US20210159904A1 - On-chip phase-locked loop response measurement - Google Patents
On-chip phase-locked loop response measurement Download PDFInfo
- Publication number
- US20210159904A1 US20210159904A1 US16/697,613 US201916697613A US2021159904A1 US 20210159904 A1 US20210159904 A1 US 20210159904A1 US 201916697613 A US201916697613 A US 201916697613A US 2021159904 A1 US2021159904 A1 US 2021159904A1
- Authority
- US
- United States
- Prior art keywords
- phase
- clock signal
- pll
- phase error
- digital
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000004044 response Effects 0.000 title claims abstract description 51
- 238000005259 measurement Methods 0.000 title abstract description 21
- 238000003780 insertion Methods 0.000 claims abstract description 20
- 230000037431 insertion Effects 0.000 claims abstract description 20
- 238000012360 testing method Methods 0.000 claims description 31
- 238000000034 method Methods 0.000 claims description 16
- 238000001514 detection method Methods 0.000 claims description 10
- 238000012544 monitoring process Methods 0.000 description 17
- 238000012546 transfer Methods 0.000 description 11
- 238000013016 damping Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 8
- 230000001052 transient effect Effects 0.000 description 7
- 238000004088 simulation Methods 0.000 description 5
- 238000013459 approach Methods 0.000 description 4
- 230000000875 corresponding effect Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 244000145845 chattering Species 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 239000013078 crystal Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000005316 response function Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03L—AUTOMATIC CONTROL, STARTING, SYNCHRONISATION OR STABILISATION OF GENERATORS OF ELECTRONIC OSCILLATIONS OR PULSES
- H03L7/00—Automatic control of frequency or phase; Synchronisation
- H03L7/06—Automatic control of frequency or phase; Synchronisation using a reference signal applied to a frequency- or phase-locked loop
- H03L7/08—Details of the phase-locked loop
- H03L7/099—Details of the phase-locked loop concerning mainly the controlled oscillator of the loop
- H03L7/0991—Details of the phase-locked loop concerning mainly the controlled oscillator of the loop the oscillator being a digital oscillator, e.g. composed of a fixed oscillator followed by a variable frequency divider
- H03L7/0992—Details of the phase-locked loop concerning mainly the controlled oscillator of the loop the oscillator being a digital oscillator, e.g. composed of a fixed oscillator followed by a variable frequency divider comprising a counter or a frequency divider
- H03L7/0993—Details of the phase-locked loop concerning mainly the controlled oscillator of the loop the oscillator being a digital oscillator, e.g. composed of a fixed oscillator followed by a variable frequency divider comprising a counter or a frequency divider and a circuit for adding and deleting pulses
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03L—AUTOMATIC CONTROL, STARTING, SYNCHRONISATION OR STABILISATION OF GENERATORS OF ELECTRONIC OSCILLATIONS OR PULSES
- H03L7/00—Automatic control of frequency or phase; Synchronisation
- H03L7/06—Automatic control of frequency or phase; Synchronisation using a reference signal applied to a frequency- or phase-locked loop
- H03L7/08—Details of the phase-locked loop
- H03L7/085—Details of the phase-locked loop concerning mainly the frequency- or phase-detection arrangement including the filtering or amplification of its output signal
- H03L7/095—Details of the phase-locked loop concerning mainly the frequency- or phase-detection arrangement including the filtering or amplification of its output signal using a lock detector
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03L—AUTOMATIC CONTROL, STARTING, SYNCHRONISATION OR STABILISATION OF GENERATORS OF ELECTRONIC OSCILLATIONS OR PULSES
- H03L7/00—Automatic control of frequency or phase; Synchronisation
- H03L7/06—Automatic control of frequency or phase; Synchronisation using a reference signal applied to a frequency- or phase-locked loop
- H03L7/07—Automatic control of frequency or phase; Synchronisation using a reference signal applied to a frequency- or phase-locked loop using several loops, e.g. for redundant clock signal generation
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03L—AUTOMATIC CONTROL, STARTING, SYNCHRONISATION OR STABILISATION OF GENERATORS OF ELECTRONIC OSCILLATIONS OR PULSES
- H03L7/00—Automatic control of frequency or phase; Synchronisation
- H03L7/06—Automatic control of frequency or phase; Synchronisation using a reference signal applied to a frequency- or phase-locked loop
- H03L7/08—Details of the phase-locked loop
- H03L7/0805—Details of the phase-locked loop the loop being adapted to provide an additional control signal for use outside the loop
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03L—AUTOMATIC CONTROL, STARTING, SYNCHRONISATION OR STABILISATION OF GENERATORS OF ELECTRONIC OSCILLATIONS OR PULSES
- H03L7/00—Automatic control of frequency or phase; Synchronisation
- H03L7/06—Automatic control of frequency or phase; Synchronisation using a reference signal applied to a frequency- or phase-locked loop
- H03L7/08—Details of the phase-locked loop
- H03L7/085—Details of the phase-locked loop concerning mainly the frequency- or phase-detection arrangement including the filtering or amplification of its output signal
- H03L7/087—Details of the phase-locked loop concerning mainly the frequency- or phase-detection arrangement including the filtering or amplification of its output signal using at least two phase detectors or a frequency and phase detector in the loop
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03L—AUTOMATIC CONTROL, STARTING, SYNCHRONISATION OR STABILISATION OF GENERATORS OF ELECTRONIC OSCILLATIONS OR PULSES
- H03L7/00—Automatic control of frequency or phase; Synchronisation
- H03L7/06—Automatic control of frequency or phase; Synchronisation using a reference signal applied to a frequency- or phase-locked loop
- H03L7/08—Details of the phase-locked loop
- H03L7/099—Details of the phase-locked loop concerning mainly the controlled oscillator of the loop
- H03L7/0991—Details of the phase-locked loop concerning mainly the controlled oscillator of the loop the oscillator being a digital oscillator, e.g. composed of a fixed oscillator followed by a variable frequency divider
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03L—AUTOMATIC CONTROL, STARTING, SYNCHRONISATION OR STABILISATION OF GENERATORS OF ELECTRONIC OSCILLATIONS OR PULSES
- H03L2207/00—Indexing scheme relating to automatic control of frequency or phase and to synchronisation
- H03L2207/50—All digital phase-locked loop
Definitions
- This invention relates to phase-locked loops (PLLs) and more particularly to measuring PLL response characteristics.
- the PLL closed loop transfer function is often measured on a tester using a sinusoidal signal generator to modulate the reference clock signal and a spectrum analyzer/oscilloscope to measure the PLL response by sending the PLL output clock signal off-chip to the test equipment.
- a sinusoidal signal generator to modulate the reference clock signal
- a spectrum analyzer/oscilloscope to measure the PLL response by sending the PLL output clock signal off-chip to the test equipment.
- a method for determining response of a phase-locked loop includes inserting a digital phase offset into a digital representation of a clock signal coupled to a phase and frequency detector (PFD) of the PLL, the digital phase offset representing a phase step.
- the method further includes determining a length of time between insertion of the digital phase offset and a first zero crossing of a phase error supplied from the PFD after insertion of the digital phase offset, the length of time being inversely related to bandwidth of the PLL.
- the method may further include determining a maximum phase error overshoot resulting from insertion of the digital phase offset and determining a peaking characteristic of the PLL based in part on the maximum phase error overshoot.
- an integrated circuit in another embodiment, includes a summing circuit to add a digital phase offset to a digital representation of a first clock signal, the digital phase offset corresponding a phase step in the first clock signal.
- a phase and frequency detector of a phase-locked loop (PLL) receives the digital representation of the first clock signal and a digital representation of a second clock signal and supplies a phase error signal indicative of a difference between the first clock signal and the second clock signal.
- a first zero crossing detection circuit receives the phase error and detects a first zero crossing of the phase error after insertion of the phase offset.
- a counter determines an elapsed time between insertion of the phase offset and the first zero crossing.
- a maximum phase error circuit receives the phase error signal and to detect a maximum phase error overshoot resulting from insertion of the digital phase offset.
- FIG. 1 illustrates a high level block diagram of a system that includes on-chip PLL response measurement capability.
- FIG. 3 illustrates a functional block diagram of an embodiment of a phase lock monitoring circuit.
- FIG. 4 illustrates PLL jitter transfer with different PLL bandwidths and peaking.
- FIG. 5A illustrates PLL phase error response to a phase step.
- FIG. 5B illustrates PLL phase error response to a phase step showing the phase error ringing settling to a steady state.
- FIG. 6A illustrates a functional block diagram of an embodiment of an on-chip PLL response measurement circuit.
- FIG. 6B shows a comparator used as a zero crossing detector.
- FIG. 7 illustrates a graph of crossover time versus inverse of closed loop 3 dB bandwidth.
- FIG. 8 illustrates a graph of maximum phase error overshoot versus closed loop peaking.
- FIG. 9 is a timing diagram showing PLL response testing using an applied phase step.
- FIG. 10 is a flow chart showing a high level control flow for determining PLL response.
- PLL response on-chip provides faster and more accurate measurement of PLL characteristics as compared to traditional approaches using off-chip test equipment.
- PLL response measurement includes measuring PLL bandwidth and PLL peaking.
- on-chip measurement provides more versatility in terms of being able to measure PLL response in a production environment, in a lab or debug environment, and even in the field.
- FIG. 1 illustrates a high level block diagram of a system that includes a phase-locked loop (PLL) 100 and a lock detector 101 .
- the PLL 100 includes on on-chip response measurement circuit 102 that receives the phase error 119 and supplies a maximum overshoot indication 104 and zero crossover time 106 , which can be used to determine the PLL response in terms of peaking and bandwidth as explained further herein.
- portions of the lock detector can be advantageously used in determining the PLL response on-chip.
- the PLL 100 receives a reference clock signal 103 , which can be supplied, e.g., from a crystal oscillator, a microelectromechanical structure (MEMS) oscillator, or other suitable low-jitter clock source.
- the digitally controlled oscillator (DCO) 110 supplies an output clock signal 105 .
- a feedback divider 107 receives the output clock signal 105 and supplies a feedback clock signal 109 .
- Time to digital converters (TDCs) 111 and 115 convert the reference clock signal 103 and the feedback clock signal 109 to digital representations of the clock signals, which are supplied, e.g., as time stamps (indicating the time a rising edge or falling edge (or both) occur) to a digital phase and frequency detector (PFD) 117 .
- PFD 117 supplies a phase error 119 indicating the phase difference between the reference clock signal 103 and the feedback clock signal 109 .
- a digital loop filter 121 and the lock detector 101 both receive the phase error from digital PFD
- Lock detector 101 also receives the time stamps from TDCs 111 and 115 .
- TDC 111 and TDC 115 are implemented by corresponding counter circuits.
- the lock detector uses these time stamps to measure the frequency offset between the clocks and provide the frequency lock status.
- the lock detector utilizes both time resolution and frequency resolution in determining whether the system is in lock or out of lock.
- the lock detector 101 provides a loss of lock indicator when there is either a phase loss of lock or a frequency loss of lock.
- the lock detector tracks the phase error information from the digital PFD 117 to determine if the PLL is phase locked.
- the frequency loss of lock indicates that the frequency offset between the reference clock signal 103 and the feedback signal 109 is larger than a frequency offset threshold.
- the phase loss of lock indicates whether the phase error is larger than a phase error threshold.
- the lock detector 101 provides a quantitative value of the frequency offset and the phase error.
- Most existing techniques provide either phase or frequency lock monitoring but not both. In addition, existing techniques do not address the stochastic behaviors of the reference clock and how to provide a faster unlock detection during those transient events and avoid any chattering of lock status for steady state operation.
- FIG. 2 illustrates a block diagram of the lock detector 101 .
- Lock detector 101 includes a frequency lock monitoring circuit 201 , phase lock monitoring circuit 203 , and a validate circuit 204 and timer-based loss of lock clearing circuit 205 that tracks the loop filter bandwidth settings. Additional details of the frequency lock monitoring circuit 201 , the validate circuit 204 , and the timer-based loss of lock clearing circuit 205 can be found in the application entitled “Accurate and Reliable Digital PLL Lock Indicator”, naming Kannanthodath V. Jayakumar et al. as inventors, application Ser. No. 16/580,161, filed Sep. 24, 2019, which application is incorporated herein by reference in its entirety.
- the phase lock monitoring circuit 203 can be used advantageously to implement part of the on-chip PLL response measurement circuit 102 (see FIG. 1 ).
- the phase lock monitoring circuit receives phase errors 119 that are oversampled in the digital PFD 117 (see FIG. 1 ).
- Integrator 231 integrates the samples to a rate corresponding to a non-oversampled PFD rate and supplies the integrated phase error 232 .
- Comparator 233 receives the integrated phase error 232 and compares the phase error 232 to a threshold phase error value and supplies a phase loss of lock indicator 235 indicative of the comparison along with the phase error amount 237 .
- a low pass filter 239 receives the phase error 232 and supplies an average phase error 241 . While FIG. 2 shows filter 239 as a second order filter, in other embodiment a higher order filter may be used to get better filtering of the noise present in the phase error information and hence more accurate detection of phase lock status. The choice of filter depends on system requirements.
- the comparator circuit 233 compares the average phase error 241 to a threshold average phase error value to determine a loss of lock condition.
- Comparisons of either or both of the average phase error 241 and the phase error 232 to respective thresholds may be used to determine when to assert the phase loss of lock signal 235 . As explained further herein, those thresholds can also be used to determine a zero crossing of the phase error and thus the crossover time as part of determining PLL bandwidth.
- FIG. 3 illustrates additional details of an embodiment of the phase monitoring circuit 203 .
- the phase monitoring circuit uses phase error information from the digital PFD circuit to detect whether the error is within the phase error threshold to determine if the PLL is phase locked.
- the phase error 119 provided by the digital PFD is at an interpolated PFD rate.
- the phase error information provided by the PFD is at N ⁇ PFD frequency, where N is the interpolation factor.
- the phase error that the PFD measures is distributed in smaller portions at higher speed to the lock detection circuit.
- the PFD sends the phase error in this way due to interface requirements with the loop filter.
- the phase error circuit 203 utilizes an integrate and dump circuit 301 to integrate the phase error 119 from the digital PFD and determine if the phase error between the feedback and reference clocks is within or has exceeded the phase lock threshold.
- the integrated and dump circuit accumulates samples over a chosen measurement time frame (N samples) and then “dumps” the resulting value.
- the saturation block 302 ensures the value of the PFD error does not exceed the limit and roll over from, e.g., a high value to a low value.
- the PFD provides the phase error information at N ⁇ PFD frequency.
- the integrator accumulates the phase error for N samples.
- the down sampler 303 dumps the integrated phase error at the PFD frequency, which is at the down sampled rate compared to the input rate.
- phase lock monitoring circuit 203 allows the phase lock monitoring circuit 203 to assert a loss of phase lock for single event transients of the reference or feedback clock that cause a phase error larger than the phase error threshold.
- the phase error detect circuit can be used for fast loss of lock triggering.
- the reference clock signal includes noise due to, e.g., jitter and wander and the feedback clock signal can also include noise and the threshold settings should accommodate an appropriate margin for such noise so as to not trigger too often due to transient events.
- the phase monitor circuit 203 also provides average phase error information.
- a phase error average circuit 305 uses the information from the integrate and dump circuit and passes it through a second order CIC filter. This filter has a programmable averaging window length (by adjusting the down sample ratio as desired) to adjust the window length. The CIC filter removes high frequency noise. As averaging is increased the phase error will approach DC (static) phase error information, which can be considered as the true phase lock condition.
- a selector circuit 309 selects between the single cycle phase error 304 or the average phase error 307 and supplies the selected phase error to comparator 311 .
- the comparator 311 supplies the phase lock indicator 315 after comparing the selected phase error to programmable thresholds to see if the phase error is greater than a set threshold causing the phase lock indicator to be asserted to indicate an out of lock condition or deasserted if the phase error is below a threshold to indicate a lock condition.
- the programmability of the thresholds allows the comparator to be used to help determine the zero crossings as described further herein.
- PLL loop response is specified by the closed loop transfer function of phase modulation.
- the PLL behaves as a low pass filter of reference modulation since its output follows its reference at low modulation frequency and cannot track high modulation frequencies.
- ⁇ in (t) 2 ⁇ f ref t+A in sin(2 ⁇ f m t), where ⁇ in (t) is the phase of the reference clock signal, f ref is the frequency of the reference clock, and f m is the reference clock modulation frequency.
- ⁇ out (t, f m ) N(2 ⁇ f ref t)+A out (f m )sin(2 ⁇ f m t+ ⁇ m (f m )), where ⁇ out (t, f m ) is the phase of the DCO output signal, N is the feedback divisor, A out (f m ) and ⁇ m (f m ) are the magnitude and phase response functions of the reference clock modulation.
- the PLL response is summarized by its bandwidth and peaking characteristics.
- the PLL bandwidth measured at the ⁇ 3 dB point is chosen by balancing the effects of reference input noise and internally generated PLL noise to achieve the lowest PLL output clock jitter.
- Lower bandwidth attenuates more noise in the reference clock input at the expense of rejecting less noise generated by PLL circuitry while higher bandwidth achieves the opposite.
- Maximum value in the transfer function is referred to as the gain peaking. Higher peaking is undesirable from a jitter perspective since a PLL amplifies the phase modulation at the peaking frequency.
- FIG. 4 illustrates PLL jitter transfer functions with different bandwidths and different peaking.
- CL(z) is the closed loop magnitude in dB and the x-axis is the normalized frequency of the reference clock in Hz.
- Transfer function 401 shows larger peaking (lower damping) and lower bandwidth relative to transfer function 403 , which shows less peaking (higher damping) and a wider bandwidth.
- FIG. 4 illustrates that different PLL configurations can provide different bandwidth and different peaking characteristics for identical reference clock and output frequencies.
- FIG. 4 shows the PLL transfer response has a low pass characteristic for the input clock phase modulation.
- the on-chip response measurement circuit can determine if actual results on-chip match the expected results such as those shown in FIG. 4 .
- FIG. 5A shows a typical PLL transient response to an induced input phase step at time 0 .
- the phase step is being provided in accordance with embodiments described herein.
- the PLL tends to overcorrect (or overshoot) as it tries to eliminate the added phase step.
- the PLL simulation model was configured to be underdamped and as a result exhibits more ringing to show the ringing effects more clearly.
- a key metric in the PLL step response is T crossover 501 defined as the elapsed time from when the input step is applied to the onset of the initial phase overshoot at the zero crossing 502 .
- Another key metric shown here is the maximum overshoot 503 , which indicates the maximum overcorrection to the applied phase step.
- FIG. 5B shows another example of PLL response to a phase step showing that after the phase step, the system eventually eliminates the added phase step.
- FIG. 6A shows additional details of the on-chip circuit to measure PLL response.
- Embodiments may also include the lock detector 101 shown in FIG. 1 .
- the on-chip response measurement circuit performs time-domain measurements of PLL output phase in response to an induced phase step. These measurements are correlated to bandwidth and peaking in the frequency domain.
- the phase of the reference clock is instantaneously advanced (phase step) and the resultant phase error transient is recorded.
- phase step a digital phase offset is stored in storage 601 and the digital phase offset, which can be a positive or negative offset, is supplied to a summer 603 . Note that rather inserting the phase step in the digital representation of the reference clock signal, other embodiments insert the phase offset into the digital representation of the feedback clock signal.
- the summer 603 is bypassed for the remainder of the test using switches or an offset code of zero is supplied to the summing circuit 603 .
- PFD 117 supplies digital phase error information 119 to the integrate and dump circuit 605 .
- the integrate and dump circuit 605 integrates the phase error 119 from the digital PFD 117 .
- the integrated and dump circuit 605 accumulates samples over a chosen measurement time frame (N samples) and then “dumps” the resulting value.
- the saturation block 607 ensures the value of the PFD error does not exceed the limit and roll over from, e.g., a high value to a low value.
- the PFD provides the phase error information at N ⁇ PFD frequency.
- the integrator accumulates the phase error for N samples.
- the down sampler 609 dumps the integrated phase error at the down sampled rate.
- the integrated and dump circuit 301 shown in FIG. 3 may be used during the PLL response testing.
- the integrate and dump circuit 605 can be omitted depending on the particular PFD configuration and sampling approach (e.g., the PFD supplies the phase error at the PFD frequency not an interpolated rate), the integrate and dump circuit 605 can be omitted.
- the maximum phase error circuit 615 and the zero crossing detection circuit 617 receive the phase error 611 supplied by integrate and dump circuit 605 .
- a test enable signal 619 starts the PLL response testing. On assertion of the test enable signal 619 , the phase offset is inserted into the reference clock digital representation (or the feedback clock signal).
- counter 621 Concurrent with insertion of the phase step, counter 621 begins counting.
- the zero crossing detector is looking for the phase error to go to zero at 502 the first time after insertion of the phase step.
- a comparator 631 shown in FIG. 6B implements this function by setting a compare threshold 633 to a value corresponding to zero to compare to the phase error. The phase error going above that threshold value for the first time indicates the first zero crossing 635 . Detecting the zero crossing causes the count value of counter 621 to be stored in storage 623 (or is stored by stopping counter 621 ) and that count value is used to determine PLL bandwidth.
- the maximum phase error logic circuit 615 determines the maximum overshoot by comparing the current phase error to a current maximum phase error being stored and replaces the current maximum if the current phase error is greater than the current maximum.
- the maximum phase error logic begins to monitor the current phase error after the zero crossing event has occurred.
- the maximum phase error logic stops monitoring the phase error at the end of the measurement cycle (end of test) once a steady state lock has been achieved indicating the phase error has settled and converged to near zero after inducing the phase step for the measurement or after a predetermined time period after the zero crossing or a predetermined time period after enable test 619 is asserted.
- the average phase error 307 may be used to determine end of the test.
- the maximum phase error is the maximum overshoot used to determine PLL peaking.
- the maximum overshoot and zero crossover time can be used to determine the PLL response based on the relation of the zero crossover time to the reciprocal of bandwidth and based on the relation of maximum phase error overshoot to PLL peaking.
- Dedicated circuits can be used to determine the PLL response, or a central processing unit (CPU) in a microcontroller unit (MCU) or other processor on the integrated circuit can be used to process the maximum overshoot and crossover time data to characterize the PLL response.
- Bandwidth of a second order type 2 PLL can be defined as shown below in equation (1) where ⁇ 3db is the 3-dB bandwidth and ⁇ n is the natural frequency of the PLL and ⁇ is the damping factor.
- ⁇ 3db ⁇ n (2 ⁇ 2 +1+ ⁇ square root over (((2 ⁇ 2 +1) 2 +1))) ⁇ 1/2 (1)
- the length of time to crossover (T crossover ) during an induced phase step is inversely related to the bandwidth of the PLL.
- the crossover time (T crossover ) is shown to be linearly proportional to the reciprocal of the PLL's 3-dB closed loop bandwidth as shown in the simulation results shown in FIG. 7 .
- FIG. 7 shows that the lower the bandwidth, the higher the T crossover .
- T crossover is not very sensitive to the magnitude of the phase step.
- FIG. 7 illustrates that the crossover time is directly proportional to the inverse of band width provided all the other PLL parameters are held constant. For example, the slope of the graph changes when the damping factor is different. In one example, the crossover time is approximately equal to 2.5/BW for a damping factor of 1.
- Simulations and loop equations also predict maximum overshoot is proportional to the maximum peaking in the closed loop transfer function. In particular, the larger the overshoot, the greater the peaking. The magnitude of the overshoot is also proportional to the magnitude of the phase step. Maximum overshoot or peak output value during an induced phase step can be defined approximately as shown below.
- Max ⁇ ⁇ overshoot phase step * ( 1 + exp ⁇ ( - ⁇ 1 - ⁇ 2 ) * T p * ( cos ⁇ T p - ⁇ 1 - ⁇ 2 ⁇ sin ⁇ T p ) ( 2 )
- FIG. 8 shows simulation results of the response for phase steps of two different magnitudes and shows the directly proportional relationship between the maximum phase overshoot to the PLL closed loop jitter peaking with the set of other PLL parameters kept constant in simulation.
- the response 801 is the best fit line for a first phase step of 85% with respect to the reference clock, i.e., 85% of the period of the reference clock.
- the response 803 illustrates the best fit line for a phase step of 50% with respect to the reference clock.
- FIG. 6A shows a separate on-chip response measurement circuit 102
- on-chip capabilities described in FIGS. 1, 2, and 3 can be used in the PLL response testing.
- the input reference clock is transferred into a time code representation by the Time to Digital converter (TDC).
- TDC Time to Digital converter
- the time codes from the TDC are passed through an adder/subtractor circuit 603 (see FIG. 6A ) before it reaches the digital PFD and lock detector circuit 101 (see FIG. 1 ).
- the test is initiated by asserting the test enable signal 619 shown in FIG. 6A , which in turn activates the required circuits needed to measure the PLL transient responses for an induced phase step.
- test enable signal 619 enables a timecode value to be added to or subtracted from the actual time code from the TDC that represents the reference clock, thereby creating a phase step when seen from digital PFD 117 and lock detector 101 .
- the induced phase step is seen by the PFD as phase error between reference and feedback clocks and the PLL responds to the phase step based on its characteristics.
- the response of the PLL is continuously monitored and captured by the lock detector.
- the phase error information is processed by the phase lock monitoring circuit 203 (see FIG. 3 ) and the phase lock monitoring circuit 203 supplies information used to determine the bandwidth and maximum overshoot produced by the PLL under test.
- the phase lock phase lock monitoring circuit 203 has programmable thresholds to determine if the phase error in the PLL loop is greater or lower than the limits. That is achieved in one embodiment by setting a trigger threshold to determine if the phase error introduced in the loop is greater than the set limit (out of lock) and a clear threshold that causes the lock indication to assert if the phase error is below the clear threshold.
- the circuit has an internal counter (not shown in FIG. 3 ) that starts counting in response to a phase error greater than a trigger threshold (in response to phase loss of lock) and stops counting when the phase error is less than the clear threshold indicating the phase is locked.
- those thresholds can be used by setting the trigger threshold at just below the phase step and setting the clear threshold close to zero.
- the internal counter in the phase lock monitoring circuit can be used to start counting on application of the phase step and stop counting when the phase error crosses zero.
- the counter runs on an internally calibrated oscillator which allows conversion of the count value to time representation.
- other embodiments may use dedicated circuits for the on-chip response measurement circuit or a different combination of available circuits and dedicated circuits.
- FIG. 9 shows the timing diagram illustrating the PLL response test operation.
- FIG. 9 shows the reference clock (ref) and feedback clock (fb) with their associated time stamp information in time codes.
- the phase step time code is also shown.
- the phase step time code is summed into the reference time code once at 901 . That introduces a phase step into the loop and PLL operates to eliminate the introduced phase error.
- FIG. 9 also shows the effect at 903 on the reference clock as well as the assertion of the test enable signal at 905 .
- FIG. 9 illustrates the phase error in time codes. Note that the shape of the phase error shown in FIG. 9 may look like the phase error shown in FIGS. 5A and 5B but depends on the damping factor of the PLL.
- FIG. 10 illustrates a high level flow chart for on-chip measurement of the PLL response.
- the PLL test is enabled resulting in the phase step being inserted and the counter for the zero crossing detector beginning its count.
- the zero crossing detector evaluates the phase information and looks for the phase error zero crossover in 1003 .
- the zero crossing may be detected, e.g., using the comparator to determine the first zero crossing.
- the test waits for the phase error to be zero in 1003 and on detection of the zero crossing, in 1005 stores the counter value at the zero crossing. Detection of the zero crossover causes the maximum error detector to begin looking for the maximum phase error.
- the logic looks to see if a current phase error is larger than the current maximum and if so, replaces the current maximum.
- the maximum phase error logic checks if the test is over based on, e.g., a predetermined number of phase errors being less than a threshold (or average phase error being less than a threshold), a predetermined time period has expired, or some other indicator that the maximum phase error has already been detected. If the test is over the maximum phase error is the maximum overshoot and is available to determine PLL peaking and the zero crossover time is available to calculate the bandwidth.
- an on-board microcontroller 108 (see FIG. 1 ) is programmed to calculate the bandwidth based on the linear relationship shown, e.g., in FIG. 7 .
- the on-board microcontroller can calculate PLL peaking based on the relationships shown in FIG.
- the settling time, peak undershoot, and the time difference between consecutive peaks of the transient response can be used to further characterize the PLL. That information can be used to determine the damping factor of the PLL, natural frequency of the PLL, and settling time for the PLL.
- the zero crossover time and the maximum phase error overshoot is provided off-chip for additional processing.
Landscapes
- Stabilization Of Oscillater, Synchronisation, Frequency Synthesizers (AREA)
Abstract
An integrated circuit includes an on-chip PLL response measurement capability. The PLL response is determined in terms of PLL bandwidth and PLL peaking. A digital phase offset is inserted to a digital representation of a first clock signal to create a phase step. A phase and frequency detector of a phase-locked loop (PLL) supplies a phase error signal indicative of a difference between the first clock signal and a second clock signal. The elapsed time between the phase step insertion and the first zero crossing of the phase error as the PLL tries to deal with the is used to determine PLL bandwidth. The maximum phase error overshoot resulting from insertion of the digital phase offset is determined for use in determining PLL peaking.
Description
- This invention relates to phase-locked loops (PLLs) and more particularly to measuring PLL response characteristics.
- The PLL closed loop transfer function is often measured on a tester using a sinusoidal signal generator to modulate the reference clock signal and a spectrum analyzer/oscilloscope to measure the PLL response by sending the PLL output clock signal off-chip to the test equipment. Such an approach to measuring PLL characteristics is time consuming requiring many seconds and even minutes to complete the testing and uses a complicated external test setup. The external measurement circuit often influences the test results and cause the test results to deviate from the actual PLL closed loop transfer function. Accordingly, improved techniques to measure PLL characteristics would be desirable.
- In an embodiment, a method for determining response of a phase-locked loop (PLL) includes inserting a digital phase offset into a digital representation of a clock signal coupled to a phase and frequency detector (PFD) of the PLL, the digital phase offset representing a phase step. The method further includes determining a length of time between insertion of the digital phase offset and a first zero crossing of a phase error supplied from the PFD after insertion of the digital phase offset, the length of time being inversely related to bandwidth of the PLL. The method may further include determining a maximum phase error overshoot resulting from insertion of the digital phase offset and determining a peaking characteristic of the PLL based in part on the maximum phase error overshoot.
- In another embodiment an integrated circuit includes a time to digital converter to generate a digital representation of a clock signal. A summing circuit adds a digital phase offset to the digital representation of the clock signal, the digital phase offset representing a phase step. A counter counts a time period beginning when the digital phase offset is inserted. A first zero crossing detection circuit detects a first zero crossing of a phase error between the clock signal and a second clock signal after insertion of the digital phase offset and supplies a zero crossing indication. A count value of the counter is stored responsive to assertion of the zero crossing indication.
- In another embodiment, an integrated circuit includes a summing circuit to add a digital phase offset to a digital representation of a first clock signal, the digital phase offset corresponding a phase step in the first clock signal. A phase and frequency detector of a phase-locked loop (PLL) receives the digital representation of the first clock signal and a digital representation of a second clock signal and supplies a phase error signal indicative of a difference between the first clock signal and the second clock signal. A first zero crossing detection circuit receives the phase error and detects a first zero crossing of the phase error after insertion of the phase offset. A counter determines an elapsed time between insertion of the phase offset and the first zero crossing. A maximum phase error circuit receives the phase error signal and to detect a maximum phase error overshoot resulting from insertion of the digital phase offset.
- The present invention may be better understood, and its numerous objects, features, and advantages made apparent to those skilled in the art by referencing the accompanying drawings.
-
FIG. 1 illustrates a high level block diagram of a system that includes on-chip PLL response measurement capability. -
FIG. 2 illustrates a functional block diagram of a lock detector circuit that includes a frequency lock monitoring circuit and a phase lock monitoring circuit. -
FIG. 3 illustrates a functional block diagram of an embodiment of a phase lock monitoring circuit. -
FIG. 4 illustrates PLL jitter transfer with different PLL bandwidths and peaking. -
FIG. 5A illustrates PLL phase error response to a phase step. -
FIG. 5B illustrates PLL phase error response to a phase step showing the phase error ringing settling to a steady state. -
FIG. 6A illustrates a functional block diagram of an embodiment of an on-chip PLL response measurement circuit. -
FIG. 6B shows a comparator used as a zero crossing detector. -
FIG. 7 illustrates a graph of crossover time versus inverse of closedloop 3 dB bandwidth. -
FIG. 8 illustrates a graph of maximum phase error overshoot versus closed loop peaking. -
FIG. 9 is a timing diagram showing PLL response testing using an applied phase step. -
FIG. 10 is a flow chart showing a high level control flow for determining PLL response. - The use of the same reference symbols in different drawings indicates similar or identical items.
- Measuring PLL response on-chip provides faster and more accurate measurement of PLL characteristics as compared to traditional approaches using off-chip test equipment. PLL response measurement includes measuring PLL bandwidth and PLL peaking. In addition, on-chip measurement provides more versatility in terms of being able to measure PLL response in a production environment, in a lab or debug environment, and even in the field.
-
FIG. 1 illustrates a high level block diagram of a system that includes a phase-locked loop (PLL) 100 and alock detector 101. ThePLL 100 includes on on-chipresponse measurement circuit 102 that receives thephase error 119 and supplies a maximumovershoot indication 104 and zerocrossover time 106, which can be used to determine the PLL response in terms of peaking and bandwidth as explained further herein. In an embodiment, portions of the lock detector can be advantageously used in determining the PLL response on-chip. - The
PLL 100 receives areference clock signal 103, which can be supplied, e.g., from a crystal oscillator, a microelectromechanical structure (MEMS) oscillator, or other suitable low-jitter clock source. The digitally controlled oscillator (DCO) 110 supplies anoutput clock signal 105. Afeedback divider 107 receives theoutput clock signal 105 and supplies afeedback clock signal 109. Time to digital converters (TDCs) 111 and 115 convert thereference clock signal 103 and thefeedback clock signal 109 to digital representations of the clock signals, which are supplied, e.g., as time stamps (indicating the time a rising edge or falling edge (or both) occur) to a digital phase and frequency detector (PFD) 117. PFD 117 supplies aphase error 119 indicating the phase difference between thereference clock signal 103 and thefeedback clock signal 109. Adigital loop filter 121 and thelock detector 101 both receive the phase error fromdigital PFD 117. -
Lock detector 101 also receives the time stamps from TDCs 111 and 115. In at least oneembodiment TDC 111 and TDC 115 are implemented by corresponding counter circuits. The lock detector uses these time stamps to measure the frequency offset between the clocks and provide the frequency lock status. The lock detector utilizes both time resolution and frequency resolution in determining whether the system is in lock or out of lock. - In an embodiment, the
lock detector 101 provides a loss of lock indicator when there is either a phase loss of lock or a frequency loss of lock. The lock detector tracks the phase error information from thedigital PFD 117 to determine if the PLL is phase locked. The frequency loss of lock indicates that the frequency offset between thereference clock signal 103 and thefeedback signal 109 is larger than a frequency offset threshold. The phase loss of lock indicates whether the phase error is larger than a phase error threshold. In addition, thelock detector 101 provides a quantitative value of the frequency offset and the phase error. Most existing techniques provide either phase or frequency lock monitoring but not both. In addition, existing techniques do not address the stochastic behaviors of the reference clock and how to provide a faster unlock detection during those transient events and avoid any chattering of lock status for steady state operation. -
FIG. 2 illustrates a block diagram of thelock detector 101.Lock detector 101 includes a frequencylock monitoring circuit 201, phaselock monitoring circuit 203, and a validatecircuit 204 and timer-based loss oflock clearing circuit 205 that tracks the loop filter bandwidth settings. Additional details of the frequencylock monitoring circuit 201, the validatecircuit 204, and the timer-based loss oflock clearing circuit 205 can be found in the application entitled “Accurate and Reliable Digital PLL Lock Indicator”, naming Kannanthodath V. Jayakumar et al. as inventors, application Ser. No. 16/580,161, filed Sep. 24, 2019, which application is incorporated herein by reference in its entirety. - The phase
lock monitoring circuit 203 can be used advantageously to implement part of the on-chip PLL response measurement circuit 102 (seeFIG. 1 ). The phase error/time error of a clock is the difference between the time indicated by the clock under test and a reference clock which can be denoted as phase error (t)=Tfb(t)−Tref(t), where Tfb is the time of the feedback clock provided as a time stamp and Tref is the time of the reference clock provided as a time stamp. In an embodiment the phase lock monitoring circuit receivesphase errors 119 that are oversampled in the digital PFD 117 (seeFIG. 1 ).Integrator 231 integrates the samples to a rate corresponding to a non-oversampled PFD rate and supplies theintegrated phase error 232. -
Comparator 233 receives theintegrated phase error 232 and compares thephase error 232 to a threshold phase error value and supplies a phase loss oflock indicator 235 indicative of the comparison along with thephase error amount 237. In addition, in embodiments, alow pass filter 239 receives thephase error 232 and supplies anaverage phase error 241. WhileFIG. 2 shows filter 239 as a second order filter, in other embodiment a higher order filter may be used to get better filtering of the noise present in the phase error information and hence more accurate detection of phase lock status. The choice of filter depends on system requirements. Thecomparator circuit 233 compares theaverage phase error 241 to a threshold average phase error value to determine a loss of lock condition. Comparisons of either or both of theaverage phase error 241 and thephase error 232 to respective thresholds may be used to determine when to assert the phase loss oflock signal 235. As explained further herein, those thresholds can also be used to determine a zero crossing of the phase error and thus the crossover time as part of determining PLL bandwidth. -
FIG. 3 illustrates additional details of an embodiment of thephase monitoring circuit 203. The phase monitoring circuit uses phase error information from the digital PFD circuit to detect whether the error is within the phase error threshold to determine if the PLL is phase locked. Thephase error 119 provided by the digital PFD is at an interpolated PFD rate. In an embodiment, the phase error information provided by the PFD is at N×PFD frequency, where N is the interpolation factor. The phase error that the PFD measures is distributed in smaller portions at higher speed to the lock detection circuit. The PFD sends the phase error in this way due to interface requirements with the loop filter. Thephase error circuit 203 utilizes an integrate and dumpcircuit 301 to integrate thephase error 119 from the digital PFD and determine if the phase error between the feedback and reference clocks is within or has exceeded the phase lock threshold. The integrated and dump circuit accumulates samples over a chosen measurement time frame (N samples) and then “dumps” the resulting value. Thesaturation block 302 ensures the value of the PFD error does not exceed the limit and roll over from, e.g., a high value to a low value. As discussed earlier, the PFD provides the phase error information at N×PFD frequency. The integrator accumulates the phase error for N samples. The downsampler 303 dumps the integrated phase error at the PFD frequency, which is at the down sampled rate compared to the input rate. That allows the phaselock monitoring circuit 203 to assert a loss of phase lock for single event transients of the reference or feedback clock that cause a phase error larger than the phase error threshold. In order to deal with the stochastic nature of the input clock, the phase error detect circuit can be used for fast loss of lock triggering. However, the reference clock signal includes noise due to, e.g., jitter and wander and the feedback clock signal can also include noise and the threshold settings should accommodate an appropriate margin for such noise so as to not trigger too often due to transient events. - In one or more embodiments, the
phase monitor circuit 203 also provides average phase error information. A phase erroraverage circuit 305 uses the information from the integrate and dump circuit and passes it through a second order CIC filter. This filter has a programmable averaging window length (by adjusting the down sample ratio as desired) to adjust the window length. The CIC filter removes high frequency noise. As averaging is increased the phase error will approach DC (static) phase error information, which can be considered as the true phase lock condition. In an embodiment, aselector circuit 309 selects between the singlecycle phase error 304 or theaverage phase error 307 and supplies the selected phase error tocomparator 311. Thecomparator 311 supplies thephase lock indicator 315 after comparing the selected phase error to programmable thresholds to see if the phase error is greater than a set threshold causing the phase lock indicator to be asserted to indicate an out of lock condition or deasserted if the phase error is below a threshold to indicate a lock condition. The programmability of the thresholds allows the comparator to be used to help determine the zero crossings as described further herein. - PLL loop response is specified by the closed loop transfer function of phase modulation. The PLL behaves as a low pass filter of reference modulation since its output follows its reference at low modulation frequency and cannot track high modulation frequencies. In particular:
- θin(t)=2πfreft+Ain sin(2πfmt), where θin(t) is the phase of the reference clock signal, fref is the frequency of the reference clock, and fm is the reference clock modulation frequency.
θout(t, fm)=N(2πfreft)+Aout(fm)sin(2πfmt+θm(fm)), where θout(t, fm) is the phase of the DCO output signal, N is the feedback divisor, Aout(fm) and θm(fm) are the magnitude and phase response functions of the reference clock modulation. - The resulting magnitude response of the PLL transfer function is therefore
-
- which is the normalized feedback divisor to account for the frequency and phase multiplying action of the PLL.
- The PLL response is summarized by its bandwidth and peaking characteristics. The PLL bandwidth measured at the −3 dB point is chosen by balancing the effects of reference input noise and internally generated PLL noise to achieve the lowest PLL output clock jitter. Lower bandwidth attenuates more noise in the reference clock input at the expense of rejecting less noise generated by PLL circuitry while higher bandwidth achieves the opposite. Maximum value in the transfer function is referred to as the gain peaking. Higher peaking is undesirable from a jitter perspective since a PLL amplifies the phase modulation at the peaking frequency.
-
FIG. 4 illustrates PLL jitter transfer functions with different bandwidths and different peaking. CL(z) is the closed loop magnitude in dB and the x-axis is the normalized frequency of the reference clock in Hz.Transfer function 401 shows larger peaking (lower damping) and lower bandwidth relative totransfer function 403, which shows less peaking (higher damping) and a wider bandwidth.FIG. 4 illustrates that different PLL configurations can provide different bandwidth and different peaking characteristics for identical reference clock and output frequencies.FIG. 4 shows the PLL transfer response has a low pass characteristic for the input clock phase modulation. The on-chip response measurement circuit can determine if actual results on-chip match the expected results such as those shown inFIG. 4 . -
FIG. 5A shows a typical PLL transient response to an induced input phase step attime 0. The phase step is being provided in accordance with embodiments described herein. The PLL tends to overcorrect (or overshoot) as it tries to eliminate the added phase step. InFIG. 5 , the PLL simulation model was configured to be underdamped and as a result exhibits more ringing to show the ringing effects more clearly. A key metric in the PLL step response isT crossover 501 defined as the elapsed time from when the input step is applied to the onset of the initial phase overshoot at the zerocrossing 502. Another key metric shown here is themaximum overshoot 503, which indicates the maximum overcorrection to the applied phase step.FIG. 5B shows another example of PLL response to a phase step showing that after the phase step, the system eventually eliminates the added phase step. -
FIG. 6A shows additional details of the on-chip circuit to measure PLL response. Embodiments may also include thelock detector 101 shown inFIG. 1 . The on-chip response measurement circuit performs time-domain measurements of PLL output phase in response to an induced phase step. These measurements are correlated to bandwidth and peaking in the frequency domain. The phase of the reference clock is instantaneously advanced (phase step) and the resultant phase error transient is recorded. In order to insert the phase step, a digital phase offset is stored instorage 601 and the digital phase offset, which can be a positive or negative offset, is supplied to asummer 603. Note that rather inserting the phase step in the digital representation of the reference clock signal, other embodiments insert the phase offset into the digital representation of the feedback clock signal. Once the offset is inserted, thesummer 603 is bypassed for the remainder of the test using switches or an offset code of zero is supplied to the summingcircuit 603.PFD 117 supplies digitalphase error information 119 to the integrate and dumpcircuit 605. As explained in relation toFIG. 3 , the integrate and dumpcircuit 605 integrates thephase error 119 from thedigital PFD 117. The integrated anddump circuit 605 accumulates samples over a chosen measurement time frame (N samples) and then “dumps” the resulting value. Thesaturation block 607 ensures the value of the PFD error does not exceed the limit and roll over from, e.g., a high value to a low value. As discussed earlier the PFD provides the phase error information at N×PFD frequency. The integrator accumulates the phase error for N samples. The downsampler 609 dumps the integrated phase error at the down sampled rate. In other embodiments, the integrated anddump circuit 301 shown inFIG. 3 may be used during the PLL response testing. In still other embodiments, depending on the particular PFD configuration and sampling approach (e.g., the PFD supplies the phase error at the PFD frequency not an interpolated rate), the integrate and dumpcircuit 605 can be omitted. The maximumphase error circuit 615 and the zerocrossing detection circuit 617 receive thephase error 611 supplied by integrate and dumpcircuit 605. A test enablesignal 619 starts the PLL response testing. On assertion of the test enablesignal 619, the phase offset is inserted into the reference clock digital representation (or the feedback clock signal). Concurrent with insertion of the phase step,counter 621 begins counting. With reference toFIG. 5A , the zero crossing detector is looking for the phase error to go to zero at 502 the first time after insertion of the phase step. In an embodiment acomparator 631 shown inFIG. 6B implements this function by setting a comparethreshold 633 to a value corresponding to zero to compare to the phase error. The phase error going above that threshold value for the first time indicates the firstzero crossing 635. Detecting the zero crossing causes the count value ofcounter 621 to be stored in storage 623 (or is stored by stopping counter 621) and that count value is used to determine PLL bandwidth. - The maximum phase
error logic circuit 615 determines the maximum overshoot by comparing the current phase error to a current maximum phase error being stored and replaces the current maximum if the current phase error is greater than the current maximum. The maximum phase error logic begins to monitor the current phase error after the zero crossing event has occurred. The maximum phase error logic stops monitoring the phase error at the end of the measurement cycle (end of test) once a steady state lock has been achieved indicating the phase error has settled and converged to near zero after inducing the phase step for the measurement or after a predetermined time period after the zero crossing or a predetermined time period after enabletest 619 is asserted. For example, the average phase error 307 (seeFIG. 3 ) may be used to determine end of the test. The maximum phase error is the maximum overshoot used to determine PLL peaking. - The maximum overshoot and zero crossover time can be used to determine the PLL response based on the relation of the zero crossover time to the reciprocal of bandwidth and based on the relation of maximum phase error overshoot to PLL peaking. Dedicated circuits can be used to determine the PLL response, or a central processing unit (CPU) in a microcontroller unit (MCU) or other processor on the integrated circuit can be used to process the maximum overshoot and crossover time data to characterize the PLL response.
- Bandwidth of a
second order type 2 PLL can be defined as shown below in equation (1) where ω3db is the 3-dB bandwidth and ωn is the natural frequency of the PLL and ζ is the damping factor. -
ω3db=ωn(2ζ2+1+√{square root over (((2ζ2+1)2+1)))}1/2 (1) - For a PLL with a defined damping factor and natural frequency we have a theoretical 3-dB loop bandwidth. The natural frequency of the PLL ωn is inversely related to PLL time constant loop filter
-
- Thus, the length of time to crossover (Tcrossover) during an induced phase step is inversely related to the bandwidth of the PLL. Referring to
FIG. 7 , the crossover time (Tcrossover) is shown to be linearly proportional to the reciprocal of the PLL's 3-dB closed loop bandwidth as shown in the simulation results shown inFIG. 7 .FIG. 7 shows that the lower the bandwidth, the higher the Tcrossover. Note that Tcrossover is not very sensitive to the magnitude of the phase step.FIG. 7 illustrates that the crossover time is directly proportional to the inverse of band width provided all the other PLL parameters are held constant. For example, the slope of the graph changes when the damping factor is different. In one example, the crossover time is approximately equal to 2.5/BW for a damping factor of 1. - Simulations and loop equations also predict maximum overshoot is proportional to the maximum peaking in the closed loop transfer function. In particular, the larger the overshoot, the greater the peaking. The magnitude of the overshoot is also proportional to the magnitude of the phase step. Maximum overshoot or peak output value during an induced phase step can be defined approximately as shown below.
-
- where Tp=ωn√{square root over (1−ζ2)}. Here the phasestep is the induced phase step to the PLL, ωn is the natural frequency of the PLL and ζ is the damping factor.
FIG. 8 shows simulation results of the response for phase steps of two different magnitudes and shows the directly proportional relationship between the maximum phase overshoot to the PLL closed loop jitter peaking with the set of other PLL parameters kept constant in simulation. Theresponse 801 is the best fit line for a first phase step of 85% with respect to the reference clock, i.e., 85% of the period of the reference clock. Theresponse 803 illustrates the best fit line for a phase step of 50% with respect to the reference clock. - Although
FIG. 6A shows a separate on-chipresponse measurement circuit 102, in embodiments, on-chip capabilities described inFIGS. 1, 2, and 3 can be used in the PLL response testing. The input reference clock is transferred into a time code representation by the Time to Digital converter (TDC). The time codes from the TDC are passed through an adder/subtractor circuit 603 (seeFIG. 6A ) before it reaches the digital PFD and lock detector circuit 101 (seeFIG. 1 ). The test is initiated by asserting the test enablesignal 619 shown inFIG. 6A , which in turn activates the required circuits needed to measure the PLL transient responses for an induced phase step. - Assertion of the test enable
signal 619 enables a timecode value to be added to or subtracted from the actual time code from the TDC that represents the reference clock, thereby creating a phase step when seen fromdigital PFD 117 andlock detector 101. The induced phase step is seen by the PFD as phase error between reference and feedback clocks and the PLL responds to the phase step based on its characteristics. The response of the PLL is continuously monitored and captured by the lock detector. The phase error information is processed by the phase lock monitoring circuit 203 (seeFIG. 3 ) and the phaselock monitoring circuit 203 supplies information used to determine the bandwidth and maximum overshoot produced by the PLL under test. - The phase lock phase
lock monitoring circuit 203 has programmable thresholds to determine if the phase error in the PLL loop is greater or lower than the limits. That is achieved in one embodiment by setting a trigger threshold to determine if the phase error introduced in the loop is greater than the set limit (out of lock) and a clear threshold that causes the lock indication to assert if the phase error is below the clear threshold. The circuit has an internal counter (not shown inFIG. 3 ) that starts counting in response to a phase error greater than a trigger threshold (in response to phase loss of lock) and stops counting when the phase error is less than the clear threshold indicating the phase is locked. Thus, those thresholds can be used by setting the trigger threshold at just below the phase step and setting the clear threshold close to zero. In that the way the internal counter in the phase lock monitoring circuit can be used to start counting on application of the phase step and stop counting when the phase error crosses zero. In an embodiment, the counter runs on an internally calibrated oscillator which allows conversion of the count value to time representation. Of course, other embodiments may use dedicated circuits for the on-chip response measurement circuit or a different combination of available circuits and dedicated circuits. -
FIG. 9 shows the timing diagram illustrating the PLL response test operation.FIG. 9 shows the reference clock (ref) and feedback clock (fb) with their associated time stamp information in time codes. The phase step time code is also shown. The phase step time code is summed into the reference time code once at 901. That introduces a phase step into the loop and PLL operates to eliminate the introduced phase error.FIG. 9 also shows the effect at 903 on the reference clock as well as the assertion of the test enable signal at 905. Finally,FIG. 9 illustrates the phase error in time codes. Note that the shape of the phase error shown inFIG. 9 may look like the phase error shown inFIGS. 5A and 5B but depends on the damping factor of the PLL. - The processed phase error information from the lock detector or dedicated circuits can be streamed out continuously to an internal MCU memory for post processing by an MCU or other on-board processor to determine the zero crossover time and the maximum overshoot rather than using dedicated circuitry to determine the maximum overshoot and zero crossover time in real time. Additionally, the programmed processor can determine the bandwidth from the zero crossover time and PLL peaking from the maximum overshoot. That data can be provided like other diagnostic or error data off-chip through a serial port or other reporting path. Additional characterization of the PLL can also be determined for the PLL under test such as settling time (see settling
time 505 inFIG. 5B ), peak undershoot (maximum negative phase error after the first zero crossing), and the time difference between consecutive peaks (e.g., 507, 509, and 511 inFIG. 5B ) of the transient response. That information can be used to determine the damping factor of the PLL, natural frequency of the PLL, and settling time for the PLL. Dedicated circuitry can be used or an on-board programmed processor can be used to determine various characteristics of the PLL described herein. -
FIG. 10 illustrates a high level flow chart for on-chip measurement of the PLL response. At 1001 the PLL test is enabled resulting in the phase step being inserted and the counter for the zero crossing detector beginning its count. The zero crossing detector evaluates the phase information and looks for the phase error zero crossover in 1003. The zero crossing may be detected, e.g., using the comparator to determine the first zero crossing. The test waits for the phase error to be zero in 1003 and on detection of the zero crossing, in 1005 stores the counter value at the zero crossing. Detection of the zero crossover causes the maximum error detector to begin looking for the maximum phase error. In 1007 the logic looks to see if a current phase error is larger than the current maximum and if so, replaces the current maximum. In 1009 the maximum phase error logic checks if the test is over based on, e.g., a predetermined number of phase errors being less than a threshold (or average phase error being less than a threshold), a predetermined time period has expired, or some other indicator that the maximum phase error has already been detected. If the test is over the maximum phase error is the maximum overshoot and is available to determine PLL peaking and the zero crossover time is available to calculate the bandwidth. In embodiments, an on-board microcontroller 108 (seeFIG. 1 ) is programmed to calculate the bandwidth based on the linear relationship shown, e.g., inFIG. 7 . In addition, the on-board microcontroller can calculate PLL peaking based on the relationships shown inFIG. 8 or another curve suitable for the particular size of the phase step. In addition to characterizing the PLL based on bandwidth and peaking, if the phase error is stored, or if additional dedicated circuitry is provided, the settling time, peak undershoot, and the time difference between consecutive peaks of the transient response can be used to further characterize the PLL. That information can be used to determine the damping factor of the PLL, natural frequency of the PLL, and settling time for the PLL. In embodiments, if an on-board processor is not available, the zero crossover time and the maximum phase error overshoot is provided off-chip for additional processing. - Thus, techniques have been described for determining PLL response using on-chip circuitry. The description of the invention set forth herein is illustrative and is not intended to limit the scope of the invention as set forth in the following claims. Variations and modifications of the embodiments disclosed herein may be made based on the description set forth herein, without departing from the scope of the invention as set forth in the following claims.
Claims (20)
1. A method for determining response of a phase-locked loop (PLL), comprising:
inserting a digital phase offset into a digital representation of a clock signal coupled to a phase and frequency detector (PFD) of the PLL, the digital phase offset representing a phase step;
determining a length of time between insertion of the digital phase offset and a first zero crossing of a phase error supplied from the PFD after insertion of the digital phase offset, the length of time being inversely related to bandwidth of the PLL; and
determining a maximum phase error overshoot that occurs after the zero crossing and results from insertion of the digital phase offset.
2. (canceled)
3. The method as recited in claim 1 , further comprising:
determining a peaking characteristic of the PLL based in part on the maximum phase error overshoot.
4. The method as recited in claim 1 , further comprising:
evaluating phase errors for the maximum phase error overshoot responsive to assertion of the zero crossing indication.
5. The method as recited in claim 1 , wherein determining the length of time comprises:
enabling a counter coincident with the digital phase offset being inserted;
detecting a first zero crossing of a phase error after insertion of the digital phase offset; and
determining a count value of the counter at a time of the first zero crossing.
6. The method as recited in claim 5 , further comprising:
determining a PLL bandwidth based on the count value of the counter at the time of the first zero crossing.
7. The method as recited in claim 5 , further comprising:
asserting a test signal to cause the phase step to be inserted and the counter to start counting.
8. The method as recited in claim 1 , further comprising:
wherein the clock signal is a reference clock signal being supplied to the PLL.
9. The method as recited in claim 1 , further comprising:
wherein the clock signal is a feedback clock signal coupled to a feedback divider of the PLL.
10. An integrated circuit comprising:
a time to digital converter to generate a digital representation of a clock signal;
a summing circuit to add a digital phase offset to the digital representation of the clock signal, the digital phase offset representing a phase step;
a counter to count a time period beginning when the digital phase offset is inserted;
a first zero crossing detection circuit to detect a first zero crossing of a phase error between the clock signal and a second clock signal and supply a zero crossing indication, the first zero crossing occurring after insertion of the digital phase offset;
wherein a count value of the counter is stored responsive to assertion of the zero crossing indication; and
a maximum phase error circuit to detect a maximum phase error overshoot resulting from insertion of the digital phase offset.
11. The integrated circuit as recited in claim 10 ,
wherein the maximum phase error circuit begins evaluating phase errors for the maximum phase error overshoot responsive to assertion of the zero crossing indication.
12. The integrated circuit as recited in claim 10 , further comprising:
wherein a peaking characteristic of a phase-locked loop is determined based on the maximum phase error overshoot and a size of the phase step.
13. The integrated circuit as recited in claim 10 , further comprising:
wherein a bandwidth of a phase-locked loop is determined based on the count value of the counter at the first zero crossing.
14. The integrated circuit as recited in claim 10 , further comprising:
wherein the clock signal is a reference clock signal being supplied to a phase-locked loop and the second clock signal is a feedback clock signal supplied by a feedback divider of the phase-locked loop.
15. The integrated circuit as recited in claim 10 , further comprising:
wherein the clock signal is a feedback clock signal coupled to a feedback divider of a phase-locked loop and the second clock signal is a reference clock signal.
16. The integrated circuit as recited in claim 10 , further comprising:
a test signal coupled to the counter, the counter responsive to assertion of the test signal to start counting.
17. An integrated circuit comprising:
a summing circuit to add a digital phase offset to a digital representation of a first clock signal, the digital phase offset corresponding a phase step in the first clock signal;
a phase and frequency detector of a phase-locked loop (PLL) coupled to receive the digital representation of the first clock signal and a digital representation of a second clock signal and supply a phase error signal indicative of a difference between the first clock signal and the second clock signal;
a first zero crossing detection circuit coupled to receive the phase error and to detect a first zero crossing of the phase error after insertion of the phase offset;
a counter to determine an elapsed time between insertion of the phase offset and the first zero crossing; and
a maximum phase error circuit coupled to receive the phase error and to detect a maximum phase error overshoot resulting from insertion of the digital phase offset.
18. The integrated circuit as recited in claim 17 , further comprising a processor configured to receive a count value indicative of the elapsed time and the maximum phase error overshoot and to determine a bandwidth of the PLL and a peaking characteristic of the PLL.
19. The integrated circuit as recited in claim 17 , wherein the maximum phase error circuit is responsive to begin evaluating phase errors for the maximum phase error overshoot responsive to the first zero crossing.
20. The integrated circuit as recited in claim 17 , wherein the first clock signal is a reference clock signal and the second clock signal is a feedback clock signal.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US16/697,613 US11018679B1 (en) | 2019-11-27 | 2019-11-27 | On-chip phase-locked loop response measurement |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US16/697,613 US11018679B1 (en) | 2019-11-27 | 2019-11-27 | On-chip phase-locked loop response measurement |
Publications (2)
Publication Number | Publication Date |
---|---|
US11018679B1 US11018679B1 (en) | 2021-05-25 |
US20210159904A1 true US20210159904A1 (en) | 2021-05-27 |
Family
ID=75974401
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US16/697,613 Active US11018679B1 (en) | 2019-11-27 | 2019-11-27 | On-chip phase-locked loop response measurement |
Country Status (1)
Country | Link |
---|---|
US (1) | US11018679B1 (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11283459B1 (en) * | 2021-03-30 | 2022-03-22 | Skyworks Solutions, Inc. | Calibration of a time-to-digital converter using a virtual phase-locked loop |
KR20240021038A (en) * | 2022-08-09 | 2024-02-16 | 삼성전자주식회사 | A digital phase locked loop and an operating method of the same |
Family Cites Families (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5656977A (en) | 1996-01-05 | 1997-08-12 | International Business Machines Corporation | Frequency-lock indicator circuit with frequency-only detection |
US6278746B1 (en) | 1998-05-12 | 2001-08-21 | Montreal Networks Limited | Timing recovery loop circuit in a receiver of a modem |
US6993101B2 (en) | 2000-04-07 | 2006-01-31 | Broadcom Corporation | Method of determining a start of a transmitted frame in a frame-based communications network |
US7323946B2 (en) | 2005-10-20 | 2008-01-29 | Honeywell International Inc. | Lock detect circuit for a phase locked loop |
US7570182B2 (en) * | 2006-09-15 | 2009-08-04 | Texas Instruments Incorporated | Adaptive spectral noise shaping to improve time to digital converter quantization resolution using dithering |
US7405628B2 (en) | 2006-09-29 | 2008-07-29 | Silicon Laboratories Inc. | Technique for switching between input clocks in a phase-locked loop |
US8493113B2 (en) * | 2011-09-12 | 2013-07-23 | International Business Machines Corporation | PLL bandwidth correction with offset compensation |
US8552767B1 (en) | 2012-03-30 | 2013-10-08 | Broadcom Corporation | Systems, circuits, and methods for a digital frequency synthesizer |
US8791734B1 (en) | 2013-02-13 | 2014-07-29 | Silicon Laboratories Inc. | Cascaded PLL for reducing low-frequency drift in holdover mode |
EP3422569B1 (en) | 2017-06-30 | 2021-11-10 | Nxp B.V. | An amplifier circuit |
US10608649B1 (en) | 2018-11-19 | 2020-03-31 | Silicon Laboratories Inc. | Relative frequency offset error and phase error detection for clocks |
-
2019
- 2019-11-27 US US16/697,613 patent/US11018679B1/en active Active
Also Published As
Publication number | Publication date |
---|---|
US11018679B1 (en) | 2021-05-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7999584B2 (en) | Method and apparatus for on-chip phase error measurement to determine jitter in phase-locked loops | |
CA2220622C (en) | Method and circuit for built in self test of phase locked loops | |
US8284886B2 (en) | Radio frequency built-in self test for quality monitoring of local oscillator and transmitter | |
US20040062301A1 (en) | Jitter measurement apparatus and jitter measurement method | |
US10819354B2 (en) | Accurate and reliable digital PLL lock indicator | |
US7158899B2 (en) | Circuit and method for measuring jitter of high speed signals | |
US5295079A (en) | Digital testing techniques for very high frequency phase-locked loops | |
US7957458B2 (en) | Jitter measurement apparatus, jitter measurement method, test apparatus and electronic device | |
US9116204B2 (en) | On-die all-digital delay measurement circuit | |
US11018679B1 (en) | On-chip phase-locked loop response measurement | |
JP4204201B2 (en) | A method to adapt sampling rate based on power system frequency | |
TW201235681A (en) | Method and circuit of clock data recovery with built in jitter tolerance test | |
US7460499B2 (en) | Modulation noise estimation mechanism | |
Ferrero et al. | A low-cost frequency multiplier for synchronous sampling of periodic signals | |
Yamaguchi et al. | Extraction of peak-to-peak and RMS sinusoidal jitter using an analytic signal method | |
JP5328096B2 (en) | Jitter measuring apparatus, jitter measuring method, test apparatus, and electronic device | |
JP5113368B2 (en) | Jitter measuring apparatus, jitter measuring method, test apparatus, and electronic device | |
US20210176020A1 (en) | Jitter self-test using timestamps | |
Park et al. | All-digital PLL frequency and phase noise degradation measurements using simple on-chip monitoring circuits | |
Hsu et al. | Timing jitter and modulation profile extraction for spread-spectrum clocks | |
US20020039027A1 (en) | Dynamically periodical detecting method and detector | |
Fischette et al. | An on-chip all-digital measurement circuit to characterize phase-locked loop response in 45-nm SOI | |
Egan et al. | A framework for the characterization and verification of embedded phase-locked loops | |
GB2387662A (en) | Measurement of transfer function or step response for a phase locked loop | |
Lee et al. | Low-cost technique for measuring clock duty cycle on FPGAs |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
FEPP | Fee payment procedure |
Free format text: ENTITY STATUS SET TO UNDISCOUNTED (ORIGINAL EVENT CODE: BIG.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
STCF | Information on status: patent grant |
Free format text: PATENTED CASE |
|
AS | Assignment |
Owner name: SKYWORKS SOLUTIONS, INC., CALIFORNIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:SILICON LABORATORIES INC.;REEL/FRAME:057033/0579 Effective date: 20210723 |