Home

Mains Frequency Fluctuation Metering

1. CH1 Freg 50 00Hz CH2 Freg 50 00Hz CH1 1 Max 3 60 CH2 Pk PE 21 04 CH2 RMS T 43V CH1 5 00 CH2 500 M 10 0rns CH2 400m 20 Oct 13 12 48 50 0053Hz Figure 12 Oscilloscope output of frequency tracking pulse generator 3 3 Trimble Copernicus II oi R3 3 XSTBY Figure 13 Trimble Copernicus Il DIP module 12 The Trimble Copernicus II is a GPS receiver module The Copernicus II used in this project came factory mounted to a DIP module A 3V magnetic mount SMA antenna was purchased to connect to the Copernicus Il The SMA antenna boosts the receivers gain by 26dB Specification Value Mode PPS Accuracy 60 ns RMS Static 350 ns RMS Stationary Mode Warm Start Time 35 secs Cold Start 38 secs Hot Start 3 secs No Battery Backup Tracking Sensitivity 160 dBm Acquisition Sensitivity 142 dBm Standard P 148 dBm High Sensitivity Table 7 Trimble Copernicus Il GPS receiver specifications 33 30 The Copernicus Il was chosen due to its specifications which are shown in table 7 The cost and high level of configurability made the receiver suitable for this project Several other Arduino shield based GPS receivers were considered but either lacked features precision specifications or were not easily adaptable for use with the Arduino Due An older revision of the Copernicus was also considered due to the price but lacked the precision and functionality the
2. which provide an excellent timing reference 16 The main issue that arises with time synchronisation at the receiving end is radio signal transfer Due to the nature of RF wave propagation significant jitter delays and signal loss may be encountered when transmitting the signal over long distances Broadcasting stations transmit at a frequency range of 25kHz to 25 MHz 21 with the exception of radio station STFS which transmits at approximately 2 6GHz 21 Between 25kHz and 25MHz signals fall into the Low 3 30kHz Medium 0 3 3MHz and High 3 30MHz frequency categories 22 Low frequency transmissions primarily travel over surface waves which travel slightly further than the visible horizon 22 Past the radio horizon the signals may reflect off the sky Medium frequency transmissions are primarily surface waves during the day with some sky wave reflection during the night 22 High frequency signals propagate as sky waves over long ranges using ionospheric returns 22 Radio clock technology was chosen to not be relied upon for the purpose of this project due to the propagation distance from the nearest radio clock station to Perth Western Australia While in principle this technology can be used to synchronise clocks with an accurate reference the location this study was conducted at had highly unreliable reception The nearest radio time signal station to Perth is call sign JJY located at Mount Otadakoya Fukushima Japan at a d
3. 0001Hz with a standard deviation of 0 0263Hz This is as expected as the grid frequency should not vary greatly in order to maintain nominal values Contingency Day Start Time Finish Time Total Duration Under frequency Sunday 02 44 44 AM 02 46 39 AM 1 min 55 Event 1 seconds Under frequency Monday 05 29 11 PM 05 32 45 PM 3 min 34 Event 2 seconds Table 14 Under frequency events detected during frequency meter performance tests Two under frequency events were detected on the grid The first under frequency event lasted a short duration at the times given in table 14 It is likely that the grid had very few loads connected at this time compared to the amount connected during the day and a large number of loads connecting to the grid at this early hour affected the grid in an unexpected way but this could only be confirmed by obtaining frequency data from a more accurate and verified source The second under frequency event was after typical working day hours on a Monday andis likely caused by a very large number of loads connecting to the grid suddenly thereby slowing down the large generators that maintain the grid frequency The recovery times on both under frequency events were well within the 15 minute specification for return to nominal grid frequency 60 Under Frequency Event 1 Data Under frequency Event 1 Frequency Hz 13 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45
4. 1766 593508000 3203 96635 11550 22761 00010 5B Reset Configuration Packet SPTNLSRT H 2 2 0000000000 18 Figure 23 NMEA Configuration VI user interface The configuration VI shown in figure 23 allows the user to configure the NMEA packets for the Trimble Copernicus II prior to run time on the Control Panel VI All of the sentences listed in table 8 are able to be configured through this VI with practical defaults already set such as 41 Receiver Configuration 15 elevation mask to reduce ionospheric jitter Land dynamics WAAS enabled Even though WAAS corrections are not available in Australia 50 future implementations of WAAS will allow the frequency meter to perform with even less jitter on the PPS output PPS Configuration Only generate outputs if a GPS fix is available 500ms pulse length active HIGH 10ns cable delay compensation due to standard SMA connector antenna Acquisition Sensitivity Standard lowest PPS jitter setting Serial Communications 19200 Baud Already the default for NMEA in the Arduino program Initial Position GPS week and time are automatically generated by LabVIEW based on the system time Latitude and Longitude are set to the Engineering and Energy building location at Murdoch University and altitude 10m above sea level This location has to be within 100km of the true location and within 5 minutes of the specified UTC time to allow the receiver to lock on as fast as possible Reset
5. 20Code 202005 pdf Accessed 19 August 2013 67 Appendices Appendix A Arduino Program See attached folder named MFFM Arduino Appendix B LabVIEW Program See attached folder named MFFM LabVIEW Appendix C Referenced Material See attached folder named MFFM References Appendix D Logging Session Data Session Data for Jitter Logging Trimble vs Arduino 29 10 2013 7 08 PM Logging Started 29 10 2013 7 08 PM First Fix 31 10 2013 7 10 PM Finish Used NMEA Packets SPTNLSPS 2 5000000 0 0000000 51 SPTNLSFS S 0 23 SPTNLSPT 019200 8 N 1 4 4 1C SPTNLSKG 1764 241680000 3203 96635 S 11550 22761 E 00010 52 SPTNLSCR 15 0 1 1 5C 68 Session Data for Jitter Logging EM406A vs Arduino 31 10 2013 7 16 PM Logging Started 31 10 2013 7 17 PM First Fix 02 11 2013 7 20 PM Finish Used NMEA Packets SPSRF100 1 19200 8 1 0 38 SPSRF105 1 3E SPSRF103 00 00 01 01 25 SPSRF104 32 066142 115 837122 10 96000 142774 2787 12 1 34 Session Data for Frequency Logging Trimble Copernicus Il PPS source on Arduino Due MCU Start 12 19PM 10 11 2013 Finish 1 00pm 12 11 2013 Initialisation and PPS Packets SPTNLSKG 1766 087588000 3203 96635 5 11550 22761 E 00010 53 SPTNLSPS 2 5000000 1 10 51 All other packets were LabVIEW default settings All logged data can be found in the folder MFFM_Logged_Data 69 Appendix E Annotated Bibliography Fundamentals of Quartz Oscillators 9 This document covers the n
6. 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 101 103 105107 109111 113 115 Time Elapsed Seconds Figure 42 Under frequency Event 1 graph The first under frequency event can be seen in figure 42 The data seems to dip and rise over the course of 12 seconds from 50Hz to below 49 8Hz and then to 49 85Hz Successive dips in the grid frequency are minor and end 1 minute 55 seconds after the initial dip below the grid frequency Under Frequency Event 2 Data 433 Under Frequency Event 2 Frequency Hz Time Elapsed Seconds Figure 43 Under frequency Event 2 graph The under frequency event displayed in figure 43 dips very quickly over the course of seconds similar to the under frequency event in figure 42 The initial dip similarly experiences slight swing back closer to nominal frequency and then gradually recovers to the nominal value The data in event 2 was expected at the time it was recorded and the waveform while slightly noisy demonstrates that minor frequency fluctuations are be detected on the mains supply line 61 7 Recommendations and Future Improvements The timing source jitter analysis for the KX 7 quartz crystal was performed on two GPS receivers that induced their own jitter into the measurements as well thereby reducing the precision of the analysis To improve this a higher standard of timing such as an atomic standard could be used to analyse
7. 56 6 1 LiHardware Components n eee Herde cs sade A A a A a aAa 57 6 1 2 Program Parameters rete eorr e reae eere en et ee Pene resonet eue teas en e ee eee eed i ae Eee eaten iega 57 6 2 Pertormance Results trt ater oe urere aet ure drin o eltanwstersut epe exceda vet davedouseneendeben heats 60 7 Recommendations and Future Improvements eese nennen nnns nne en terea nisse tte te aas sss ensi ranas 62 S CONCIUSION S 63 9 ReferBcBs 6 A AAA uae eura eter er hen ee no wa a Ad re Pn inh tone e eeu beu vu Pee erae 64 Appendices ee Hye eet DR dex ee ties ii 68 Appendix A Arduino Program c sscccccccsssssssssececscsssscsssesecessesseneesecsesssessneeseceesssesaeaeseceessaesaeassesecsseesesasaesecees 68 Appendix B EabVIEW Program d e eee tene ie ee e eh pee eee exea ta ea ev 68 Appendix C Referenced Material trii eto etie treo HERE HER tbe e PRENSA Ee URS WERE R YE Ee I re kien iaa TER URS 68 Appendix D Logging Session Data coccccnononocnconenononenonnnnnnnnnnnononnnnnnnnnnnnnennnnnnnnnnnnnnnnnnnncnnnnn no nnnnnnnnannnnnnnnnnnnnos 68 Appendix E Annotated Bibliography ccccccccccessssssceceeececsesesaeeeeccecsesesaeseeecsceeseeassesecsceesaaeaeseeseseseeaaeeeeees 70 Fundamentals of Quartz Oscillators 9 oooooooocnccononoooonannnonononannonncnncnnnnnnnonnnncnnnnnonnnnnnncnnnnnnonnnnncnnnannns 70 Relative timing characteristics of the one pulse per second 1PPS output pul
8. Copernicus ll provided The stock module allows TTL level serial communications on 6 ports 3xTX and 3xRX and allows communicating in three different formats e TSIP Trimble Standard Interface Protocol this interface is Trimble s primary packet transmission standard in their GPS receivers e TAIP Trimble ASCII Interface Protocol primarily suited to vehicle tracking applications Considerably powerful in networked environments due to the ability of communicating through a unique ID in packet based communication e NMEA National Marine Electronics Association this packet standard is supported by the TinyGPS library and can easily be parsed to the LabVIEW program for packet analysis The chosen protocol for this project was NMEA due to its simplicity and to favour a set standard between the Trimble Copernicus ll and the GlobalSat EM406 A modules The configured messages used by the receiver can be seen below in table 8 Some of these messages are fixed while others vary as time changes Packet Sentence Description Automatic Message SPTNLSNM 0021 01 54 Configures receiver to output GGA Output messages every second Receiver SPTNLSCR 15 0 1 1 5C 15 elevation mask Stationary Configuration mode WAAS enabled PPS Configuration SPTNLSPS 2 5000000 1 0000010 51 Fix Based PPS 500ms pulse Active HIGH 10ns cable delay compensation Acquisition SPTNLSFS S 0 23 Standard sensitivit
9. Crystal 19 October 2012 Online Available http www geyer electronic de uploads tx_userartikelfrequenz GEYER KX 7_01 pdf Accessed 17 October 2013 B A Forouzhan and F Mosharraf Computer Networks A Top Down Approach in Physical Layer and Transmission Media New York McGraw Hill 2012 pp 548 560 International SEMATECH Manufacturing Initiative Using Network Time Protocol NTP Introduction and Recommended Practices 28 February 2006 Online Available http www sematech org docubase document 4736aeng pdf Accessed 16 October 2013 27 J Burch K Green J Nakulski and D Vook Verifying the Performance of Transparent Clocks in 28 PTP Systems in International IEEE Symposium on Precision Clock Synchronisation for Measurement Control and Communication Brescia 2009 G D Krebs GPS 2A Navstar 2A 27 January 2013 Online Available 65 29 30 31 32 33 34 35 36 37 38 39 40 http space skyrocket de doc sdat navstar 2a htm Accessed 5 Novermber 2013 Lockheed Martin U S Air Force Awards Lockheed Martin GPS III Flight Operations Contract 31 May 2012 Online Available http www lockheedmartin com us news press releases 2012 may 0531 ss gpslll html Accessed 18 August 2013 Lockheed Martin Corporation GPS III The Next Generation Global Positioning System 2011 Online Available http www lock
10. EBS BREE Se Sos un m 00 DIN FMN e O c 00 Sexo na 0S Dmaoa SUR Nena smo foe DUN Oo n NAN uuu O0 OO C O e e r4 m m in in Y 5000 v e M B e Elapsed Time Seconds Figure 31 Arduino 48 hour mean centered jitter graph Copernicus II PPS source Figure 31 above displays how the Arduino s jitter changes over the 48 hour logging period relative to the PPS signal the Copernicus II GPS receiver provides The initial aim was to allow the ambient temperature to naturally change and directly affect the Arduino s timing jitter Unfortunately while the location where this data was collected was suitable for frequency logging due to excellent GPS signal reception and the ability to leave the laptop running for 48 hours to collect the data air conditioning was turned on at 9AM each day and skewed the results This is displayed by the fact that the jitter is primarily positive around the mean value indicating a cooler environment when compared to the data collected in 5 1 2 51 48 Hour Clock Time Interval Error 900000 800000 700000 800000 500000 Cumulative TIE 300000 200000 Linear Cumulative TIE 100000 O Cumulative Time Interval Error ps un Cx 00 R 27310 36413 45516 54619 63722 81928 91031 100134 109237 118340 127443 136546 145649 154752 1633 Elapsed Time Seconds Figure 32 Arduino 48 hour TIE graph Copernicus Il PPS
11. Equation 2 Determination of frequency using a gating period Equation 3 Fractional error of frequency measurement using a gating period Equation 4 Calculation of parts per million specification based on the center frequency Hz and peak frequency variation Hz VII List of Abbreviations ASCII American Standard Code for Information Interchange BIPM International Bureau of Weights and Measures BJT Bipolar Junction Transistor DMA Direct Memory Access ERA Economic Regulation Authority GNU Gnu s Not Unix GPS Global Positioning System IEEE Institute of Electrical and Electronics Engineers IETF Internet Engineering Task Force IRQ Interrupt Request ISR Interrupt Service Routine LCD Liquid Crystal Display MCU Micro Controller Unit NMEA National Marine Electronics Association NTP Network Time Protocol OCXO Oven Controlled Crystal Oscillator PC Personal Computer PCB Printed Circuit Board PPM Part s Per Million PPS Pulse Per Second output of a GPS PTP Precision Time Protocol RF Radio Frequency RMS Root Mean Square SV Space Vehicle SWIS South West Interconnected System TAI International Atomic Time TCXO Temperature Compensated Crystal Oscillator TIE Time Interval Error TTL Transistor Transistor Logic USB Universal Serial Bus UTC Coordinated Universal Time VI Virtual Instrument WAAS Wide Area Augmentation Sy
12. IMGAaSUFEMENE UncertaiNty ee 9 1 2 FREQUENCY DeteCctlOn iren eiie eai ero rao rir A e ARE erba sa aaia a 10 1 2 1 Measurement Error Sources ssesseeeeeeen ener enne seen nennen enn enne ne saroien ENT 10 1 22 Counting Method ioter RN 12 12 3 Frequency Counters suo odere teo iii ida 13 1 2 4 HeterodVrilllg ice ertet NN A 13 LAC MEHDEERETITUL 13 1 2 6 Accuracy of Modern Systems erret erii o re nh nodo ae raa Fea eaae Fa Chee UP rav eA Te de eov Ra YAN E ido d Ee EYE ER aE 14 1 3 Thesis PUrpOSE ienis a FR FR ACER RD 14 14 IESU E 15 Pd s ol dd aU o o NEAR E EET T IR DEEP 16 2 ET eT TS 1d A NA 16 2 1 1 Atomic CIOCKS e NON 16 251 2 Radio eoru m M 17 21 3 CrystallOscillatOrS iii errore tilda andar Ferran ed rri NUR RR FE FE FRUI a cadivi 18 21 4 Time Protocols O 19 2 1 5 Global Positioning SySteM ccccocococooncnncnnnononnnnnnncnnnnnononnnnnnonnnnonnnnnnnnnnonnnnonnnnnnnnnnnnnnennnncnnannnnnnnnnnnnninnnns 20 2 2 Grid ParatfietelSi A entre dieto chin temseccevstawets tins nee erras 23 3 Hardware ImplemiertatiOr 5 o oor a a aaa e eaaa aae LN ERR Foe REOS sd devs sae iesavaancesness sane 24 3 1 Mitro Controller D Jo fm 24 3 2 Frequency Detection Shield cccconcoconcnncnocononooncnncnncnnonnnnonncnonnnononnonnnncnnnnnnnnnnnnncnnnonennnnnnncnnnnnonnnn
13. SWIS region this value must be less than 10 seconds for 9996 of the time Event Frequency Band Hz Target Recovery Time Single Contingency 48 75 51 00 Normal range 15 mins Over frequency events 50 5Hz within 2 mins Multiple Contingencies 47 00 52 00 Normal range 15 mins Under frequency events 1 Above 47 5Hz within 10 secs 2 Above 48 0Hz within 5 mins 3 Above 48 5Hz within 15 mins Over frequency events 1 Below 51 5Hz within 1 min 2 Below 51 0Hz within 2 mins 3 Below 50 5Hz within 5 mins Table 3 SWIS target recovery times for grid frequency variations due to contingencies 35 The parameters in table 3 are primarily used to compare contingencies detected on the frequency meter against the specified standard to assure the recovery times are within the specified range 23 3 Hardware Implementation 3 1 Micro Controller Unit The Arduino Due 10 seen in Figure 8 was chosen as the prototyping Microcontroller Unit MCU for the project amongst other MCUs due to its hardware specifications cost large collection of open source libraries instant availability and its ability to meet the requirements of the project A A HO rcm orn A mig 0s dT Figure 8 Arduino Due MCU 10 Murdoch University s Engineering amp Information Technology department provided an Arduino Due for prototyping the metering unit The Arduino website and Atmel datasheet list the following s
14. accounting for the ISR processing time in the Arduino code this mean error is reduced to 4 5425ppm No compensation is made for the jitter contribution made by the Copernicus ll as it is assumed that while the PPS jitter is 60ns its mean value is sufficiently close to O to make it negligible The Copernicus II GPS receiver exhibited a 100 up time during the data logging session with only 4 successive timing outliers that had to be removed due to incorrect timing values being recorded Appendix 4 lists the data collection dates and times and what packets were used to initialise each GPS receiver Histogram 120000 4 99633 100000 4 80000 4 E 60000 55631 g E Frequency 40000 4 17677 20000 1 4 9 0 o T T T T T T 999990 999991 999992 999993 999994 999995 More Microseconds Elapsed Between PPS Interrupts Figure 30 Histogram of PPS generated time intervals on the Arduino Due Copernicus Il PPS source Figure 30 displays the recorded timing distribution in a histogram The time interval data appears to largely be centered around 999 993us with mostly 1us sway to each side The initial assumption was that the timing jitter would be largely erratic given the nature of the ppm specifications found in literature however this data implies otherwise 50 48 Hour Clock Jitter from the Mean Value Seriesl Cumulative Clock itter us cooDmr un Rzuonz9 2ozxcoc2g2mth 5 OMG REA RPRRSBSE
15. any satellite signals below the set angle PPS was configured to output a 1Hz pulse with a 5096 duty cycle only when the receiver has a fix Because the Copernicus Il uses a 2m SMA connected RG 174 type antenna 47 it s propagation delay is equivalent to 10 12ns based on GPS source coaxial cable propagation delay data sheet 48 hence the receiver was configured to output it s PPS 10ns earlier However this effect can be effectively ignored as the transmission delay will stay the same and have no significant change at room temperature In standard acquisition mode the receiver has an acquisition sensitivity of 148dBm and 160dBm once the receiver has a fix 33 High sensitivity mode should only be used under obscured signal conditions but at the cost of an increased time to first fix The only parameters that can be changed in the serial communications packet are the Baud rate input protocol and output protocol A baud rate of 19200 bps with NMEA in out was set as they were suitable for communications with the Arduino Due This communication had no effect on the PPS signal as they were wired to separate pins hence this baud rate could freely be changed as the data that s sent through the GGA message to the LabVIEW terminal is well within 4800 bps 32 The initial position packet is configured to decrease the time to first fix by providing ephemeris location data to the receiver which includes latitude longitude and altitude above sea
16. detectable on the order of uHz based on the timing jitter results found in section 5 The system had the limitation of noisy frequency readings at the uHz resolution however This may be due to mains supply noise triggering the pulse generation shield at slightly incorrect times due either to mV level variations on the grid or noise induced by the mains adapter Future improvements in regards to this and other recommendations have been outlined in section 7 63 9 References 1 2 3 4 5 6 7 8 9 10 11 12 13 International Organisation for Standardisation ISO 5725 1 1994 2012 Online Available http www iso org iso catalogue detail htm csnumber 11833 Accessed 09 September 2013 International Organisation for Standardisation ISO 3534 1 2006 2010 Online Available http www iso org iso catalogue_detail htm csnumber 40145 Accessed 09 September 2013 Wikipedia Accuracy and Precision 28 October 2013 Online Available http en wikipedia org wiki Accuracy and precision Accessed 29 October 2013 Royal Society of Chemistry AMC Technical Brief September 2003 Online Available http www rsc org images terminology part 1 technical brief 13 tcm18 214863 pdf Accessed 9 September 2013 Wikipedia Frequency Measurement 2013 Online Available http en wikipedia org wiki Frequencytt Measurement Accessed 24 October 2013 J Carr Radio Rec
17. importance because it can be compared to the variations in jitter on the KX 7 crystal since temperature will have the greatest effect 5 1 Arduino Frequency Stability Data Performance tests were created to quantify the clock drift of the Arduino Due s 12MHz KX 7 Quartz Crystal _ 10 xAf 4 ppm To attempt to correlate the logged data to a similar standard the EM406 A and the Trimble Copernicus ll GPS receivers both provided a PPS signal as a reference timer on the Arduino Due Deviation from this PPS signal would come from the crystal oscillator s frequency jitter and a significantly smaller portion of this deviation is quantified as the PPS signal jitter itself 5 1 1 Clock Drift Relative to Trimble Copernicus II The clock drift of the Arduino was logged through a 48 hour PuTTY session into a comma separated value file It s also possible to log the clock drift of the Arduino now with the latest implementation of the LabVIEW programs Mean us Max Value Mean ps Min Value Mean us Standard Deviation Mean ps 999993 2195 999995 999990 0 613217 Table 12 PPS triggered Arduino 1 second timing interval data Copernicus Il PPS source From the mean value in table 12 the mean jitter can be calculated on the Arduino for the 48 hour period The mean error value is calculated at 6 7805 ppm with a standard deviation of 0 613 49 around the mean value corresponding to the crystal s jitter By
18. is chosen to detect the frequency the sampling period must be considered carefully to avoid aliasing Aliasing is an undesirable effect caused by sampling a periodic waveform below the Nyquist sampling rate 7 The Nyquist theorem states that the sampling frequency should be at least twice the sampled signals frequency 8 By sampling at less than twice the input frequency a false frequency may be sampled In practical applications this value should be 5 10 times higher than the sampled frequency as a minimum so that the reconstructed signal is more defined and less prone to noise Figure 5 illustrates an aliased sinusoid due to an under sampled AMNIS Figure 5 Aliased sinusoidal waveform due to an under sampled signal 5 signal 10 13 1 2 6 Accuracy of Modern Systems Many electronic systems rely on a XO which has a known internal oscillation frequency to provide a continuous time base XOs are relatively cheap and effective but are subject to frequency stability variations especially in long term use and environments with significant temperature variation 9 The electronics that rely on XOs for a stable time base are usually precise in the short term Long term stability may be significantly affected depending on the crystal s cut temperature and material 9 To compensate for the frequency instability in electronics that rely on XOs a more precise timing signal could be used to either steer the electronic clock to the m
19. seen in appendix A and a structured flow chart of the program is given Variable Declarations Function Definitions GetRisingEdge ISR Setup Routine Open serialO Print SysTick LOAD register value Attach PPS interrupt PPS Interrupt Print Number of SysTick gt VAL gone past 0 SysTick gt VAL call Figure 26 ISR clock cycle quantifying program 44 The mean processing time for the SysTick VAL storage call and the PPS GetRisingEdge ISR were 7 clock cycles and 194 97 clock cycles respectively with a standard deviation of 0 on both This time corresponds to the execution cycles undertaken when a rising edge has occurred This translates to an offset of 2 321 us for the PPS ISR and a final adjusted offset of 2 238 us when the clock cycles that SysTick gt VAL calls require are factored in 4 2 2 Alternate Microsecond Function Implementation One significant factor in the design of the Arduino programs was the calls to the micros function Micros returns an unsigned 32 bit integer time value upon each call equal to the number of microseconds that have elapsed since the Arduino Due was turned on This function had the issue of gaining a millisecond during some ISR calls because the system timing ISR was not able to be called uint32 t micros void uint32 t tricks uint32 t count SysTick gt CTRL do ticks SysTick gt VAL count GetTickCount while SysTick gt
20. source The TIE graph in figure 32 attempts to illustrate the low long term effect that the temperature had on timing jitter of the Arduino was negligible with relatively small temperature variations The R value is very high indicating a high data correlation to the linearly fit trend line After 172 800 seconds 48 hours the TIE accumulated to 799 08ms or an average of 399 54ms lost to the Arduino s timing offset per day This value is largely based on the mean error from the PPS signal time and is minimally affected by the clock s jitter 52 5 1 2 Clock Drift Relative to GlobalSat EM406 A The EM406 A provided a PPS timing source for the Arduino in an ambient temperature affected environment Air conditioning was kept off to prevent artificial modification to the jitter logging data A PuTTY client recorded the 48 hour session similarly to section 5 1 1 Mean us Max Value ps Min Value ps Standard Deviation ps 999993 8166 999995 999992 0 540911 Table 13 PPS triggered Arduino 1 second timing interval data EM406 A PPS source The mean error value was logged at 6 1834 ppm over the 48 hour logging session with jitter analysed from the standard deviation value of 0 541ppm around the mean error The PPS ISR compensated mean error was 3 9454ppm Histogram 140000 117115 100000 4 E 80000 3 y 60000 4 E Frequency 43450 40000 4 EE 12148 242 0 o T T T 999992 999993 999994 999995 M
21. Antenna 1 June 2006 Online Available http php2 twinner com tw files onshine ANT555 2006 NEW pdf Accessed 17 October 2013 48 GPS Source Calculating the propagation delay of coaxial cable 25 January 2011 Online Available http www gpssource com files Cable Delay FAQ pdf Accessed 17 October 2013 49 National Instruments LabVIEW System Design Software National Instruments 2013 Online Available http www ni com labview Accessed 29 July 2013 50 GPS Oz WAAS May 2008 Online Available http www gpsoz com au WAAS htm Accessed October 2013 51 stimmer GitHub April 2013 Online Available https github com arduino Arduino pull 1388 files Accessed 16 October 2013 52 M P J Relative timing characteristics of the one pulse per second 1PPS output of three GPS receivers in The 6th International Symposium on Satellite Navigation Technology Including Mobile Positioning amp Location Services Melbourne 2003 53 K Ozsoy A Bozkurt and Tekin Indoor positioning based on global positioning system signals Microwave and Optical Technology Letters vol 55 no 5 pp 1091 1097 2013 54 Economic Regulation Authority Electricity Industry Network Quality and Reliability of Supply Code 2005 6 December 2006 Online Available http www era wa gov au cproot 2372 2 D04 20Electricity 20Industry 20 28Network 20 Quality 20and 20Reliability 200f 20Supply 29
22. CTRL SysTick CTRL COUNTFLAG Msk return count 1000 SysTick gt LOAD 1 ticks SystemCoreClock 1000000 Figure 27 Previous Arduino library implementation of micros 10 The code in figure 27 was unsuitable for use in interrupts due to relatively frequent error of 1ms The user stimmer 51 submitted a more suitable micros function that does not suffer from the SysTick register rollover issue and this was implemented within the ISRs generated in the code Figure 28 provides the code submitted by stimmer and did not have any observed millisecond sized fluctuations during any logging periods thereby relieving the initially detected issue 45 uint32 t micros void uint32_t ticks ticks2 uint32_t pend pend2 uint32_t count count2 ticks2 SysTick gt VAL pend2 SCB gt ICSR amp SCB_ICSR_PENDSTS SCB gt SHCSR amp SCB_SHCSR_SYSTICKACT_Msk count2 GetTickCount do ticks ticks2 pend pend2 count count2 ticks2 SysTick gt VAL pend2 SCB gt ICSR amp SCB_ICSR_PENDSTSET_Msk SCB gt SHCSR amp SCB SHCSR SYSTICKACT Msk count2 GetTickCount while pend pend2 count count2 ticks lt ticks2 return count pend 1000 SysTick gt LOAD ticks 1048576 F_CPU 1000000 gt gt 20 this is an optimization to turn a runtime division into two compile time divisions and a runtime m
23. Configuration Hot Start store user configuration to flash memory on reset stand by request wake up on activity on Port B NMEA IN Each of these sentences can be modified with minimal work by the user and the VI details any comments relating to expected input format The output type of all sentences is a string NMEA Checksum vi This VI performs an 8 bit exclusive OR on all the ASCII byte value components of the literal string input The output type is string created by taking the final exclusively OR ed byte and converting it to a string representation of the hexadecimal byte equivalent String Example Exam Ascii Decimal Values d i ii 112 108 101 Binary Values 01000101 01111000 01100001 01101101 01110000 01101100 01100101 Exclusively ORed Output 01001000 Output Hex Representation 48 Figure 24 NMEA checksum generation illustration Figure 24 displays how a string called Example may be broken into its individual characters and each character has a decimal or base 2 binary representation When each of these binary values are exclusively ORed together another binary value is output and this can be represented as a hexadecimal value that is finally represented as a string in NMEA messages This string is used for error checking from and to the GPS receiver 42 NMEA Packet Decoder vi lad m Figure 25 NMEA Packet Decoder VI user interface The VI shown in figure 25 is similar to the NMEA Configur
24. EA packets for the EM406 A however will always need to be manually entered if it is chosen as the PPS ISR generating GPS receiver as the Trimble Copernicus Il GPS receiver is configured for default settings The program will wait for the periodic output GGA message to confirm the GPS has a fix The number of satellites required to make the GPS fix valid will depend on whether the Copernicus II is operating in stationary mode or an alternative dynamic such as Land Sea or Air In stationary mode only one satellite is required to get a time signal fix but the jitter increases from 60ns to 350ns 33 In all other modes the jitter is at the nominal 60ns data sheet specified value 33 As soon as the GPS fix is valid the appropriate interrupts are attached and the main loop is run The main loop is fairly simple comma separated data Pulses Counted Gating Error at Start Gating Error at End is printed to the active serial line and stored to the SD card as backup 48 5 Timing Precision Timing precision results were generated over 48 hours to ensure the data could be analysed over a cyclical period to determine if the KX 7 crystal s timing jitter varied periodically Date Temperature C Min Max Mean 29 10 2013 13 31 22 30 10 2013 18 37 28 31 10 2013 14 26 20 01 11 2013 14 27 20 02 11 2013 16 29 22 Table 11 Temperature data for the jitter logging time interval The temperature data in table 11 is of
25. L2C signal for more robust civilian use 30 e Block IIF 12 satellites are due to launch in this series with the second being sent in July 2011 IIF has all of IIR M s capabilities introduces a 3 Civilian Signal L5 29 In May 2012 the contract for the next generation of satellites has been awarded to Lockheed Martin to provide Block IIIA satellites 29 The primary benefits of the new generation are higher accuracy improved anti jamming increased lifetime and backward compatibility with older systems 30 The first satellite in this generation is due to launch in 2014 29 and will also introduce 4 civilian signal L1C 30 In the past civilian use of GPS suffered from selective availability which was discontinued on May 2 2000 31 Selective availability affected all non military GPS receivers by increasing the location error up to 100m away from the true position This error was unacceptable for high precision location and timing applications In timing applications that rely on a GPS receiver s PPS this error caused significant additional timing jitter A 100m location error generated by selective availability is equivalent to 333 6 nanoseconds PPS jitter Fortunately this is no longer an issue GPS satellites typically transmit at two frequencies the L1 frequency band 1575 42 MHz and the L2 frequency band 1227 6 MHz These frequencies are in the ultra high frequency band 300 3000 MHz Radio waves in this frequency
26. Mains Frequency Fluctuation Metering 1 wu Murdoch UNIVERSITY A thesis submitted to the School of Engineering and Information Technology Murdoch University in partial fulfilment of the requirements for the degree of Bachelor of Engineering By Dusan Sibanic November 2013 I Acknowledgements would like to thank Dr Gareth Lee for his supervision helpful feedback and guidance throughout the duration of this thesis My deepest thanks go out to my family partner and friends for their continuous support Your patience and understanding through the duration of this thesis gave me motivation and momentum from beginning to end would like to extend a special thank you to Mark Purvis for his support with the SD Card shield and relevant libraries and Michael Chapman for his generosity with additional components in the project Thank you to all the people working alongside me in the E amp E building s project room and pilot plant You folks made the long days seem shorter That s what it s all about you have to enjoy life Acknowledgements go to Arduino and all the hard working open source library developers that allow interesting projects to continue to come to fruition For all future thesis students reading this wish you luck and encourage you to persist in your studies II Abstract Detecting the frequency of the mains supply is a crucial component of maintaining the grid frequency at its nominal lev
27. Period T Figure 2 Illustration of jitter on a periodic waveform The cumulative time interval error TIE is depicted in figure 3 If the cumulative TIE reaches over 50 of the nominal period the error will not be recognisable i e a 51 error will be taken as 4996 To ensure the cumulative TIE doesn t reach this threshold a clock source with a quantified jitter should be used and periodically calibrated to a more precise source if required Ideal Waveform Time 1 Interval ul Error Real Waveform Cumulative Time Interval Error MW N z Y m E W D Mh E kr Figure 3 TIE generated by the real waveforms jitter relative to the ideal waveform 11 1 2 2 Counting Method Counting is a method for frequency detection and involves recording the number of waveform periods during a set gating period which is simply a chosen constant time interval 5 By counting the number of input signal cycles over a gating period it is possible to determine the frequency by dividing the number of counted cycles over the gating period as shown in equation 2 The fractional error associated with this is given in equation 3 and is inversely proportional to the sampled waveforms frequency as shown in figure 4 Cycles Counted f E Gating Period sec 2 Af 1 3 f 2 f Tm Gating Error Gating Period T Gating Error Gating Error Gating Period To Gating Error Figure 4 Gating error magn
28. ansitions at 6 834 682 610 904324 Hz 18 Both NIST and BIPM have defined the standard second based on the caesium 133 standard as the 9 192 631 770 periods of the radiation corresponding to the transition between two hyperfine levels of the ground 16 state of the caesium 133 atom 19 This means that atomic clocks can achieve accuracy on the order of parts per billion which translates to better performance than any other available timing source 2 1 2 Radio Clocks Radio clocks are synchronised by the RF signal containing time data that timing signal stations send The list of broadcasting stations is maintained by the BIPM 19 The broadcasting stations are spread internationally A limitation of radio clocks is that many locations have poor signal reception or no reception at all Radio clock stations all vary in the frequency bands they may output their timing signals 20 Antennas vary proportionally in size to their output frequency which affects the length of the propagated RF signal Stations also vary in their transmission times where some stations may transmit the time signal continuously and others have downtime The length of the pulse per second signals can also vary between stations 20 The lack of a standardised timing signal format and time interval between signals may potentially make radio clocks unsuitable for some applications Radio clock stations are primarily connected to atomic clocks such as caesium standards
29. ation VI in it s function because it is in charge of generating an NMEA message However this VI is separate because it is not focused on initialisation sentences but instead is focused on allowing the user to select from an array of message instruct the GPS module choices to generate the appropriate automated message at the chosen rate The receiver can configure all of the messages in figure 25 a number of chosen messages or none 43 4 2 Arduino Two program files were created for use with the Arduino Due 1 MFFM Arduino ino 2 QuantifyClockCycles ino Logging the Arduino s clock drift can be done through the MFFM Arduino program in addition to the ability to modify this program to alternately send frequency data The relevant LabVIEW configurations have already been made so choosing the right main loop function and relevant ISRs is simple The QuantifyClockCycles program is detailed in the PPS ISR Processing Time sub section below and its primary purpose is to ascertain the number of clock cycles that elapse during ISR calls in order to offset collected data values and bring the measurement trueness closer to the actual value 4 2 1 PPS ISR Processing Time The PPS time collection is an ISR that is triggered by the rising edge of a digital pin connected to the PPS output of a GPS receiver The purpose of the ISR is to quantify the ISR s processing time in order to offset clock jitter data from the Arduino Due The code for this can be
30. atural and induced effects of clock sway in crystal oscillators The varying effects of time temperature natural forces gravity pressure voltage and other factors are analysed on three types of crystal oscillators Room Temperature Temperature Controlled and Oven Controlled The analysis provides a useful measure of the time interval at which the clocks should be disciplined to stay true to a more precise time source This study is directly applicable to the measuring MCU Arduino Due as it uses an on board crystal oscillator to keep track of time which is subject unwanted clock sway effects over longer periods of time Relative timing characteristics of the one pulse per second 1PPS output pulse of three GPS receivers 52 Three GPS receivers are analysed to determine the confidence associated with the precision of the Pulse Per Second output of each As atomic clocks are extremely expensive and difficult to obtain the author used a scientific grade GPS as the reference clock which specified a much more precise PPS deviation compared to the other two GPS receiver units The clock deviation of the other two GPS units was measured relative to the reference clock Statistical analysis must be performed to determine the trueness of the disciplined clock in this thesis This paper provides an understanding of what type of analyses must be performed Allan Deviation excluded Accurate measurement of the mains electricity frequency 44 The a
31. bal Positioning System The Global Positioning Network had its inception in 1973 to replace the Navy Navigation Satellite Systems 22 The GPS satellite network was operational on 27 April 1995 with 24 satellites orbiting the globe twice a day RF waves propagate at the speed of light 299 793 077 ms The GPS signals are sent from space at a height of 20 200 km but this distance varies as the satellites follow an elliptical path 22 GPS satellites have an orbital period of 11 hours and 58 minutes 22 Each GPS SV is equipped with four atomic clocks two rubidium and two caesium 22 The initial generation of GPS SVs was Block II with the first satellite launched into orbit in February 1989 and final on October 1990 Since then several other satellites were launched to provide improvements to the existing infrastructure e Block IIA 13 satellites of this series still orbit the Earth with the final satellite being launched on November 1997 This block was designed to allow a longer period of independent operation with control segment contact 180 days 28 Satellites in this block only operated on the L1 frequency 29 e Block IIR 12 satellites from this series were launched since July 1997 as replenishment satellites to replace older satellites that were about to fail or already failed e Block IIR M 8 satellites were launched in this series with the final being launched in August 2009 These satellites included the
32. band primarily propagate as space waves which require a direct line of sight GPS broadcasts a Pulse Per Second signal to GPS receivers This signal is generated by an atomic clock on board each GPS satellite and is subject to transmission jitter and processing jitter Transmission jitter comes from several sources the largest being from the space wave propagating through space and Earth s atmosphere As an RF wave passes through the troposphere and 20 ionosphere its speed is reduced At a height of 80 400km the RF waves pass through the ionosphere which refracts the GPS satellites signals 32 Because the velocity variations through the ionosphere are known at GPS transmission frequencies GPS receivers mostly correct the error associated with ionospheric delays 32 Tropospheric delays are caused by refraction and a further change in the propagation medium WAAS enabled receivers may receive atmospheric condition data over different regions which allows the receiver to operate at a much greater accuracy in its atmospheric delay calculations 32 GPS Receiver E lonosphere mm Troposphere Satellite Visibility Region Signal Transmission Path gt satellite Orbital Path Figure 7 GPS Satellite signal transmission path diagram Figure 7 depicts the orbital path of a single GPS space vehicle on a fixed axis As the satellites traverses its orbital path between the apogee and perigee the signal that travels to t
33. bytes to read 50 0002 115200 at 50 0001 data bits stop bits 50 0001 a 10 B uy 3 gt Data Read i flow control Rate ms None 5 Looping Clear Data Time Sent Serial Messages NMEA Packets Sent Autoscroll Protocol packet numbers 1 1 Communications Configuration Packet 32 Location Initialisation Packet 33 PPS Configuration Packet 4 Sensitivity Mode Packet Max Conecion 5 Receiver Configuration Packet y t 6 Reset Configuration Packet SRLS Tee 7 Available Drift Logging CSV File Location A o 8 Available 14 S E 39 Receiver Message Output Packet Always 9 in case Max Correction more protocols are added this should always be the Frequency Logging CSV File Location Under frequency Time ms final message y lo i A Figure 22 Control Panel VI user interface As seen in the main project GUI in figure 22 the user can select which serial port to connect on corresponding to the port they connected to the Arduino either via mini USB or RS232 connection Baud rate is set to 115200 as default and will cause errors if it is changed The LabVIEW program expects 1 byte to be read at a time and hence the data read rate has been set to 5ms so that it can collect all the data at the port in time The data that is sent to the Arduino is shown in the Sent Serial Message string indicator and the data that are received back including all handshake characters are displayed in the In
34. eivers in RF Components and Circuits Elsevier Newnes 2002 pp 41 46 Wikipedia Aliasing 2013 Online Available http en wikipedia org wiki Aliasing Accessed 24 October 2013 B A Olshausen Aliasing 10 October 2000 Online Available http redwood berkeley edu bruno npb261 aliasing pdf Accessed 28 October 2013 Hewlett Packard Fundamentals of Quartz Oscillators 3 September 2000 Online Available http literature agilent com litweb pdf 5965 7662E pdf Accessed 1 August 2013 Arduino Arduino ArduinoBoardDue 23 October 2012 Online Available http arduino cc en Main arduinoBoardDue Accessed 29 July 2013 GlobalSat EM406 A Product User Manual GlobalSat Technology Corporation 16 February 2010 Online Available http elmicro com files sparkfun em406a ug pdf Accessed 20 August 2013 SparkFun Electronics Copernicus Il DIP Module 2013 Online Available https www sparkfun com products 11858 Accessed 17 August 2013 D L Mills Network Time Protocol Version 3 Specification Implementation and Analysis March 1992 Online Available http www ietf org rfc rfc1305 txt Accessed 9 October 2013 64 14 15 16 17 18 19 20 21 22 23 24 25 26 D L Mills Network Time Protocol Version 4 Protocol and Algorithms Specification June 2010 Online Available http tools ietf org html
35. el Most frequency counters enable the user to monitor frequencies but monitoring frequency variations at a high resolution is often expensive Electronic systems that measure frequency also have to generate a local time base to calculate the frequency upon All time bases suffer from the effect of frequency jitter which makes the timing source deviate from the nominal second by a quantified amount Modern systems have improved drastically and have relatively insignificant jitter for most timing applications but high precision applications require a quantification of this source of timing error The purpose of this thesis is to document the background implementation testing results and identified future improvements for a frequency meter that can record minor fluctuations of the grid frequency By achieving this objective the grid supply and demand data can be logged and used for several applications such as network forecasting or maintaining nominal grid frequency An extensive research period was required to determine key design facets pertaining to the frequency meter Key identified tasks included choosing a timing source finding a suitable software development platform and associated hardware developing a graphical software implementation that displays real time frequency fluctuations contingency alarming for nominal frequency deviation events communications design between the frequency meter and the PC quantifying clock precision and eval
36. en establishes a connection 26 PTP is a more recent timing protocol implementation designed to provide a higher standard of precision than NTP PTP is specified under IEEE1588 2008 PTP is primarily intended to provide a time base more accurate than NTP in areas where GPS signals are either inaccessible or too costly PTP works on a similar principle to NTP but has additional protocol provisions for estimating propagation and synchronisation delay between the server and the client Hardware provisions however must be made to provide this and can be costly for simple applications Both protocols are susceptible to the same transmission related delays like any other networking protocol Latency is the measure of transmitted signal s delay and is typically quantified using algorithms that compute the delay 25 The number of hops is a significant contributor to the effect of latency as it reduces end to end synchronisation performance 27 The data rate limit is another factor that may limit the transmission of the NTP packet but in most modern networks is not an issue Line coding delay 25 comes from both the client and server and is the time that the sender and receiver take to compute and assemble an outgoing packet as well as the time taken to decode generate checksums and error check an incoming packet Precision on the order parts per million is typical of NTP but the jitter may vary on the order of tens of thousands of ppm 19 2 1 5 Glo
37. erial print PrevMicros Serial print Serial printIn countedges Ie if EdgeChanged true EdgeChanged false Serial print PulsesCounted Serial print Serial print Initial_ Gap Serial print Serial print Final_ Gap Figure 40 Statements given in main Arduino loop 58 The commented out code in the main loop pertains to the jitter analysis variables The rest of the code is used in frequency metering Further comments pertaining to this are available within the program s comments found in Appendix A The LabVIEW setup involves setting up three VI files Control Panel NMEA Configuration and NMEA Packet Decoder The rest of the project library will provide support functions and must not be modified The Control Panel is relatively simple to set up simply select the Frequency Logging Mode button If the Drift Logging Mode button is concurrently selected Frequency Logging will take precedence The operator must specify the location of where you want to save the data set and append the file name with the csv extension The frequency change threshold value on the Control Panel is recommended to be set at 0 2Hz but this may be changed depending on the range of outlier data the user may experience Results on the Arduino indicated that any change over 0 2Hz tends to be an outlier with typical values averaging around 1Hz The NMEA Configuration file allows the user to change the NMEA packets i
38. es all the technical specifications and while lacking a set up diagram relevant to the Copernicus Il DIP module it infers enough to be able to connect it safely and for reliable operation Indoor positioning based on global positioning system signals 53 This paper analyses the issues with indoor placement of GPS receivers and inability for signals to propagate well through solid walls The author proposes a repeater based indoors GPS system where the repeaters are placed outside and carry the signal indoors via a cable This document is significant as it covers the primary issues of signal propagation to indoors systems and analyses the GPS error ability to capture signals clock bias and positioning accuracy when the GPS is placed indoors ISO 5725 1 44 This document is an international standard It provides set definitions on accuracy precision trueness bias and other key terms relevant to scientifically accurate measurements The terms used in the project that are listed in this standard will carry the same meaning to avoid confusion in terms such as accuracy and precision It will also serve as an excellent point of note for defining errors as they are measured establishing resolution uncertainty and commenting on the differences between measured values and expected values 71
39. for a physical connection to log the data This would enable the user to simply attach the instrument to a power socket let it record for as long as is needed and return to obtain the frequency meter when the measurement point is no longer required 62 8 Conclusion The developed frequency meter achieved all the primary objectives and all set minor objectives The Arduino Due MCU crystal oscillator was found to be performing well within it s data sheet timing jitter specification allowing it to perform well in most applications where a stable time base is critical The GPS timing implementations did not vary greatly in the resultant Arduino timing jitter data sets inferring the GPS signal jitter is negligible at the Arduino Due s sampling rate specification The studies undertaken in this thesis were broad and developed necessary engineering skills A broad range of learning outcomes were gained such as time management task prioritisation physical and theoretical electronic design implementations and their respective limitations hardware cost benefit margin analysis awareness of available software implementations project schedule management and the ability to work independently to a stakeholder s project specifications The frequency meter was able to detect the mains frequency at a specification higher than the set objective of metering mHz level fluctuations on the grid By utilising an MCU with a high sampling rate frequency changes were
40. form of frequency detection to maintain correct operation There are many systems available both commercially and for home use to detect the frequency of various periodic waveforms 1 2 1 Measurement Error Sources Frequency measurement can be performed in several ways depending on the frequency range that has to be measured and the shape of the waveform Modern forms of frequency detection include counting involving a gating period 5 frequency counters and heterodyning 5 frequency conversion Each method is subject to several issues that affect the accuracy precision and measurement error of a measurand In modern systems timing source jitter is an issue that creates measurement error and contributes to the cumulative time interval error Jitter shown as the interval j in figure 2 is the periodic deviation from the nominal period of the source waveform It is usually expressed in parts per million PPM as expressed in equation 1 A ppm specification defines how many microseconds the signal may be off the nominal value For example a 1 part in 20 million 0 05 ppm specification will correspond to a 50ns jitter at a frequency of 1 Hz whereas a 30 parts per million jitter specification will correspond to 30us from the nominal signal period Because jitter is often quantified on the order of micro seconds or less this specification becomes useful 1086 xAf Frequency Source Jitter ppm 1 10 Period T Period T
41. function wipes all buffered data on the UARTs e NMEA Response short int type returns 1 0 1 or 2 Verifies the NMEA packet that was sent to the Copernicus II was correct Returns 1 for unknown message returns O for invalid packet returns 1 for valid packet and returns 2 for response time out e GetNMEA void type returns nothing Gets an NMEA message from LabVIEW and stores it in a buffer e SendNMEA void type returns nothing Sends the buffered NMEA message to the Copernicus ll and appends the carriage return and line feed characters e Serialit Protocol Request void type returns nothing is replaced by 0 or 2 corresponding to the active Arduino to LabVIEW serial line Requests the NMEA packets from LabVIEW sequentially and communicates in a custom protocol e Reset Copernicus Comms void type returns nothing Changes the Copernicus Il baud rate to the original 4800bps e GetFrequency void type returns nothing ISR function Generated when a pulse is generated on digital pin 9 by the periodic pulse generator that has the same frequency as the mains supply The Arduino will wait for 10 seconds to elapse before it enters manual configuration mode In automatic setup mode a protocol has been developed to allow the Arduino and LabVIEW programs to communicate NMEA packet data for initialisation of the Trimble Copernicus Il GPS receiver The Arduino will poll LabVIEW for the NMEA packets it requires The NM
42. gure 14 Various elevation mask angles of GPS Satellites referenced to a North Pole positioned receiver Figure 15 EM406 A GPS receiver module Figure 16 EM406 A connector cable Figure 17 Figure 18 Figure 19 EM406 A cable connection diagram Arduino Due RS232 communication compatibility shield EAGLE schematic GPS relative frequency stability analysis circuit Eagle Schematic Figure 20 Inverted NAND gate voltage output Figure 21 Figure 22 Figure 23 GPS receivers in phase default 1us and 4us length PPS signals Control Panel VI user interface NMEA Configuration VI user interface Figure 24 NMEA checksum generation illustration Figure 25 NMEA Packet Decoder VI user interface Figure 26 ISR clock cycle quantifying program Figure 27 Figure 28 Figure 29 Previous Arduino library implementation of micros New interrupt functioning implementation of micros Arduino frequency metering program flow chart Figure 30 Histogram of PPS generated time intervals on the Arduino Due Copernicus Il PPS source Figure 31 Figure 32 Figure 33 Arduino 48 hour mean centered jitter graph Copernicus II source Arduino 48 hour TIE graph Copernicus II PPS source Histogram of PPS generated time intervals on the Arduino Due EM406 A PPS source Figure 34 Arduino 48 hour mean centered jitter graph EM406 A PPS source Figure 35 Arduino 48 hour TIE graph EM406 A PPS source Fi
43. gure 36 Physical frequency meter setup Figure 37 Figure 38 Figure 39 Trimble Copernicus Il connected pins diagram PPS ISR for timing precision analysis PPS ISR for frequency metering Figure 40 Statements given in main Arduino loop Figure 41 Figure 42 Figure 43 48 Hour frequency log graph Under frequency Event 1 graph Under frequency Event 2 graph V List of Tables Table 1 GPS system error source table Table 2 SWIS grid operational parameters Table 3 SWIS target recovery times for grid frequency variations due to contingencies Table 4 Arduino Due specifications Table 5 Arduino Due pin connections Table 6 Pulse generation circuit s frequency tracking offset with variations in mains supply power quality Table 7 Trimble Copernicus Il GPS receiver specifications Table 8 Copernicus Il project default NMEA packet configuration with checksums Table 9 EM 406A GPS receiver specifications Table 10 EM406 A project default NMEA packet configuration with checksums Table 11 Temperature data for the jitter logging time interval Table 12 PPS triggered Arduino 1 second timing interval data Copernicus Il PPS source Table 13 PPS triggered Arduino 1 second timing interval data EM406 A PPS source Table 14 Under frequency events detected during frequency meter performance tests VI List of Equations Equation 1 Jitter calculation based on the nominal frequency f and peak frequency variation Af
44. hapter heavily as the recommendations are drawn from the concluded findings It outlines future improvements that may not have been able to be implemented in this project due to various factors but would be viable in further studies 15 2 Background 2 1 Timing Methods Timing is of crucial importance in many applications and time tracking has numerous methods Modern systems can derive their time from various sources such as computer network protocols GPS signals radio transmissions the known period of the mains power supply signal or various types of crystal oscillators In frequency measurement systems the ability to specify measurement precision accuracy and error comes by relating the systems performance to a standard with known timing characteristics such as an atomic clock No perfect system exists to keep track of time but the clock drift clock deviation from the perfect time model of all systems is able to be quantified relative to very accurate and precise timing standards With the advent of atomic clocks many technologies have been developed that synchronise their timers to rubidium or caesium standards More recently ytterbium clocks have been developed that outperform previous clock standards 17 While caesium clocks take five days to reach peak performance ytterbium clocks can achieve this in one second 17 Precise timing technology has drastically changed over the recent years and further improvements are continuall
45. he receiver will undertake a non linear path due to the refractive index changes between atmospheric layers This results in a variation of the signal s transmission path length to the receiver which proportionally creates a variation in timing signal jitter 21 Error Source Error Variance lonospheric effects 5 meters Satellite orbital shifts 2 5 meters Satellite clock errors 2 meters Multipath effects 1 meter Tropospheric effects 0 5 meters Calculation and rounding error 1 meter Table 1 GPS system error source table 32 Table 1 explains the variation in GPS signal error due to multiple sources Variations in the ionosphere and orbital altitude of the GPS space vehicle account for the largest component of the GPS error Modern GPS receivers especially those with WAAS enabled correction can account for most of these errors to improve the accuracy of the received signal data These GPS error sources contribute to 15 meters of dilution of precision In WAAS corrected GPS receivers if a WAAS correction is able to be obtained this error goes down to 3 5 meters 32 This enables GPS receivers to have PPS accuracy on the order of parts per billion 33 22 2 2 Grid Parameters It is important to know the frequency of the grid as all electronic equipment that is connected to it has a certain operating frequency requirement The frequency may dictate the electronics efficiency operating lim
46. heedmartin com au content dam lockheed data space documents gps GPSII _FactSheetFINAL1 pdf Accessed 11 November 2013 U S Department of Defense DoD Permanently Discontinues Procurement Of Global Positioning System Selective Availability 18 September 2007 Online Available http www defense gov releases release aspx releaseid 11335 Accessed 19 August 2013 Kowoma Sources of Error in GPS 2013 Online Available http www kowoma de en gps errors htm atmospheric Accessed 11 October 2013 Trimble Copernicus Il GPS Receiver Reference Manual 2 September 2011 Online Available https dIlnmh9ip6v2uc cloudfront net datasheets Sensors GPS 63530 10 Rev B Manual Copernicus Il pdf Accessed 17 August 2013 Standards Australia AS 60038 2012 Standard Voltages SAI Global Limited Sydney 2012 Western Power Technical Rules 23 December 2011 Online Available http www westernpower com au documents aboutus accessarrangement 2011 WE_n6800 863 v9E TECHNICAL RULES OF 23 DECEMBER 2011 pdf Accessed 29 July 2013 Atmel AT91SAM ARM based Flash MCU 9 September 2012 Online Available http www atmel com Images doc11057 pdf Accessed 21 August 2013 Arduino Arduino Compare 2013 Online Available http arduino cc en Products Compare Accessed 8 October 2013 Freescale Semiconductor Inc Microcontrollers 2013 Online Available http www freescale com webapp sps site ho
47. in PPS jitter The performance of this design was tested in both ICAPS and physically 37 Om Ove da o KD 560 240 9 00 360 400m 7 00 zs 2 E Sie 1 60 160 5 00 Ss E 3 2 jf e c 400m 3 60 3 00 240 5 60 1 00 1 50 2 50 3 50 450 time in seconds 500m Gun 010 v 0 0 gt Y Home x 1 00000 Y 159 252m deltaX 374 444m Guna tox v 1 1 Y Home X 1 37444 Y 3 30881 delta Y 3 14956 Figure 20 Inverted NAND gate voltage output This design however had the major limitation of requiring both signals being in phase with each other which would give only the most narrow signal as the output It s proposed that this design may still be able to work if it is modified in future works to delay the phase of one signal by 180 Tek Ma O Scan CH2 Coupling B Limit 1 40MHz Volts Div Coarse voltage Invert CH1 2 00 CH2 2 00 M 5 frnns CH1 00 v 30 Sep 13 20 27 lt 10Hz Figure 21 GPS receivers in phase default 1us and 4ps length PPS signals 38 4 Software 4 1 NI LabVIEW 2013 National Instruments LabVIEW 49 is a dataflow programming environment based on the G programming language LabVIEW 49 offers the standard functionality of most programming languages and incorporates a gra
48. istance of 5033km Under the assumption of reception being available in Perth a latency of 16 78ms would be observed due to a transit delay of approximately 1ms for every 300km the signal has to traverse 16 Surface wave signals paths typically propagate up to 1500km 23 At distances greater than this the signal becomes a sky wave signal and refracts off the ionosphere At distances of 5000km or greater the signal s reliability becomes extremely poor and unusable due to the signal s irregular pathways 23 Due to the lack of signal integrity in Perth alternate technologies were considered 17 2 1 3 Crystal Oscillators Crystal oscillators XOs have been used in many electronic devices to keep track of time The quartz crystal oscillator has the property of piezoelectricity which provides a link between electronics and mechanical distortion of the crystal lattice The XO has stiffness and some elasticity in its bonds which allow the crystal to resonate like a tuning fork The frequency at which the crystal oscillates is determined by the size shape and cut of the crystal and the frequency drift that the crystal may experience with temperature is determined by the size of the cut C1 b C2 L1 R1 ION Figure 6 XO circuit model a and passive element equivalent model b The equivalent model in Figure 6 has four parameters where C1 is the capacitance due to the electrode holder and leads C2 is the notional capac
49. itance the inductance L1 is related to the oscillator s mass and the resistance R1 is due to bulk losses The XO is typically inserted into an electronic feedback loop where it oscillates at it s resonant frequency and is amplified at the output The XO model in figure 6 demonstrates that the XO behaves like a band pass filter so when coupled to an external amplifier it is possible to create a system with gain and positive feedback Because C2 and L1 behave like a second order electronic system they will have a defined resonance frequency f 1 fo 2nvL1 C2 4 XO frequency stability can be reduced due to the effects of aging varying power quality gravitational force vibrations electromagnetic interference retrace essentially a cold start temperature and pressure 9 The temperature of a crystal is of greatest importance as it has the 18 greatest effect on oscillator stability 9 Three commonly used variations of XOs are affected by temperature in different lengths The room temperature XO RTXO has no method of temperature compensation the temperature compensated XO TCXO is cut in a way to minimise changes to its frequency stability due to temperature changes and is encased to minimise abrupt ambient temperature changes The oven controlled XO OCXO has the most precise method of oscillatory frequency stability control 9 OCXOs control the temperature variation the crystal is exposed to through a feedback temperature contro
50. its or it may provide an alternate use such as providing a time base in digital timers While generally the mains frequency is not used to provide a time base due to the low cost and high availability of XOs it s nonetheless important for many applications To measure the frequency of the grid the parameters of the grid must be known In order to design a metering system that will not damage itself due to fluctuations in the grid information was taken from Western Power s website and the SAlGlobal Standard Voltages document 34 Western Power is the power utility company operating in the SWIS region of Western Australia While Standards Australia defines the nominal voltage and frequency values for all of Australia 34 in AS60038 Western Power specifies it s own operating standards in the Technical Rules document 35 Tolerance Nominal Value Min Max Min Max Mode Voltage 240V RMS 10 6 226 V 254 4 V Frequency 50 Hz 49 8 Hz 50 2 Hz SWIS Frequency 50 Hz 49 5 Hz 50 5 Hz Islanded Table 2 SWIS grid operational parameters 35 Table 2 shows the operating frequencies for standard and islanded grid connections and the operating limits for the grid voltage The accumulated synchronous time error is defined as the difference between Western Australian Standard Time and the time measured by integrating the instantaneous operating frequency of the power system 35 In the
51. itude increase due to lower sampled waveform frequency 12 1 2 3 Frequency Counters High frequencies can be measured through frequency counters and several modern technologies allow this such as data acquisition cards and microcontrollers Most frequency counters derive their time base from a crystal oscillator XO which oscillates at a known frequency 5 The measured input frequency is then ascertained by counting the number of periods in a time period generated by local frequency counter s XO The frequency counter method is generally very precise in the short term but long term measurements will be affected by the jitter of the instrument s time base source Modern frequency counters can currently cover up to a range of 100GHz 5 but are typically expensive for high range frequency measurements 1 2 4 Heterodyning Heterodyning is the process of mixing two different frequencies to produce a frequency that can be used in signal processing 6 The output frequency that is produced is called the heterodyne Historically heterodyning was used to process high frequency signals by mixing them into a heterodyne that could be processed by the technology that was available Heterodyning is still used in RF applications 6 but as frequency counter technology keeps improving to provide higher sampling rates and costs go down heterodyning is more suited to fill very high frequency detection applications 1 2 5 Aliasing No matter which method
52. l system which allows the crystal to perform with significantly less variation in operating frequency Most consumer electronics utilise RTXOs due to their very low cost and ability to keep a timing accuracy within the order of parts per million 24 2 1 4 Time Protocols NTP and PTP are protocols designed to synchronise computers over a general purpose computer network to a high precision clock standard Both protocols use a server client architecture to transmit UTC time over packet switched networks As with any networking protocol packet errors throughput size latency variation and packet loss can cause the performance of the system to drop 25 Applications that require reliable precise timing will be affected by this performance drop NTP is the most common time synchronisation standard in computers today The IETF maintains NTPv3 the most common implementation of NTP RFC 1305 13 provides the specification implementation and analysis of NTPv3 The newest implementation of NTP is NTPv4 14 NTP has several topologies including server client where the client periodically polls the server for the time and calculates its own clock offset symmetric active passive mode NTP data is polled via peers on the network broadcast multicast mode a server sending NTP packets periodically to a group of clients or the entire networks and manycast mode a client polls several NTP servers to determine the server with least latency to connect to th
53. level Reset packet data provides a configuration package that allows the receiver to enter stand by mode when the GPS receiver is not required If the ephemeris data is less than 4 hours old a system hot start is possible and the receiver will find a fix within 3 seconds 33 The system will activate from stand by mode through activity on the NMEA IN port RX B The LabVIEW NMEA Configuration program has been designed to automatically calculate modify and append checksums to each packet required for the Copernicus ll 33 3 4 GlobalSat EM406 A Figure 15 EM406 A GPS receiver module 11 The GlobalSat EM406 A shown in Figure 15 is a GPS receiver with the specifications listed in table 9 The receiver was provided by Murdoch University s Engineering amp Information Technology department for PPS jitter analysis in this project Chipset SiRF Star III Input Voltage 4 5V 6 5V DC Communication Protocols SiRF NMEA USER1 Channels 20 All in view tracking Sensitivity 159dBm Logic Level OV Low 3 3V High Table 9 EM 406A GPS receiver specifications 11 The EM406 A GPS receiver was communicated to through a TTL serial connection from the Arduino Due The chosen communication protocol was NMEA to maintain a set standard among the GPS receivers It was however discovered that while the automated output messages of the EM406 A are the same as the Trimble Copernicus II the configuration packet
54. lications in the analysis of load management network forecasting generator response to load variation and contingency analysis 14 1 4 Thesis Outline In addition to the abstract introduction background and conclusion the thesis has five key chapters e Hardware Implementation This section discusses the hardware chosen for the project the specifications that are relevant to each component how it will contribute to reaching the project s goals and how the hardware is connected for various analysis purposes e Software The libraries used in the software implementation their purpose in the project and any additional libraries developed are discussed in this section to detail the approach taken to meet the project s goals The two primary programming languages used are G LabVIEW s graphical programming language and the Arduino programming language a Wiring language derivative e Timing Precision One of the primary goals of the project was to quantify the precision of the frequency meter This is done in this section by analysing the relative clock drift data between several implementations such as the Arduino Due s XO the Trimble Copernicus II GPS receiver and the EM406 A GPS receiver e Frequency Meter The frequency metering system is described in the final chapter in the main body including its overall performance and limitations e Recommendations and Future Improvements This chapter ties into the conclusion c
55. lifier Periodic Pulse Gen 21 10 2013 12 42 19 Sheet 1 1 4 Figure 11 Frequency tracking pulse generation circuit Eagle schematic The frequency tracking circuit is designed to periodically generate digital pulses that are at the same frequency as the incoming 50Hz sinusoidal waveform A low pass filter attenuates the incoming signal s frequency past the cut off point of 500Hz in order to reduce high frequency noise while minimising attenuation at the 50Hz frequency A 1N4148 diode is connected with the anode to ground and the cathode connected to the T1 transistor s base This diode allows current to flow through the capacitor C1 and resistor R4 during the negative cycle of the input waveform The diode s action prevents damage to Transistor T1 as the Emitter Base voltage cannot exceed more than 6V 46 T1 switches on when the base emitter voltage is above 0 7V 46 Due to the positive non zero voltage that the transistor turns on at the square wave that is produced has a mark space ratio that is slightly less than 5096 but still easily long enough on the order of milliseconds to be measured by the Arduino which can measure on the order of microseconds 36 Figure 12 displays the transistor s pulse triggering but it appears that the square wave s positive and negative edges is very close to zero due to the larger AC signal voltage 29 Tek AL Trig d M Pos 200 0 us MEASURE
56. llent clock rate and considerably large amount of SRAM Flash memory In the future however improved MCU models will be released that operate at the same clock rate or higher with even more powerful specifications The main improvement of a new MCU would be a faster sampling rate as right now the resolution is set at a maximum of 1ys This is effective for long term jitter logging but for analysing very small short term changes in clock jitter such as that claimed by GPS receivers this is much too high Hence it is recommended that hardware improvements are performed when a cost effective upgrade is available Among other hardware upgrades improved filtering could be designed to replace the first order low pass filter on the Arduino frequency detection shield A second order or better band pass filter could be created to attenuate signals outside the nominal range but generally an improved low pass filter design should be just as good as the mains frequency is relatively low The transistor amplifier circuit on the frequency detection shield performs well and detects the frequency of the mains supply but could also be improved by designing a zero crossing detector circuit that performs with minimal propagation delay and is reliable Lastly the data is currently transmitted either only to an SD card or directly to a PC terminal A GSM shield would add wireless data collection capability and could send data periodically to a database to remove the need
57. mepage jsp code PCMCRO1 Accessed 8 October 2013 Microchip Technology Inc Microchip MCUs 2013 Online Available http www microchip com pagehandler en us products picmicrocontrollers Accessed 8 October 2013 Freescale Semiconductor HC11 Microcontrollers 07 2005 Online Available http www freescale com files microcontrollers doc data_sheet M68HC11E pdf Accessed 2 November 2013 66 41 Arduino Arduino FAQ 2013 Online Available http arduino cc en Main FAQ Accessed 9 October 2013 42 Creative Commons Creative Common Attribution Share Alike 3 0 Australia 2013 Online Available http creativecommons org licenses by sa 3 0 au Accessed 9 October 2013 43 Free Software Foundation GNU Lesser General Public License 29 June 2007 Online Available http www gnu org licenses Igpl html Accessed 9 October 2013 44 D Ibrahim Accurate Measurement of the Mains Electricity Frequency in International Conference on Electrical and Electronics Engineering Bursa 2011 45 CADSoft Eagle 6 5 0 Download Page CADSoft USA 7 August 2013 Online Available http www cadsoftusa com download eagle language en Accessed 13 August 2013 46 Multicomp BC547B General Purpose Transistor 12 May 2008 Online Available http www farnell com datasheets 410427 pdf Accessed 2 October 2013 47 OnShine ONSHINE G P S Antenna ANT 555 GPS Active
58. n order to speed up the Copernicus II initialisation time and modify its functionality It is recommended that the values be left as default for the most part The user should primarily use this VI to insert their GPS coordinates in order to obtain a GPS fix faster As long as the coordinates are within 100km of the correct location they will be valid If poor signal reception is experienced it is possible to change the receiver to High Sensitivity Mode If a lock onto less than 4 satellites is established for a long period the NMEA configuration can be changed to Stationary dynamic rather than Land Sea or Air This provides a PPS time base from 1 satellite but at the cost of increased PPS jitter The NMEA Packet Decoder may provide more automated GPS messages than the GGA message that is typically used The user will have to modify the LabVIEW or Arduino program s to parse this data depending on it s intended use It is recommended that the default settings are used and the GGA message is always selected for output Following these setup steps the programs are ready to run if a physical connection to the Arduino is established 59 6 2 Performance Results 48 Hour Frequency Log Data Frequency Hz Frequency Hz 6 Time Elapsed Seconds Figure 41 48 Hour frequency log graph Figure 41 demonstrates the overall 48 hour mains supply frequency data set that the Arduino logged The mean data set value was 50
59. nnncnninnons 27 3 3 Trimble Copernicus ll ueseeseeseseseses eese en nnn nnns aE tern nass sisse states asas esses satanas asas ensi na 30 AS A cetsitechextocsatecudceducteenseh as e Ene esa nE a e a rE a E 34 3 5 MAX232 Communications Shield ttn rh tiet nasasaad adane inikan iiaeaa 36 36 GPS Jitter Analysis CIFCUIE sssini ae aia aE aaa e a LER vea Pen e LAN RS RAR 37 NON 39 MINES 2093 sica 39 41 1 Control Panel Mica did dba Ee dpa REOR 40 74 1 2 Supporting FUncklOns ie dass 41 1a PA aN OUI o soe ERE EP 44 42 1 PPS ISR Processirig Titme eni eor ohnehin ore aE Eumene ALE Aaea RASA NERA RU REM NAR ar Leur ai EEEE 44 4 2 2 Alternate Microsecond Function Implementation ooooconcnncnccononaonnnnnnncnnonanonononcnnnononnnnnnnncnnnnnnnanonncos 45 4 2 3 Frequency Meterifig uiii eren ete tones i rient dastaki Erian iai An ka sho tna dni ausge ENA KARARAN EKARO Ent KE iaia 46 5 TIMING PEC CISIO AETA AEA E TAA T A A E A AA A EAE A 49 5 d Arduino Frequency Stability Data encadena e ERE 49 5 1 1 Clock Drift Relative to Trimble Copernicus ll coconooconocncconononoonnnncnononononnnoncnncnnonnnncnncnnnnnonnnnnnncnncnnnns 49 5 1 2 Clock Drift Relative to GlobalSat EM406 A oooocccocccccocaconcnononccononononanonononnnnnnn nono nennen enne enn snnt 53 o FrEQUEN CY Meter zit RR REPERTA a e RE RARE a a EE baa deedsnae PER buaceetasuancodsiuadswacnlandeascye 56 o TTE PI p nin aniinds
60. ore Microseconds Elapsed Between PPS Interrupts Figure 33 Histogram of PPS generated time intervals on the Arduino Due EM406 A PPS source The histogram data in figure 33 displayed a similar result to section 5 1 1 with a large distribution being centered around one value 999 994 and mostly a 1 us jitter long term about this value Similarly no erratic variations in Arduino timing jitter were detected 53 48 Hour Clock Jitter from the Mean Value 4000 2000 e 2000 4000 Seriesl 6000 8000 Cumulative Clock Jitter us 10000 12000 Elapsed Time Seconds Figure 34 Arduino 48 hour mean centered jitter graph EM406 A PPS source In an ambient temperature affected environment the Arduino s jtter around its mean timing value displayed a very similar start to the data in 5 1 1 but was dissimilar in the fact that it appeared similar to a sinusoidal waveform indicating a periodically repeating nature The temperature data in the start of section 5 indicates that over the 48 hour period a repeating set of data should appear over the first 48 hours and a larger trough should be displayed due to the highest temperature data being on the final day of the recording This can be seen by the trough around the 138381 second mark dipping lower than the previous trough While observations can be made upon this data improvements could be made in the future to simulatenousl
61. ore precise time source or simply quantify the error associated with the XO and compensate for this error respectively 1 3 Thesis Purpose This project envisages building a precise metering device to monitor small mains supply frequency fluctuations on the order of mHz or better While power utility companies internationally choose to keep the mains supply frequency at either 50Hz or 60Hz they have no control over the time at which customers may connect or disconnect loads As loads are connected and disconnected from the grid the generators that provide power to the grid are adjusted to either slow down or speed up to maintain the nominal grid frequency There is a delay involved in the generator s corrective response actions and this delay period gives way to typically minor frequency fluctuations on the mains supply A frequency meter has been designed that has a quantified timing precision The developed meter is based upon an open source electronics prototyping board the Arduino Due 10 Appropriate electronics have been developed that connect to this MCU and various methods of keeping an accurate time base have been considered such as GPS 11 12 NTP and PTP 13 14 atomic clocks 15 radio clocks 16 and crystal oscillators 9 The frequency metering unit is able to store grid frequency data in real time and transmit this data to a computer for analysis of the supply and demand ratio on the grid This high precision meter has app
62. pecifications for the Arduino Due 10 summarised in table 4 CPU Atmel AT91 SAM3X8E CPU Clock 84 MHz Static RAM 96 kB Core Resolution 32 bit Flash Memory 512 kB DMA Availability Yes Operating Voltage Range 7 12V Digital I O Pins 54 Analog Input Pins 12 Analog Output Pins 2 Analog Input Range 0 3 3V Analog Output Range 0 3 3V Analog I O Resolution 10 12 bit 1028 4096 values Sampling Rate 1 MS s Table 4 Arduino Due specifications 36 10 24 Most MCUs available on the market are either 8 bit or 16 bit typically produced by Arduino 37 Freescale 38 or Microchip 39 A Motorola 68HC11 68HC12 40 was also considered for the project Due to the simplicity availability of support and extensive libraries available on the Arduino platform the Arduino Due was a more suitable development platform The Arduino Due is a low cost MCU which can perform 32 bit operations at a clock rate of 84 MHz No other MCU with these specifications or better could be found at a reasonable cost These specifications outperformed most competitors on the market and greatly outperformed all considered competition for its cost The Arduino Due is an open source electronics prototyping platform released under the Creative Commons Attribution Share Alike license and its public libraries fall under the GNU Lesser General Public License 41 42 43 Under the share alike license all work created upon
63. peration Pin Description DO RXO LabVIEW TX via USB D1 TXO LabVIEW RX via USB D7 GPS PPS Signal D9 Mains Pulses Frequency Measurement D10 SD card Power D11 SD card MOSI D12 SD card MISO D13 SD card SCK D14 TX3 EM406A RX D15 RX3 EM406A TX D16 TX2 MAX232 Shield D17 RX2 MAX232 Shield D18 TX1 COPERNICUS 2 D19 RX1 COPERNICUS 2 SPI See Pins D11 D12 D13 Table 5 Arduino Due pin connections 3 2 Frequency Detection Shield Figure 9 TI AC 9131 AC AC step down conversion adapter A TI AC 9131 adapter seen in Figure 9 was utilised to step down the voltage from the mains supply s 240V AC to 3 3V AC A datasheet was not available for the component The product label stated a 240V 3 3V AC AC step down conversion Tek xJ es Trig d M Pos 0 000s MEASURE CH1 Freg 43 30Hz f f Af E p RMS V z CH1 Max 11 0 CH1 5 00 M 10 0ms CH1 J 511mv CH1 vertical position 0 00 divs 0 004 Figure 10 Stepped down AC waveform oscilloscope screenshot Figure 10 displays the observed stepped down no load voltage of the adapter 27 The stepped down waveform was observed at 7 64 Vays This waveform appeared to be at the grid standard frequency of 50Hz 34 The Johnson noise due to the impedance of the output windings is unknown due to no datasheet specification and no shielding is provided hence
64. phical design environment making it ideal for visual debugging and graphical user interface design Real time data acquisition and analysis can to be displayed visually with minimal effort by the programmer due to LabVIEW s extensive libraries With an emphasis on minimising processing cycles on the Arduino so as to avoid unknown variations in processing time contributing as a source of error project relevant information can be passed to LabVIEW for analysis and storage to the PC from the Arduino through a USB MicroUSB RS232 TTL or USB TTL connection The project relevant LabVIEW files are all clustered into the MFFM Thesis lvproj project file where MFFM is an abbreviation for Mains Frequency Fluctuation Metering The VI files in this project are listed as e Control Panel vi e GPS Week and Seconds vi e GPS Fix vi e PadZeroes vi e NMEA Configuration vi e NMEA Checksum vi e NMEA Packet Decoder vi The graphical user interface for frequency metering is available through the Control Panel VI The other VI files are primarily designed for use as supporting functions 39 4 1 1 Control Panel VI The Control Panel VI is the primary graphical interface for use in metering frequency fluctuations Mains Frequency Fluctuation Metering Control Panel VI Input From Arduino Due Verbose Real Time Frequency Value Mains Frequency Hz B 50 0003 conn Port to Connect On 50 0003 1 COM7 50 0002 baud rate
65. put from Arduino Due string indicator Frequency data is displayed in real time as it is collected from the Arduino and is plotted on the Real Time Frequency Value graph The user may alter the time period they wish to display by modifying the Time axis values The Frequency axis scales itself proportionally to the input information but this may be altered by the user The frequency change threshold input allows the user to select how much the frequency is allowed to change from second to second in order to attempt to delete all outlier data that may be generated due to multiple Arduino ISR s running consecutively 40 The user may select either Drift Logging Mode to log the Arduino clock jitter to a CSV file or Frequency Logging Mode Over and under frequency data is logged maximum durations are stored and a Boolean display lights up to indicate these conditions 4 1 2 Supporting Functions While these files are documented within their respective VI programs this section attempts to give a brief description of the purpose of each VI file that supports the Control Panel VI at run time GPS Week and Seconds vi This file provides the functionality of generating the GPS time in seconds since the start of the week Sunday 0000 24 Hour Time The default parameters are UTC 8 Perth Time UTC Offset off The output type is a 32 bit signed integer GPS Fix vi The GPS Fix VI provides a GPS fix determination based on the GPGGA message o
66. ram flow chart In the declaration and initialisation of variables all the variables that are used throughout the program including in ISRs are specified Variables that may have their value changed within an ISR are set as volatile This is done by writing the volatile keyword before the variables data type is declared The advantage of this is that the correct value will be brought up when the variable is called as it is stored in RAM memory rather than a storage register Functions were developed within the program to both reduce the overhead with re writing the same code and to make the code more readable In a brief summary the functions perform the following tasks 47 e HWCDelay void type returns nothing Executes delay function for a specified millisecond value Used to allow HardWare Configuration packets to take place in the Trimble Copernicus Il GPS receiver e NMEA Packet Checker void type returns nothing Primarily used for debugging allows manual input of NMEA packet strings through SerialO to verify the reply packet is received and valid e GetRisingEdge void type returns nothing ISR function Generated when a PPS rising edge is detected on digital pin 8 Holds value from micros function when the interrupt is generated e ClearSerialtt void type returns nothing There are 4 ClearSerial functions where is replaced by O 1 2 and 3 corresponding to the 4 Serial UARTs on the Arduino This
67. rfc5905 Accessed 14 October 2013 HyperPhysics Atomic Clocks 27 April 2009 Online Available http hyperphysics phy astr gsu edu hbase acloc html Accessed 5 November 2013 Wikipedia Radio Clocks 27 August 2013 Online Available https en wikipedia org wiki Radio clock Accessed 8 October 2013 The Hindu Business Line US scientists build world s most precise clock 23 August 2013 Online Available http www thehindubusinessline com news international us scientists build worlds most precise clock article5051703 ece Accessed 8 October 2013 Wikipedia Rubidium Standard 17 August 2013 Online Available https en wikipedia org wiki Rubidium_standard Accessed 5 November 2013 Bureau International des Poids et Mesures BIPM Bureau International des Poids et Mesures 2013 Online Available http www bipm org Accessed 8 October 2013 International Bureau of Weights and Measures Time Signals March 2013 Online Available ftp ftp2 bipm org pub tai scale timesignals pdf Accessed 8 October 2013 BIPM Time Signals March 2013 Online Available ftp ftp2 bipm org pub tai scale timesignals pdf Accessed 8 October 2013 L Tetley and D Calcutt Electronic Navigation Systems Woburn Butterworth Heinemann 2001 M A Lombardi How Accurate is a Radio Controlled Clock Horological Journal p 4 2010 Geyer Quartz Technology Geyer KX 7 Quartz
68. s were slightly different and had to be adjusted These packets are visible under table 10 below Packet Sentence Description Baud Rate SPSRF100 1 19200 8 1 0 38 NMEA protocol at 19200 Baud Debug SPSRF105 1 3E Development Data ON Message SPSRF103 00 00 01 01 25 GGA Message Output output every second Navigation SPSRF104 GPSTOW GPS Initialisation 32 066142 115 837122 10 96000 GPSTOW WEEKNO 12 1 34 Time of Week seconds WEEKNO GPS Week since first Epoch Table 10 EM406 A project default NMEA packet configuration with checksums Implementation appropriate carriage return and line feed delimiters should follow all packet checksums 11 34 The EM406 A had no ability to obtain a GPS satellite fix inside the Murdoch University Engineering building but was able to easily obtain a fix inside a residential house The results were the same for the Trimble Copernicus Il except when the Copernicus Il had an SMA antenna attached In a residential setting the EM406 A had an average time to first fix of 62 seconds from a cold start while it s data sheet specification states 42 seconds 11 Figure 16 EM406 A connector cable Figure 16 displays the EM406 A connector cable which was attached to a pinless header for easier connection to the Arduino via interconnecting wires The connections used by the GPS receiver are shown in Figure 17 The top numbers display the pin number associated with the functions listed at
69. se of three GPS receivers Lx ERE 70 Accurate measurement of the mains electricity frequency 40 oooooococnccccononooncnncnocanonannnnncncnonananonnncnons 70 Electronic Navigation Systems 18 occccononocoonocncconononnnnnononnnononnnnnnncnnonnnnonnnnnnncnnnnnnnnnnnnncnnnnnenonnnnncnnnnnnns 70 Trimble Copernicus Il GPS Receiver Reference Manual 45 enne 71 Indoor positioning based on global positioning system signals 52 coconcococonccanononoonnnncnnnononannnonononanono 71 150 5725 1 40 ioo neto um Et e e a E NU cd Ee e ND ete d ERE REN 7 IV List of Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Measurement precision and trueness relative to a referenced standard Illustration of jitter on a periodic waveform TIE generated by the real waveforms jitter relative to the ideal waveform Gating error magnitude increase due to lower sampled waveform frequency Aliased sinusoidal waveform due to an under sampled signal XO circuit model a and passive element equivalent model b GPS Satellite signal transmission path diagram Arduino Due MCU TI AC 9131 AC AC step down conversion adapter Figure 10 Stepped down AC waveform oscilloscope screenshot Figure 11 Figure 12 Figure 13 Frequency tracking pulse generation circuit Eagle schematic Oscilloscope output of frequency tracking pulse generator Trimble Copernicus Il DIP module Fi
70. ss of a component s specification such as jitter from the nominal operating frequency Accuracy is an umbrella term that specifies the overall trueness and precision of measured data It is defined as the closeness of agreement between a test result or measurement result and the true value 1 This is depicted in figure 1 Reference value Probability Trueness density Value Precision Figure 1 Measurement precision and trueness relative to a referenced standard 1 3 Bias is not defined in ISO 5725 1 3 because it carries a different meaning across different scientific disciplines Bias will be defined for the purpose of this thesis as the difference between the expected measurement and the reference measurement value which is useful for calibrating instruments 3 Measurement error is the result of a difference between the obtained measurement and the true measurement 1 4 The measurement error can be broken down into two components random error and systematic error Random error is the unpredictable error detected over a course of measurements 4 Systematic error is the quantifiable error that can be predicted over a course of measurements 4 1 2 Frequency Detection Many modern systems rely on frequency detection for standard operation Quality control of mains frequency variable frequency drives frequency modulating systems in communications and a multitude of other electrical systems all use a
71. stem XO Crystal Oscillator 1 Introduction 1 1 Measurement Uncertainty This thesis involves analysis of the performance of multiple hardware components It is necessary to define the terminology that will be used in the results in order to create a common understanding between the reader and the author primarily to avoid misunderstanding and or vagueness of terminology ISO 5725 1 Accuracy trueness and precision of measurement methods 1 is the international standard used in this study to define the terminology associated with measurements All measurements that are made have an associated uncertainty to them As a general concept the uncertainty specifies validity of the result of a measurement 2 Quantitative measures of uncertainty may be specified such as variance standard deviation and range 2 The precision of measured data relates to how close together the measured values are 1 Precision can also be broken down to two components e Repeatability How closely the measurements agree under specified conditions that the measurement was originally taken under over a short time interval 3 e Reproducibility How closely the measurements agree with the original set of data under the same process but different instruments over a longer time interval 3 The trueness of a measurement specifies how far the expected measurand is from the reference value 1 The data sheets used throughout this thesis will define truene
72. t RS232 signals between 3 to 15V for false logic and 3 to 15V for true logic The output TTL signal is 0 5V To prevent damage to the Arduino Due the 5V TTL OUT signal coming from the MAX232N to the Arduino Due is reduced to 3 33V through a voltage divider To ensure the input is registered on the MAX232N a transistor amplifier circuit takes the 3 3V serial output from the Arduino Due and converts it to a 5V logic level 36 3 6 GPS Jitter Analysis Circuit In the analysis of the timing jitter on the Arduino Due it was also considered important to test the relative jitter between the two GPS receivers PPS outputs Released under the Creative Commons Attribution Share Alike 3 0 License http creativecommons org licenses by sa 3 0 EM406 A Design By Dusan Sibanic D Pulse_Width_Analyser 23 10 2013 7 01 03 AM Sheet 1 1 Figure 19 GPS relative frequency stability analysis circuit Eagle Schematic 4 This design incorporated a transistor NAND gate The Copernicus II was to generated a 500ms length PPS signal The EM406 A was to create a 1us length PPS signal that s fed into a monostable 555 timer that generates a 500ms signal The theory was to use two similar length out of phase signals feed their outputs through the inputs of a NAND gate invert this output and produce a signal who s length may vary over a long period of time with variance
73. the Arduino platform must be distributed under the same license In a conference paper by Ibrahim 44 a method for metering the mains frequency is proposed that utilises a near zero detector PIC18F4520 MCU and PC link to acquire periodic pulses compute the period between them and log the mains frequency The design utilised an 8MHz XO on board the PIC MCU This design was considered in the planning stage for the mains frequency meter in order to choose the hardware components that will meet the project s goals It is unclear whether the XO on board the PIC MCU is temperature compensated in any form Given that most RTXOs are mounted onto the MCUs PCB it was assumed to be an RTXO This is not an issue in short term frequency measurements but does pose an issue long term Clock stability is able to be quantified by examining the on board MCU drift relative to a more precise timing source such as GPS PPS or an atomic standard The Arduino Due EAGLE 45 schematic file specifies the on board 12MHz XO as a KX 7 quartz crystal with a 30ppm frequency tolerance at 25 C 24 The aging specification is rated at 2ppm year 24 but the manufacturing date of the KX 7 crystal was not able to be ascertained Given that the Arduino Due is less than 2 years old however an upper limit was set giving at most 4ppm additional jitter 25 The Arduino Due pins were assigned as outlined in table 5 for all performed experiments and standard frequency metering o
74. the bottom while the letters B and W correspond to the cable colours black and white EM406A Connector 4 B 32 BB Figure 17 EM406 A cable connection diagram 35 While the EM406 A GPS receiver was suitable for testing relative timing against the Arduino Due it did not provide a datasheet PPS jitter specification lacked an antenna port performed poorly in low signal environments and did not allow the level of functionality the Copernicus Il GPS receiver provided so it was not chosen as the primary timing standard in this project 3 5 MAX232 Communications Shield The MAX232C IC based communications shield was primarily added to provide an alternate communication method to computers While the Arduino Due provides communication through either the Native Programming USB ports or TTL serial 10 the RS232 communication method has no need for drivers and can support older machines attempting to run the metering module 1 2 3 4 A A A A A T B Wo gt T B 96 1 a ait t gt Y Y a t b Y lt c WW gt lt T C ES D D MAX232 15 10 2013 12 50 53 PI Sheet 1 1 1 2 3 4 Figure 18 Arduino Due RS232 communication compatibility shield EAGLE schematic The circuit designed in figure 18 utilises a MAX232N chip which is a 16 pin DIP module The MAX232N can convert up to two RS232 signals to TTL level and vice versa The module requires 5V DC to power it and will conver
75. the noise that may be potentially introduced to the 50Hz waveform is unknown and this is a possible source of error in the final design s metering precision Several designs were considered for the pulse generation circuitry that would be attached to the MCU input such as zero crossing detectors 44 a window comparator circuit and a BJT 46 pulse generation circuit A zero crossing detector generates a pulse every time a periodic signal crosses the zero volt mark Many zero crossing detectors were in fact near zero crossing detectors that generated a pulse at a similar input voltage to the developed transistor amplifier circuit Several considered circuits involving operational amplifiers required voltages that the Arduino could not provide When analysed for the benefit the operational amplifiers would bring over their complexity and limitations they were not necessary in the design of this project Variation from Nominal Voltage Offset us Variation from Nominal Voltage Offset us 1 3 21 1 3 28 2 6 36 2 6 62 5 15 46 5 17 08 10 29 51 10 36 07 Table 6 Pulse generation circuit s frequency tracking offset with variations in mains supply power quality The power quality variations in table 6 are given as a percentage offset from the nominal 240V in the SWIS region The given offsets are valid for the respective power quality variation over 1 second 28 Filter and Transistor Amp
76. the physical connections needed to be created as shown in figure 36 RS232 Port Stepped Arduino Due Down Pulse Shield Mains RS232 Shield Supply SD Card Shield Figure 36 Physical frequency meter setup 56 6 1 1 Hardware Components A 3V SMA connect antenna was connected to the SMA male connector on the Trimble Copernicus Il to increase signal reception A 3 3V and GND rail were connected on the breadboard to provide power to the Copernicus Il module The connections for the Copernicus Il are shown in figure 32 with the active connections bolded LNA SMA Connector Reserved 7 Reserved 1 Reserved 6 OPEN TX B NMEA SHORT Trimble TX A TSIP Reserved 2 Copernicus Il Reserved 5 VBATT 3 3V 63530 00 GPS RX A TSIP XRST 3 3V RX B NMEA VCC 3 3V PPS GND 0V Reserved 4 XSTBY 3 3V Reserved 3 Figure 37 Trimble Copernicus Il connected pins diagram The RS232 shield has digital pin 16 and 17 connected to the Arduino as soon as the shield is connected allowing communication instantly The Arduino s 5V rail powers the MAX232N DIP module The mains supply can be connected in any manner to the green screw terminal on the pulse generation shield and is not polarity sensitive The shield stacking hierarchy is as follows Arduino Due bottom RS232 Shield Pulse Generation Shield SD Card Shield top The SD card shield connections are as o
77. the timing jitter directly against As atomic standards have very low jitter on the order of parts per billion they would be a suitable candidate Data collected on the Arduino for timing jitter analysis was not monitored closely with respect to temperature variations While it can be inferred that temperature had an effect on the frequency stability of the Arduino s crystal oscillator based on theory quantifying the scale of change in jitter across different temperature ranges would allow a temperature dependent model to be developed for the crystal therefore allowing the crystal to be used without an external standard providing a time base for the MCU as its jitter could be quantified at any time based on a temperature sample External effects such as pressure aging and other effects defined in section 2 1 3 could also be considered The GPS pulse per second signal was considered in the jitter analysis of this project but was not able to be quantified due to the length of time required to log the signal jitter the very low short term jitter order of nanoseconds and the shared phase relationship between the two GPS receivers By quantifying the actual jitter of both GPS receivers PPS outputs experimentally derived corrections can be made to the measurements made on the Arduino s jitter instead of relying upon data sheet specifications The Arduino Due is an excellent MCU in today s market offering an excellent sampling speed low cost exce
78. uating the performance of the final frequency meter A GPS time source was chosen to provide an accurate source of 1 second pulses An Arduino Due microcontroller used a KX 7 quartz crystal oscillator to maintain its time base and the accuracy of the KX 7 s time base was analysed against the Trimble Copernicus Il and GlobalSat EM406 A GPS receivers time base When analysed relative to the GPS receivers accurate time base the KX 7 maintained a low time base variation well within it s data sheet specifications The Arduino Due microcontroller was programmed and provided relevant frequency data to a LabVIEW PC terminal which allowed frequency visualisation data storage grid frequency contingency detection recovery time logging GPS initialisation data and cross platform communication protocols Frequency data was logged on the frequency meter and was able to provide a microHertz resolution The primary limitation of the design was low level noise on the mains supply line as this affected the designed electronics when logging frequency measurements below the milliHertz range Multiple recommendations for future work have been identified and included in this report III Thesis Contents LACKNOWIED i n gl RETIRO 1 EA MD Mn 2 MR AN 3 IWILIS OPERA sacada 5 AoE S A E E a tdi 6 MIListof EQUATIONS uu E 7 VILLT Of AbbreviatiofiS NA 8 A E ANREDE DDR A Tcr NM 9 1A
79. ultiplication and shift saving a few cycles Figure 28 New interrupt functioning implementation of micros 51 4 2 3 Frequency Metering The metering program on the Arduino is configured to allow either manual or automatic setup In addition to this the program only needs two minor modifications to run in jitter logging mode where the Arduino will log time based on an external interrupt trigger such as a GPS PPS signal After the Arduino finishes setup the output is periodic based on the PPS generated ISR Figure 29 displays the routine the Arduino undertakes for normal frequency metering operation The stages are segmented and interdependent with LabVIEW to progress to the next stage if a LabVIEW connection is detected 46 Declare and Initialise Variables Define Functions HWCDelay NMEA_Packet_Checker GetRisingEdge ClearSerial NMEA Response GetNMEA SendNMEA Serial Protocol Request Reset Copernicus Comms GetFrequency Wait for LabVIEW Comms Found Comms on Enable corresponding SerialO or Serial port Serial2 Initiate NMEA packet transfer from LabVIEW 10 Seconds Elapsed Packets Done Manual Configuration Wait for GPS Fix Attach Interrupt Functions on D8 D9 GetRisingEdge on D8 GetFrequency on D9 Interrupt Generated Print CSV frequency data to Serial port and SD Card Figure 29 Arduino frequency metering prog
80. uthor presents a cost effective solution to metering the grid frequency This paper presents a similar methodology wherein a PIC MCU is fed a digital pulse generated by a zero crossing detector circuit and counts the time between the pulses to determine the frequency It also discusses methodology to increase accuracy in obtaining the correct frequency and is highlights the drawbacks of particular solutions such as counting the number of pulses in a second window However this method is not clock disciplined but does give insight into the type of electronics that require development Electronic Navigation Systems 22 This book covers the different factors that affect radio signals at various frequencies and while providing an introductory chapter to radio signalling it also covers Satellite Navigation systems like GPS 16 Excellent explanations are given for the various effects that affect GPS such as atmospheric effects noise operating frequency and others This allows quantification of the effects that generate systematic error in clock synchronisation 70 Trimble Copernicus II GPS Receiver Reference Manual 33 This document contains all the data specifications for the Copernicus Il GPS module except the clock accuracy when a PPS fix is not obtained NMEA packet configuration is given for this particular unit 10 as a certain configuration must be running for both demonstration purposes and setup testing Additionally this manual provid
81. utlined in table 5 6 1 2 Program Parameters To run the frequency metering correctly several program files must be configured properly In the Arduino environment s MFFM Arduino program there exist two interrupt functions named GetRisingEdge One is used for logging relative clock jitter on the Arduino as outlined by the results in section 5 The alternative is used for frequency metering which is relevant to this section Figures 38 and 39 display the GetRisingEdge ISR function code The code in figure 38 is used for timing precision analysis while the code in figure 39 has been adapted for use in metering the mains frequency 57 void GetRisingEdge PrevMicros NewMicros NewMicros micros EdgeChanged true countedges Figure 38 PPS ISR for timing precision analysis void GetRisingEdge PPS Micros micros Final_Gap PPS_Micros PulseTime PPS Started true EdgeChanged true PulsesCounted PulseCount PulseCount 0 return Figure 39 PPS ISR for frequency metering To change between the two functions simply comment out the function that is redundant by wrapping the start with the characters and the end of the redundant function with the characters This approach will be also used in the main loop Two sets of if statements exist within the main body as shown in figure 40 if EdgeChanged true EdgeChanged false Serial print NewMicros Serial print S
82. utput by the Copernicus Il module The VI expects a GPGGA string message including both the S start character and the checksum at the end The LabVIEW string library finds the separation index of the commas located throughout the message and dissects the message based on these string index values into its various components such as UTC Time GPS Fix Status Latitude Longitude and more Output types are dissected message strings and a Boolean value that determines whether the GPS has obtained a satellite fix PadZeroes vi The PadZeroes VI takes a string input and replaces all spaces with a string value of O This VI is primarily used to support the NMEA Configuration VI file The output type is a string NMEA Configuration vi Receiver Configuration Acquisition Sensitivity Reset Configuration Generated Packets Elevation Mask Degrees Mode LL Reset Type Receiver Config Packet PPS Fix On LLL Hot Software d PINISCRAS DAL SCS Dynamics Pulse Length Multiple of 100ns USO PPS Packet Station ary 5000000 Serial Communication Store User Config to Flash SPTNLSPS 2 5000000 1 0000010 51 WAAS EEr Baud Rate Latitude Acquisition Sensitivity Packet eie 19200 32 0396635 SPTNLSFS 5 0 23 High Woke Up from standby Ore Serial Communications Packet Li ongi potBaciiy imunk ions Pac Antenna Length UTC at BAGKY SPTNLSPT 019200 8 N 1 4 4 1C Compensation ns Elapsed Time Value secs Initial Position Packet 10 0 SPTNLSKG
83. y being made 2 1 1 Atomic Clocks Atomic clocks are the highest standard of clock precision available today Atomic clock standards are expensive often costing tens of thousands of dollars or more thereby making them a difficult standard to use outside of expensive projects and experiments Time synchronisation on computers and electronics is often done by polling time from an accurate source To synchronise to this accurate source several implementations exist such as e Radio clock broadcasting stations e Stratum 1 NTP servers e GPS Satellites that broadcast a PPS signal Radio clocks 16 have a local atomic clock reference that generates time data for radio broadcasting In Network Time Protocol NTP implementations 14 an atomic clock is considered a stratum 0 device Stratum O devices provide a very accurate timing signal and are used as reference clocks Stratum 1 servers are synchronised within microseconds to their respective stratum O device and may broadcast NTP time packets GPS satellites each have an atomic clock on board the space vehicle The instrumentation on board the space vehicle allows a very accurate PPS signal to be generated and broadcast to GPS receivers through radio frequencies Atomic clocks function by locking an electronic oscillator to the frequency of an atomic transition 15 Two well known and often used standards are caesium 133 which transitions at 9 192 631 770 Hz 15 and rubidium 87 which tr
84. y log ambient temperature and attempt to correlate the two sets of data Overall a cyclical nature in the Arduino crystal jitter is observed when ambient temperature is not affected by household climate control systems such as air conditioning 54 48 Hour Cumulative Time Interval Error Cumulative TIE Linear Cumulative TIE P Ww m A T E E v 2 E E 3 o N un RS ASR Elapsed Time Seconds Figure 35 Arduino 48 hour TIE graph EM406 A PPS source A final TIE for the EM406 A based set of data is generated and can be seen in figure 35 The TIE is accumulated to 672 64ms over the 48 hour period The data set again gives a linear trend with a very high R value indicating the data in the trend line fit correlates highly 336 32ms are lost by the Arduino on average in this set of data per day 15 04 less than the time lost in the Copernicus II data set This corresponds to and average of 3 89ppm well within the 30ppm specification given by the KX 7 crystal s manual 24 55 6 Frequency Meter With a quantified Arduino timing bias frequency metering was able to be performed Similar to section 5 a 48 hour set of data was obtained on the mains supply s frequency to determine if the data was cyclical and whether the frequency varied as expected 6 1 Setup Prior to logging the frequency data setup needed to be performed in the LabVIEW program settings the Arduino program MFFM Arduino ino and
85. y mode Sensitivity Serial SPTNLSPT 019200 8 N 1 4 4 1C 19200 Baud 8 data bits No parity Communications check 1 Stop bit NMEA in and NMEA out Initial Position SPTNLSKG GPSW GPSWMS 3203 96635 S 11550 22761 E 00010 FF GPSW GPS Week since first epoch GPSWMS Milliseconds accumulated since 00 00 UTC Sunday Reset Configuration SPTNLSRT H 2 2 0000000000 1B Hot Start Store User Config to Flash on reset Wake on NMEA port activity Table 8 Copernicus Il project default NMEA packet configuration with checksums Implementation appropriate carriage return and line feed delimiters should follow all packet checksums 33 The automatic message output was configured to display GGA messages GGA messages display GPS Fix data which allows PPS integrity monitoring based on the number of active satellites 31 Elevation gt S Mask Angle a 90 b 60 c 15 Horizon Figure 14 Various elevation mask angles of GPS Satellites referenced to a North Pole positioned receiver The receiver s configuration had the elevation mask set at 15 The elevation mask is the minimum elevation angle between the horizon and the satellite relative to the receiver as shown in Figure 14 At 10 elevations and higher ionospheric and tropospheric signal corruption is reduced as the atmospheric effects begin to become more predictable for the receiver The possible limitation of this approach is exclusion of

Mains Frequency Fluctuation Metering

Contents

Download Pdf Manuals

Related Search

Related Contents