-1) one clock cycle previous to the current time instant. Thus, at t612, value of window_identifier (232) considered for evaluation is a logic LOW (value in time interval t610-t612) and the value of unique_count1 (401-1) considered for evaluation is a logic HIGH. Accordingly, unique_1-generator 310-1 evaluates the above logic and continues to keep unique_count1_q (311-1) at a logic HIGH at t612.
At t614, for the determination of unique_count1_q (311-1), the value of window_identifier (232) is a logic HIGH and the value of unique_count1 (401-1) is a logic LOW. Accordingly, unique_1-generator 310-1 evaluates the above logic and de-asserts unique_count1_q (311-1) starting at t614.
Signal unique_count2 (401-2) is asserted when the value of count_1_2 (128-2) is not equal to the value of count1_q (306-2). Thus, unique_count2 (401-2) is shown as being asserted only in time intervals t610-t612 and t630-t632.
Unique_count2_q (311-2) is determined in a manner similar to that of unique_count1_q (311-1). Thus, unique_count2_q (311-2) is shown as being asserted in time interval t612-t614 (as the value of window_identifier, (232) considered for evaluation is a logic LOW and the value of unique_count2 (401-2) considered for evaluation is a logic HIGH). Similarly, unique_count2_q (311-2) is also asserted in time interval t632-t634.
At 612, when unique_count2_q (311-2) is set to a logic HIGH, window_identifier (232) also goes to logic HIGH as both unique_count1_q (311-1) and unique_count2_q (311-2) are at logic HIGH. As noted above, window_identifier (232) gets de-asserted at t614, i.e., after one clock cycle of clk_dig_tdc (107). Similarly, window_identifier (232) gets asserted again in time interval t632-t634.
Phase difference of a pair of input clocks whose frequencies are integer-related multiples is the time between occurrence of an edge of the slower clock and the nearest edge of the faster clock. Thus, window_identifier (232) pulses are shown as being generated upon occurrence of rising edges of clk_in_2 (101-2) (slower clock) in
Rollover of count values is taken into consideration while determining phase difference between inputs. As is well known in the relevant arts, a counter has a finite width and can count only up to a certain maximum value (referred to as full-scale value) as determined by the width of the counter. A rollover event is said to occur when the value of a counter reaches its full-scale value. Thereafter, the counter resets and starts counting from zero again.
If one of the counters corresponding to an input clock has rolled over and the other counter has not rolled over, this may result in calculation error in subtraction-unit 330. In order to prevent such errors, the rollover event needs to be handled. Rollover events are handled by rollover1-generator 315-1, rollover2-generator 315-2 and rollover block 325.
The manner in which rollover is handled in an embodiment of the present disclosure is described next with respect to
For ease of understanding,
In an embodiment, rollover1-generator 315-1 determines that a rollover has occurred when the MSB (most significant bit) of count value corresponding to clk_in_1 (101-1) has changed from a value of 1 in a previous clk_dig_tdc (107) (not shown) cycle to a value of 0 in the current clock cycle. Rollover1-generator 315-1 performs this determination by comparing MSB values of count_1_1 (128-1) and count1_q (306-1). Rollover2-generator 315-2 operates in a manner similar to rollover1-generator 315-1 in order to determine if a rollover has occurred for count value corresponding to clk_in_2 (101-2).
In an embodiment of the present disclosure, rollover block 325 implements the following logic to handle rollover events:
If (rollover_1==0 && rollover_2==0)
count1_sub=count1_q
count2_sub=count2_q
Else if (rollover_1==0 && rollover_2==1)
count1_sub=count1_q
count2_sub=count2_q+(2{circumflex over ( )}10−1)
Else if (rollover_1==1 && rollover_2==0)
count1_sub=count1_q+(2{circumflex over ( )}10−1)
count2_sub=count2_q
Else if (rollover_1==1 && rollover_2==1)
count1_sub=count1_q
count2_sub=count2_q (Logic 2)
Expression (rollover_1=0 && rollover_2=0) would evaluate to a ‘1’ when both rollover_1 and rollver_2 have a value of logic LOW, and would evaluate to a ‘0’ otherwise. Expression (rollover_1=0 && rollover_2==1) would evaluate to a ‘1’ when rollover_1 has a value of logic LOW and rollver_2 has a value of logic HIGH, and would evaluate to a ‘0’ otherwise. Expression (rollover_1==1 && rollover 2==0) would evaluate to a ‘1’ when rollover_1 has a value of logic HIGH and rollver_2 has a value of logic LOW, and would evaluate to a ‘0’ otherwise. Expression (rollover_1==1 && rollover_2==1) would evaluate to a ‘1’ when both rollover_1 and rollver_2 have a value of logic HIGH, and would evaluate to a ‘0’ otherwise.
Referring to
Just prior to t719, count2_q (306-2) is shown to be having a value 960 while count_1_2 (128-2) is shown to be having the value 37. The binary equivalents of the two counts are as shown below:
960: 1111000000
37: 0000100101
Thus, MSB of count_1_2 (128-2) has changed from a value of ‘1’ in a previous clock cycle (indicated by value of count2_q (306-2)) to a value of ‘0’ in the current clock cycle of clk_dig_tdc (107). Therefore, rollover1 generator 315-2 asserts signal rollover_2 (316-2) starting at t719.
Just prior to t719, both count1_q (306-1) and count_1_1 (128-1) are shown to be having the value 995. Therefore, rollover2 generator 315-1 does not assert signal rollover_1 (316-1) at t719.
At t719, inputs to rollover block 325 are accordingly a logic LOW (for rollover_1 (316-1)) and a logic HIGH (for rollover_2 (316-2)). Rollover block 325 evaluates logic 2 and generates outputs:
count1_sub=count1_q (995 in this case)
count2_sub=count2_q+(2{circumflex over ( )}10−1)=(37+1024-1)=1060. This is depicted in time interval t719-t723 in
Subtraction-unit 330 receives count1_sub (242-1) and count2_sub (242-2) and generates output (995 minus 1060), i.e., ‘−65’, as depicted in waveform running_count_diff (333) in time interval t719-t723.
Just prior to t723, count1_q (306-1) is shown to be having a value 995 while count_1_1 (128-1) is shown to be having the value 72. Therefore, rollover1-generator 315-1 determines that MSB of count_1_1 (128-1) has changed (from a value of ‘1’ in a previous clock cycle to a value of ‘0’ in the current clock cycle of clk_dig_tdc, 107), and asserts signal rollover_1 (316-1) starting at t723.
At t723, inputs to rollover block 325 are accordingly a logic HIGH (for rollover_1 (316-1)) and a logic HIGH (for rollover_2 (316-2)). Rollover block 325 evaluates logic 2 and generates outputs:
count1_sub=count1_q(72 in this example)
count2_sub=count2_q(37 in this example)
Subtraction-unit 330 receives count1_sub (242-1) and count2_sub (242-2) and generates output (72 minus 37), i.e., ‘35’, as depicted in waveform running_count_diff (333) in time interval t723-t724.
Signals rollover_1 (316-1) and rollover_2 (316-2) get de-asserted when window_identifier (232) gets de-asserted after a rollover event of counters of both inputs (count_1_1, 128-1 and count_1_2, 128-2). Thus, although window_identifier (232) is shown as being de-asserted at t720, rollover_2 (316-2) continues to remain asserted. At t725, rollover_1 (316-1) and rollover_2 (316-2) are shown as being de-asserted when window_identifier (232) gets de-asserted.
In this manner, rollover1-generator 315-1, rollover2-generator 315-2 and rollover block 325 together ensure that rollover events of respective counters do not result in corresponding calculation errors.
The various blocks of a calculation unit as described in detail above enable TDC 100 in generating digital values representing the phase difference between a pair of periodic inputs whose frequencies are related to each other by an integer value, i.e., whose frequencies are the same or that are integer-related multiples (e.g., 2×, 3×, etc.). The manner in which TDC 100 determines the phase difference for each of the above scenarios is illustrated below with respect to example diagrams.
Additionally,
Rising edges R11 through R13 of clk_in_1 (101-1) are shown to be occurring at respective time instants t810, t818 and t826. Accordingly, digital core 110 receives values of count_1_1 (128-1) as 289, 373 and 457 respective to each rising edge of clk_in_1 (101-1).
Similarly, rising edges R21 through R23 of clk_in_2 (101-2) are shown to be occurring at respective time instants t812, t820 and t828. Accordingly, digital core 110 receives values of count_1_2 (128-2) as 310, 394 and 478, respective to each rising edge of clk_in_2 (101-2). Thus, in the illustrative embodiment, clk_in_2 (101-2) is shown to be lagging in phase (by approximately 90 degrees) with respect to clk_in_1 (101-1).
Although counts (count_1_1, 128-1 and count_1_2, 128-2) corresponding to rising edges of input clocks (clk_in_1, 101-1 and clk_in_2, 101-2) are shown to be received synchronous with the respective rising edges, in practice, counts may be received slightly later than the occurrence of corresponding rising edges of clk_in_1 (101-1) and clk_in_2 (102-2).
In response to count_1_2 (128-2) changing its value (to 310) at t812, window_identifier (232) is asserted at t814 for one clock cycle of clk_dig_tdc (107). Thus, starting at t816, phase differences between clk_in_1 (101-1) and clk_in_2 (101-2) are measured.
Running_count_diff (333) is shown as having the following values in respective time intervals, based on value of (count1_q (306-1) minus count2_q (306-2)) in the durations (based on pipeline delay noted above):
t816-t818: running_count_diff(333)=289-310=−21
t818-t820: running_count_diff(333)=289-310=−21
t820-t822: running_count_diff(333)=289-310=−21
t822-t824: running_count_diff(333)=373-310=63
The phase difference in duration t816-t822 represents the phase difference between rising edge R21 of clk_in_2 (101-2) and rising edge R11 of clk_in_1 (101-1), while the phase difference in duration t822-t824 represents the phase difference between rising edge R12 of clk_in_2 (101-2) and rising edge R12 of clk_in_1 (101-1).
At t818, positive-min-generator 335 receives a logic HIGH on path 232 (window_identifier) and a value of ‘−21’ on path 333 (running_count_diff). Therefore, positive-min-generator 335 sets the value of pos_min (336) to max1 at t818.
At t820, positive-min-generator 335 receives a logic LOW on path 232 (window_identifier) and a value of ‘−21’ on path 333 (running_count_diff). Therefore, positive-min-generator retains the current value of pos_min (336), i.e., max1, at t820. In this manner, values of pos_min (336) are shown to be updated based on values of window_identifier (232), running_count_diff (333) and the current value of pos_min (336).
At t818, negative-max-generator 345 receives a logic HIGH on path 232 (window_identifier) and a value of ‘−21’ on path 333 (running_count_diff). Therefore, negative-max-generator 345 sets the value of neg_max (346) to the value of running_count_diff (333), i.e., ‘−21’.
At t820, negative-max-generator 345 receives a logic LOW on path 232 (window_identifier) and a value of ‘−21’ on path 333 (running_count_diff). Therefore, negative-max-generator 345 retains the current value of neg_max (346), i.e., ‘−21’ at t820. In this manner, values of neg_max (346) are shown to be updated/retained based on values of window_identifier (232), running_count_diff (333) and the current value of neg_max (346).
When TDC 100 operates in mode (3) (single-shot continuous mode), output generator block 290 (specifically, single-shot-data generator 355) forwards the absolute minimum of pos_min (336) and neg_max (346) on path 155-1-5 as an output of the sequence of outputs, with each output corresponding to a respective window determined by window-identifier-block 320. Thus, in the above example, the value of ‘−21’ would be output on path 155-1-5. Since the time resolution of counters is 62.5 pico-seconds in this illustrative embodiment, a value of 21 indicates that the phase difference between the pair of input clocks is (21*62.5 pico-seconds), i.e., 1312.5 pico-seconds.
Using similar operations as illustrated above, TDC 100 determines the phase difference between periodic clock inputs having equal frequencies for other values of phase difference (leading or lagging).
The manner in which TDC 100 determines the phase difference between a pair of input clocks having frequencies that are integer-related multiples (2x or greater) is illustrated next with an example.
As noted above with respect to
Also, as noted above with respect to
In response to count_1_2 (128-2) changing its value (to 281) at t854, window_identifier (232) is asserted at t855 for one clock cycle of clk_dig_tdc (107). Thus, starting at t856 (one clock cycle after window_identifier, 232, is asserted), phase differences are measured between rising edge D21 of slower clock clk_in_2 (101-2) and each rising edge (D11 through D18) of faster clock clk_in_1 (101-2) present between two successive edges D21 and D22 of slower clock clk_in_2 (101-2).
Running_count_diff (333) is shown as having the following values in respective time intervals, based on value of (count1_q (306-1) minus count2_q (306-2)) in the durations (based on pipeline delay noted above):
t856-t859: running_count_diff(333)=297−281=16
t859-t861: running_count_diff(333)=381−281=100
t861-t863: running_count_diff (333)=465−281=184
t863-t865: running_count_diff(333)=549−281=268
t865-t867: running_count_diff(333)=633−281=352
t867-t869: running_count_diff(333)=717−281=436
t869-t874: running_count_diff(333)=801−281=520
t874-t877: running_count_diff(333)=885−281=604
Thus, running_count_diff (333) is shown as having 8 distinct values, with each value representing the phase difference between rising edge D21 of clk_in_2 (101-2) and a corresponding one of rising edges D1I through D18 (of clock clk_in_1, 101-1).
Prior to t857, pos_min (336) is shown as having a value of 17 as a result of a previous (not shown) minimum phase difference between two consecutive rising edges of clk_in_2 (101-2). At t857, positive-min-generator 335 receives the value of running_count_diff (333) as 16 and a logic HIGH on path 232 (window_identifier), Therefore, positive-min-generator 335 updates pos_min (336) to the value of running_count_diff (333), i.e., 16, at t857. Between t857 and t876, since none of running_count_diff (333) values exceeds the current value of 16, positive-min-generator 335 retains the value of pos_min (336) at its current value, i.e., 16.
Prior to t857, neg_max (346) is shown as having a large negative value of (−236-1). Starting at t857, since all phase difference values are positive, negative-max-generator 345 retains neg_max (346) at the large negative value.
At t875, window_identifier (232) signal is asserted again, indicating the occurrence of the next rising edge (D22) of clk_in_2 (101-2). It may be observed that the value of 16 is the minimum phase difference amongst all 8 differences measured between two window_identifier (232) pulses. Accordingly, phase difference between clk_in_1 (101-2) and clk_in_2 (101-2) is determined to be 16 at t875. As noted above, phase difference of a pair of input clocks whose frequencies are integer-related multiples is the time between occurrence of an edge of the slower clock and the nearest edge of the faster clock.
Thus, when TDC 100 operates in mode (3) (single-shot continuous mode), Output generator block 290 (specifically, single-shot-data generator 355) forwards the absolute minimum of pos_min (336) and neg_max (346) on path 155-1-5. Thus, in the above example, the value of ‘16’ would be output on path 155-1-5. Since the time resolution of counters is 62.5 pico-seconds in this illustrative embodiment, a value of 16 indicates that the phase difference between the pair of input clocks is (16*62.5 pico-seconds), i.e., 1000 pico-seconds.
According to another aspect of the present disclosure, TDC 100 accumulates a pre-determined (provided by user) number of phase differences between input clocks (clk_in_1, 01-1 and clk_in_2, 101-2) and averages the accumulated value (i.e., accumulated value divided by the number of phase difference samples) in order to potentially improve resolution with which the phase difference can be determined. The manner in which such accumulation is implemented is described next with an example.
As noted earlier with respect to
Running_acc_data (155-1-3) is desirable in situations where input clocks (clk_in_1, 101-1 and clk_in_2, 101-2) have small frequencies (such as ˜1 Hertz) and the user has programmed a relatively high number (say, 1024) of pre-determined number of phase difference samples to be accumulated. Each accumulation cycle would accordingly be of a duration of 1024 seconds. In such cases, the user is required to wait for 1024 seconds (after start of an accumulation cycle) before accumulated data is available on output path 155-1-1 (accum_valid_data). Meanwhile, if the user desires to check an intermediate accumulated value (say at around 3 seconds or 10 seconds after start of the accumulation cycle), the user can do so by reading accumulated data available on path 155-1-3 (running_acc_data) upon the end of any desired number of windows.
Referring to
Accumulated data is available on path 155-1-1 by time instant t925 (at the end of first accumulation cycle in duration t910-t925), and is held at that value (accumulator valid data1) until the next accumulation cycle is complete (in this case, until t935). At t925, data_valid (155-1-2) is also asserted to indicate that data available on path 155-1-1 is valid. If the user reads output on path 155-1-1 during time interval t925-t935, the user is provided with value (accumulator valid data1) updated at t925.
In mode (1), referring to
In mode (2), referring to
Continuing to refer to
Thus, when the accumulation cycle ends at t875 (based on pre-determined number of phase difference samples, provided as user input), accumulator-valid-data-generator 375 determines the absolute minimum of acc_pos_q (342) and acc_neg_q (352), and provides the value on path 155-1-1 (accum_valid_data). Thus, the value of 672 is provided as accum_valid_data (155-1-1), starting at t878 until the end of the next accumulation cycle.
Running_nsample (155-1-4) value is incremented every time window_identifier (232) (not shown) is asserted. Each value of data running_acc_data (155-1-3) is also updated every time window_identifier (232) is asserted (upon the end of each window). Thus, if the user desires to check the accumulated value between the start and end of an accumulation cycle, the user may read running_acc_data on path 155-1-3. Running_nsample data available on path 155-1-4 indicates the number of accumulated samples available on path 155-1-3.
For example, referring now to
In mode (1), referring to
In this manner, TDC 100 provides accumulated data in accumulator modes (1) and (2). The description is continued to illustrate the manner in which TDC 100 generates output in continuous single-shot mode of operation.
Counts data10 through data16 respectively correspond to rising edges of clk_in_1 (101-1, not shown), while counts data20 through data25 respectively correspond to rising edges of clk_in_1 (101-1, not shown). Running_count_diff (333) is shown as being computed as diff0 through diff11, with each difference representing a difference in counts between corresponding input clocks (as described in detail with respect to
Single-shot-data-generator 355 receives pos_min on path 336, neg_max on path 346 and running_count_diff on path 333, and generates single_shot_data on path 115. Upon the end of each window (as determined by window_identifier 232), single-shot-data-generator 355 determines an absolute minimum value of pos_min (336) and neg_max (346) and provides the absolute minimum value on path 115. This mode of operation may be suitable in cases where the input clocks have the same frequency, and the average frequency of input clocks remains constant but the instantaneous phase difference changes continuously/is fluctuating.
In mode (4), calculation unit 150 operates on pulse inputs, as noted above. Thus, single-shot-data-generator 355 forwards running_count_diff (333) as a one-time phase difference between the pulses, and provides the phase difference on path 115 (single_shot_data).
In this manner, aspects of the present disclosure enable a TDC to measure phase difference between periodic inputs.
TDC 100 implemented as described above can be incorporated in a larger device or system as described briefly next.
As is well known in the relevant arts, SyncE is a physical layer (PHY)-based technology for achieving synchronization in packet-based Ethernet networks. The SyncE clock signal transmitted over the physical layer should be traceable to an external master clock (for example, from a timing card such as card 1110 or 1120). Accordingly, Ethernet packets are re-timed with respect to the master clock, and then transmitted in the physical layer. Thus, data packets (e.g., on path 1131 and 1132) are re-timed and transmitted without any time stamp information being recorded in the data packet. The packets may be generated by corresponding applications such as IPTV (Internet Protocol Television), VoIP (Voice over Internet Protocol), etc.
Thus, line card 1130 receives a packet on path 1131, and forwards the packet on output 1146 after the packet has been re-timed (synchronized) with a master clock. Similarly, line card 1150 receives a packet on path 1151, and forwards the packet on output 1166 after the packet has been re-timed (synchronized) with a master clock.
The master clock (1111/clock-1) is generated by timing card 1110. Timing card 1120 generates a redundant clock (1121/clock-2) that is to be used by line cards 1130 and 1150 upon failure of master clock 1111. Master clock 1111 and redundant clock 1121 are provided via a backplane (represented by numeral 1170) to each of lines cards 1130 and 1150.
In line card 1130, jitter attenuator PLL 1140 contains TDC 100 as the phase detector for detecting phase differences between the selected one of input clocks clock-1/clock-2, and output clock 1141, with TDC 100 being as described above in detail. PLL 1140 generates an output clock 1141 which is used to synchronize (re-time) packets received on path 1131 and forwarded as re-timed packets on path 1146.
Similarly, in line card 1150, jitter attenuator PLL 1160 contains TDC 100 as the phase detector for detecting phase difference between the selected one of input clocks clock-1/clock-2, and output clock 1161, described above in detail, and receives clocks 1111 and 1121, described above in detail, and receives clocks 1111 and 1121. PLL 1160 generates an output clock 1161 which is used to synchronize (re-time) packets received on path 1151 and forwarded as re-timed packets on path 1166. Upon being out of phase-lock, PLL 1160 is designed to provide locking of phase in the manner described above in detail.
References throughout this specification to “one embodiment”, “an embodiment”, or similar language means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present disclosure. Thus, appearances of the phrases “in one embodiment”, “in an embodiment” and similar language throughout this specification may, but do not necessarily, all refer to the same embodiment.
While in the illustrations of
While various embodiments of the present disclosure have been described above, it should be understood that they have been presented by way of example only, and not limitation. Thus, the breadth and scope of the present disclosure should not be limited by any of the above-described embodiments, but should be defined only in accordance with the following claims and their equivalents.
| Number | Name | Date | Kind |
|---|---|---|---|
| 7071751 | Kaviani | Jul 2006 | B1 |
| 9092013 | Song et al. | Jul 2015 | B2 |
| 9405024 | Frach et al. | Aug 2016 | B2 |
| 10496041 | Yao et al. | Dec 2019 | B2 |
| 10505554 | Janardhanan | Dec 2019 | B2 |
| 10567154 | Wentzloff | Feb 2020 | B1 |
| 10671025 | Mautner et al. | Jun 2020 | B2 |
| 10819355 | Abughazaleh | Oct 2020 | B1 |
| 10868541 | Choi et al. | Dec 2020 | B1 |
| 11387834 | Kumar | Jul 2022 | B1 |
| 20040161068 | Zerbe | Aug 2004 | A1 |
| 20080069292 | Straayer et al. | Mar 2008 | A1 |
| 20090296532 | Hsieh | Dec 2009 | A1 |
| 20100134083 | Trescases | Jun 2010 | A1 |
| 20200028662 | Brown | Jan 2020 | A1 |
| 20220026856 | Lee | Jan 2022 | A1 |
| Entry |
|---|
| M. Skandar Basrour et al., “Time-to-Digital Conversion based on a Self-Timed Ring Oscillator” Nov. 20, 2019, 145 pages. |
| “Variable Modulus Counter” http://www.elektronikjk.pl/elementy_czynne/IC/9305.pdf, downloaded circa Jan. 10, 2022, 04 Pages. |
| Wu Gao et al., “Integrated High-Resolution Multi-Channel Time-to-Digital Converters (TDCs) for PET Imaging”, http://citeseerxist.psu.edu/viewdoc/download?doi=10.1.1.688.658&rep=rep1&type=pdf, Jan. 8, 2011, 23 Pages. |
| Kuo-Hsing Cheng et al., “A time-to-digital converter using multi-phase-sampling and time amplifier for all digital phase-locked loop,” 13th IEEE Symposium on Design and Diagnostics of Electronic Circuits and Systems, Date of Conference: Apr. 14-16, 2010, 05 Pages. |
| Scott Tancock et al., “A Review of New Time-to-Digital Conversion Techniques,” in IEEE Transactions on Instrumentation and Measurement, vol. 68, No. 10, pp. 3406-3417, Oct. 2019. |
| Kyu-Dong Hwang et al., “An area efficient asynchronous gated ring oscillator TDC with minimum GRO stages,” Proceedings of 2010 IEEE International Symposium on Circuits and Systems, Date of Conference: May 30, 2010-Jun. 2, 2010, pp. 3973-3976. |
| “TDC7201 Time-to-Digital Converter for Time-of-Flight Applications in LIDAR, Range Finders, and ADAS” SNAS686—May 2016 DOI: https://www.ti.com/lit/ds/symlink/tdc7201.pdf?ts=1641356653664&ref_url=https%253A%252F%252Fwww.google.com%252F, May 2016, 50 Pages. |
| “Time-to-Digital Converter with Analog Front-End” MAX35101 DOI: https://datasheets.maximintegrated.com/en/ds/MAX35101.pdf, Downloaded circa Jan. 10, 2022, 66 Pages. |
| J. Christiansen, “High Performance Time to Digital Converter” Version 2.2, https://cds.cern.ch/record/1067476/files/cer-002723234.pdf, Mar. 2004, 102 Pages. |
| “MTD135 8-Channel, Multi-hit Time-to-Digital Converter” , https://teledynelecroy.com/lrs/dsheets/mtd.htm, downloaded circa Jan. 7, 2022, 10 Pages. |
| “TDC-GP1 General Purpose TDC” General Purpose TDC, Feb. 12, 2001, acam mess electronic, https://www.pmt-fl.com/uploads/2017/09/precision-measurement-db-gp1-en-1019306.pdf, Feb. 12, 2001, 61 Pages. |
| QuTag, https://www.qutools.com/wp-content/uploads/2019/02/quTAG_Datasheet.pdf, Downloaded circa Jan. 7, 2022, 04 Pages. |
