The purpose of RMOs

2010 SIM TFWG Workshop and Planning Meeting, Lima, Peru

March 9-12, 2010

The SIM Time Network and its role in the time and frequency laboratory

Michael A. LombardiNational Institute of Standards and Technology (NIST), USA

[email protected]


March 9-12, 2010

SIM is the Interamerican Metrology System, one of the world’s five major Regional Metrology

Organizations (RMOs) recognized by the BIPM


March 9-12, 2010

The purpose of RMOs The International Bureau of Weights and

Measures (BIPM) works to ensure the worldwide uniformity of measurements and their traceability to the International System of Units (SI). This allows the measurements made in one country to be accepted and trusted in other countries, which is important for international trade.

The BIPM expects RMOs to review the quality systems of NMIs, and their calibration and measurement capabilities (CMCs). RMOs should also:

Organize regional comparisons to supplement the BIPM key comparisons so that more nations can establish traceability to the SI. This is the role of the SIMTN.


March 9-12, 2010

Information about SIM

SIM consists of NMIs located in the 34 member nations of the Organization of American States (OAS), which extends throughout North, Central, and South America, and the Caribbean region.

OAS accounts for roughly 13% of the world’s population (about 910 million people as of 2009), and roughly 27% of its land mass. SIM is the largest RMO in terms of land area.

About 2 out of 3 people in the SIM region live in the United States, Brazil, or Mexico (roughly 617 million people).

Eleven SIM nations (mostly islands) have less than 1 million people.

SIM has organized metrology working groups (MWGs) in 11 different areas, including time and frequency. The SIM Time Network is operated by the T&F MWG.


March 9-12, 2010

SIM Time Network Design Goals Our design goals were:

To establish cooperation and communication between the SIM time and frequency labs now and in the future.

To build a network that allowed all SIM NMIs to compare their time standards to those of the rest of the world.

To utilize equipment that was low cost and easy to install, operate, and use, because SIM NMIs typically have small staffs and limited resources.

To be capable of measuring the best standards in the SIM region. This meant that the measurement uncertainties had to be as small, or nearly as small, as those of the BIPM key comparisons.

To report measurement results in near real-time, without the processing delays of the BIPM key comparisons.

To build a democratic network that favored no single laboratory or nation, and to allow all members to view the results of all comparisons.


March 9-12, 2010

TIME AND FREQUENCY METROLOGY WORKING GROUPWorking to support time and frequency metrology throughout the Americas

United States, 2005

Mexico, 2005

Canada, 2005

Panama, 2005

Brazil, 2006

Costa Rica, 2007

Colombia, 2007

Argentina, 2007

Guatemala, 2007

Jamaica, 2007

Uruguay, 2008

Paraguay, 2008

Peru, 2009

Trididad & Tobago, 2009

Chile, 2010

Saint Lucia, 2010


March 9-12, 2010


March 9-12, 2010

Country Date equipment was shipped

BIPM MRA Signatory?

T&F Standard Contributes to UTC?

United States 2005 Yes Time Scale (six masers and four cesiums) and Primary Standard

Yes

Mexico 04/2005 Yes Time Scale (one maser and four cesiums)

Yes

Canada 05/2005 Yes Time Scale(three masers and four cesiums)

Yes

Panama 10/2005 Yes Two cesiums Yes

Brazil 09/2006 Yes Time Scale(six cesiums)

Yes

Costa Rica 01/2007 Yes Cesium No

Colombia 02/2007 No Two cesiums No

Argentina 07/2007 Yes Cesium Yes

Guatemala 08/2007 No GPSDO No

Jamaica 12/2007 Yes Two cesiums No

Uruguay 11/2008 Yes Disciplined rubidium No

Paraguay 11/2008 Yes Rubidium No

Peru 06/2009 Yes Rubidium No

Trinidad / Tobago 08/2009 Yes GPSDO No

Chile 2010 Yes Rubidium? NMI does not, but geodetic observatory does

Saint Lucia 2010 Yes Rubidium? No


March 9-12, 2010

The SIM Measurement System Simple design makes it easy and inexpensive for

SIM labs to compare their standards. It includes: 8-channel GPS receiver (C/A code, L1 band) Time interval counter with 30 ps resolution Rack-mount PC and flat panel display Pinwheel type antenna Applies broadcast ionospheric (MDIO)

corrections

Data are not stored in CGGTTS format. The receiver measures all visible satellites and stores 1-minute and 10-minute REFGPS averages.

All systems are connected to the Internet, and send their files to a web server every 10 minutes.

The web server processes data “on the fly” in near real-time. Results can be viewed on the web in either common-view or all-in-view format.

All units are built and calibrated at NIST

Systems are paid for by either OAS or the participating NMI and become the property of the NMI.


March 9-12, 2010

tf.nist.gov/sim


March 9-12, 2010


March 9-12, 2010

Reporting results to participating SIM laboratories

Measurement results can be viewed using any Java-enabled web browser. Our web-based software does the following:

Plots the one-way GPS data (average of all satellites and tracks for each individual satellite) as recorded at each site relative to the local standard.

Plots the time and frequency difference between NMIs using the common-view method (common-view data are averaged across all satellites and are also shown for each individual satellite).

Calculates the Allan deviation and time deviation.

Makes 10 minute, 1 hour, and 1 day averages available in tabular form.

Up to 200 days of data can be retrieved at once. All old data remains available, nothing is ever deleted.

The time difference between any two laboratories can be viewed by all laboratories in the network. New results are available every 10 minutes.

Results can be processed as “classic” common-view or all-in-view.


March 9-12, 2010


March 9-12, 2010

Maximum Time Difference last 6 months of 2009

NIST CNM NRC CNMP ONRJ ICE SIC INTI BSJ

NIST 57 -95 38 25 -1066 -56 52 51

CENAM -57 -111 -81 -38 -1081 -74 68 -80

NRC 95 111 122 93 -997 118 133 138

CENAMEP -38 81 -122 -59 -1088 -88 57 56

ONRJ -25 38 -93 59 -1084 -46 61 68

ICE 1066 1081 997 1088 1084 1032 1072 1098

SIC 56 74 -118 88 46 -1032 79 100

INTI -52 -68 -133 -57 -61 -1072 -79 -71

BSJ -51 80 -138 -56 -68 -1098 -100 71


March 9-12, 2010

Average Time Difference last 6 months of 2009


NIST 10 -73 13 < 1 -480 -8 15 11

CENAM -10 -82 4 -9 -489 -18 3 < 1

NRC 73 82 86 71 -407 65 86 84

CENAMEP -13 -4 -86 -17 -492 -21 -7 -2

ONRJ < 1 9 -71 17 -476 -6 11 13

ICE 480 489 407 492 476 456 487 464

SIC 8 18 -65 21 6 -456 16 18

INTI -15 -3 -86 7 -11 -487 -16 4

BSJ -11 < 1 -84 2 -13 -464 -18 -4


March 9-12, 2010

Average Frequency Difference (× 1015) last 6 months of 2009


NIST -4 -2 3 -2 15 -1 < 1 1

CENAM 4 2 6 2 19 2 4 5

NRC 2 -2 5 < 1 17 < 1 2 4

CENAMEP -3 -6 -5 -5 13 -4 -3 -1

ONRJ 2 -2 < 1 5 17 < 1 2 4

ICE -15 -19 -17 -13 -17 -21 -15 -28

SIC 1 -2 < 1 4 < 1 21 2 3

INTI < 1 -4 -2 3 -2 15 -2 2

BSJ -1 -5 -4 1 -4 28 -3 -2


March 9-12, 2010

Benefits to the SIM Region

Improved time coordination.

Better time standards are being maintained at many of the SIM labs.

Increased awareness of the importance of time and frequency. Some SIM labs are introducing new calibration services and improving existing services to

better support local industry. New time services are also being introduced (NTP servers, web clocks, etc.).

Improved status for NMIs. Companies in SIM countries are likely to use their local NMI as a source of traceable

frequency measurements.

A more visible official timekeeper. Some SIM labs have become or are trying to become the official timekeepers in their

countries.


March 9-12, 2010


March 9-12, 2010

Selecting a Time and Frequency Standard:

Rubidium, Cesium, GPSDO, or ensemble time scale


March 9-12, 2010

Cesium Oscillators Used to define the second, one of the seven base units of the SI A true intrinsic standard Costly, probably $35 000 to $75 000 USD

Rubidium Oscillators Low cost (often under $5 000 USD) but need to be adjusted often to

compensate for aging and frequency drift

GPS Disciplined Oscillators A quartz or rubidium oscillator continuously steered to agree with

signals from the GPS satellites. Cannot be adjusted, but doesn’t need it.

Price ranges from about $1 000 to $15 000 USD Are they an acceptable choice as a primary frequency standard?

Choices in Frequency Standards


March 9-12, 2010

Rubidium Oscillators

The least expensive atomic oscillator

Typically costs about 1/10 as much as a cesium, but its unadjusted accuracy is typically about 1000 times worse $3,500 rubidium might be accurate to a few parts in 1010 after warm up $35,000 cesium is likely to be accurate to within a few parts in 1013

They need to be adjusted periodically Frequency change (due to aging) can exceed 1 x 10-11 per month, or

1 x 10-10 per year

Not acceptable for BIPM key comparisons


March 9-12, 2010


March 9-12, 2010

Cesium Oscillators

A true primary standard for a cal lab, since the SI second is defined based on cesium

Maintenance cost is high, since the beam tube (which is often more than half the cost of the oscillator) typically needs replacement after about 10 years

Can operate for many weeks, months, or years without requiring adjustment, maintaining average frequency of less than 1 × 10-13 if properly maintained.

Still needs to be checked against a secondary standard to make sure that it is working - a failed cesium becomes an OCXO.


March 9-12, 2010


March 9-12, 2010

GPS Disciplined Oscillators (GPSDO)

A self-calibrating standard

Care must be taken to ensure that GPS signals are being received and that the GPSDO is working properly

Performance varies, but even the worst units will be accurate to better than 1 × 10-12 over a one day interval

Great performance for the money, but not adjustable and not accepted in BIPM key comparisons

A good thing to have, however, as a backup or secondary standard


March 9-12, 2010


March 9-12, 2010

Ensemble Time Scales

Require multiple cesium oscillators.

Very expensive, but a time scale will be more stable than any of its individual clocks. It will also keep going if one of the clocks fails.

Requires measurement hardware, phase steppers, synthesizers, etc.

CENAM, NIST, NRC, and ONRJ now have ensemble time scales and can help other SIM labs that want to start one.


March 9-12, 2010

The NIST Time Scale consists of an ensemble of commercial clocks, currently six hydrogen masers and four cesium beam standards.

The weighted average of these clocks generates continuous signals from a high resolution frequency synthesizer that is locked to a hydrogen maser. Both 5 MHz (frequency) and 1 pps (time) signals are generated.

The clock ensemble is periodically calibrated using the NIST-F1 primary standard.

Of course, a time scale can be built with fewer clocks and without hydrogen masers.. However, at least three cesiums are required. The time scales at CENAM and ONRJ have built or redesigned in recent years and provide excellent performance.

UTC(NIST) Time Scale


March 9-12, 2010

Selecting Measurement Equipment


March 9-12, 2010


March 9-12, 2010

Essential Equipment for a Frequency Measurement Laboratory

Essential Equipment

Primary Frequency Standard Secondary Frequency Standard (if you don’t have a multi-clock time scale) Distribution Amplifier Oscilloscope Universal Counter Signal Generator


March 9-12, 2010

Distribution Amplifier

The frequency signals (typically 10 MHz sine waves) from the lab’s primary standard can be distributed throughout the work area using a distribution amplifier.

Signals from the distribution amplifier should be used as the external time base for all instrumentation in the laboratory (counters and signal generators, for example). This will ensure that the lab’s measuring instruments have the same frequency accuracy as the primary standard.


March 9-12, 2010

Oscilloscope

Allows the viewing of waveforms and pulses, an indispensable device for the time and frequency lab.

Can be used to measure frequency, but that should only be done if a counter is not available.


March 9-12, 2010

Key Oscilloscope Specifications Bandwidth

20 MHz (low end) to 70 GHz (high end) A 100 MHz unit is adequate for a timing lab

Time base range Best scopes can scale from about 50 ps to 50 s per division (10 divisions) A unit that scales from 5 ns to 50 s per division is adequate for a timing lab

Frequency Counter Resolution Typically about 1 x 10-4 (1 kHz resolution at 10 MHz), not useful for any serious

measurement, but provides quick check of frequency Time Interval Resolution

Proportional to length of the interval Typically about 1 x 10-4 (1 µs for 10 ms interval)

Number of channels Ranges from 2 to 8 plus external trigger 2 channel handles pattern drift method

t function is very useful for time interval measurements


March 9-12, 2010

Universal Counter

Counts frequency, time interval, period, and events (totalize mode). Some universal counters include other functions like phase, peak voltage, rise/fall time, etc.

Probably the most important instrument in a time and frequency lab, useful in nearly all areas of T&F metrology.


March 9-12, 2010

Universal Counter Specs (Frequency)

Frequency Range 100 MHz (low end) to 50 GHz (high end) For most purposes, a unit that can count up to 100 MHz is adequate.

Resolution (number of digits) 7 digits ($200 counters) to 13 digits 8-digit counter has 1 Hz resolution at 10 MHz:

10 000 000 Hz This allows it to detect frequency changes as small as 1 x 10-7

12-digit counter has 100 µs resolution at 10 MHz: 10 000 000 000 0 Hz This allows it to detect frequency changes as small as 1 x 10-11

12-digit counters are relatively cheap, and recommended

Any selected counter must be able to accept an external time base.


March 9-12, 2010

Universal Counter Specs (Time Interval)

Time Interval Resolution 100 ns (low end, period of 10 MHz time base) to 20 ps (high end) Relatively low cost counters have 150 ps resolution, which is more

than adequate. Smallest frequency offset which can be resolved at 1 second is 150 ps / 1 second, or 1.5 x 10-10.

Time Interval Accuracy and Range Typically, no better than 1 ns for even the best counters, but it is

not proportional to the length of the interval like it is for oscilloscopes.

A good counter can can measure intervals from 1 ns to about 100 000 seconds. The best oscilloscopes can measure shorter intervals, but it is unlikely that cal labs will need to do this.


March 9-12, 2010

Signal Generator

Generates signals at user selectable frequencies and amplitudes Some units (called function generators) generate a variety

of different waveforms, other units generate just sinewaves and/or squarewaves


March 9-12, 2010

Signal Generator Uses

Extremely useful tool for generating signals to test measurement systems. Also useful for repairing equipment.

Can generate frequencies with small offsets from nominal

Can generate a needed frequency locked to house reference, when no standalone oscillator exists

Can generate test signals needed to calibrate stopwatch calibrator or other systems


March 9-12, 2010

Key Signal Generator Specifications

Frequency Range 1 MHz or less (low end) to 100 GHz or more (high end) A device that goes a little above 10 MHz is adequate.

Amplitude Range 0.01 to 10 V peak-to-peak (50 ohms) is typical and will handle most lab

functions Resolution

Resolution tends to be lower on devices with the largest frequency range, for example a device that goes to 1 GHz probably will only have 1 or 10 Hz resolution.

1 µHz resolution is about as good as it gets, and is desirable for cal lab purposes. It allows generating a 10 MHz signal with a 1 x 10-13 frequency offset. Some units extend 1 µHz resolution out to about 80 MHz. With a unit like that you can generate a signal with a 1.25 x 10-14 frequency offset.

External time base is mandatory


March 9-12, 2010

Optional Equipment for a Frequency Measurement Laboratory

(recommended for most labs)

Optional Equipment

An automated phase comparison system Homebrew (PC and time interval counter, for example) Commercially available phase comparator (heterodyne system, for example) NIST Frequency Measurement System

Frequency Dividers and Multipliers A stopwatch calibrator Data Analysis Software (homebrew, Stable32, Excel, etc.)


March 9-12, 2010

Basic Model of a Phase Comparator for Measuring Frequency

f0 f f0 Comparator

Oscilloscope, frequency counter, time interval counter,

Dual Mixer Time Difference System

Device Under Test

Reference Frequency

Data Acquisition System

PC


March 9-12, 2010

Homebrew System

Several types of systems are possible, but the most practical would probably consist of a time interval counter, a PC, and software

Software must be written to collect a series of measurements from the counter (Basic, C, Labview, etc.)

Data can be analyzed using your own software, Stable 32, or Excel. You can download Francisco’s software for free from the SIM web site.

Frequency dividers must be built or purchased to divide the standard reference and DUT signals to a common low frequency, usually 1 Hz


March 9-12, 2010

Time Interval Method


March 9-12, 2010

Dual-Mixer Time Difference System (DMTD)


March 9-12, 2010

Commercial DMTD System Measures frequencies from 1

MHz to 20 MHz

Computes Allan deviations from 0.01 s to 106 s

Produces phase and frequency plots

Resolution is about 100 femtoseconds (0.1 ps), about 200 times better than the best time interval counters.


March 9-12, 2010

Stopwatch Calibrator

A handy instrument to have if you laboratory calibrates stop watches and timers

Several models are available, costing between about $2,000 and $3,500

They reduce the time required to perform calibrations


March 9-12, 2010


March 9-12, 2010

Software The laboratory should have software available to perform basic time and

frequency functions, such as:

Estimate average frequency accuracy from phase data Compute Allan deviation Produce graphs for calibration reports Perform uncertainty analysis

Software can written in-house with any programming language and many functions are easy enough to automate with Excel. One commercially available software package is Stable32 (www.stable32.com). The free software written by Francisco Jimenez can be downloaded here:

http://tf.nist.gov/sim/papers.htm


March 9-12, 2010

Joining the BIPM key comparisons


March 9-12, 2010

Simple format collects more data without the need for a tracking schedule.

The SIM Time Network collects more data than the CGGTTS* format used by the BIPM, but is not

compatible

* Consultative GPS and GLONASS Time Transfer Sub-committee

Method Tracks per day Track Length Satellites Minutes of data per day

CGGTTS *

single-channel

48 13 1 624

CGGTTS *

multi-channel

90 13 8 typical 9360

SIM 144 10 8 maximum 11520


March 9-12, 2010

The CGGTTS Common-view Data Format

GPS RCVR: NBS10V9809MJD= 51658 YR=00 MONTH=04 DAY=24 HMS=14:47:20 (UT)GGTTS GPS DATA FORMAT VERSION = 01REV DATE = 2000-04-03RCVR = NBS10....................CH = 01IMS = 99999LAB = NISTX = -1288398.27 mY = -4721698.10 mZ = +4078625.68 mFRAME = ITRF....COMMENTS = NO COMMENTS..............INT DLY = 53.0 nsCAB DLY = 0199.9 nsREF DLY = 0066.7 nsREF = UTCNISTCKSUM = 74

PRN CL MJD STTIME TRKL ELV AZTH REFSV SRSV REFGPS SRGPS DSG IOE MDTR SMDT MDIO SMDI CK hhmmss s .1dg .1dg .1ns .1ps/s .1ns .1ps/s .1ns .1ns.1ps/s.1ns.1ps/s 3 08 51655 105800 780 380 760 -1058301 -1131 -571 -1098 415 163 107 +2 76 +0 02 8 32 51655 111400 780 319 2933 -7071115 -3061 -246 -3082 290 074 125 -20 85 -9 34 13 28 51655 113000 780 415 3083 +6965884 -30 -94 -241 625 019 100 -12 71 -7 FB 3 74 51655 114600 780 296 530 -1058331 +929 -503 +962 470 163 133 +19 92 +24 17 31 08 51655 121800 780 498 706 -7572 -400 -197 -390 470 180 87 +4 99 +14 DD 13 32 51655 123400 780 569 2693 +6966345 +171 -440 -40 424 011 79 +0 90 +9 F0 18 68 51655 125000 780 279 1829 -341335 +18 -132 +22 698 182 141 +35 152 +44 16 31 74 51655 132200 780 283 472 -7436 +2669 -73 +2678 441 206 139 +29 190 +36 24


March 9-12, 2010

You need to have a cesium oscillator

You need to have a CGGTTS compatible GPS receiver

Your country must be a signatory of the CIPM MRA

You must contact the BIPM and provide information on the name/address of the laboratory, clocks (model, serial number), time transfer equipment in the laboratory, and any other relevant information. They will then assign an acronym and a code to your laboratory, and a code to each clock.

You must submit a data file once per month (by the 5 th day of the month) by FTP

Steps required in order to appear on the BIPM Circular-T and contribute to UTC


March 9-12, 2010

BIPM-Compatible Common-view Receivers

There are a few receivers that you can buy. The cost is usually between $15,000 and $35,000 USD.

AOS TTS-2 (single frequency)

AOS TTS-3 and TTS-4 (dual frequency)

Dicom GTR50 (dual frequency)

Novatel (dual frequency)

PolaRx2eTR (dual frequency)


March 9-12, 2010

SIM system with conversion software (used by INTI in Argentina but not 100% compatible)

Make the SIM system simultaneously produce both the SIMTN and the CGGTTS formats

This is difficult because some information that is necessary for CGGTTS compatibility cannot easily be extracted from the SIM receivers.

Design a new CGGTTS receiver for SIM. This has been discussed at NIST and CENAM.

The currently available receivers cost between $15,000 and $35,000 USD. A single frequency receiver could probably be designed and sold for $10,000 USD (perhaps

less). A dual frequency receiver will cost more because of the high price of the parts. Please let us know if there is any interest.

Three possibilities for the future

The purpose of RMOs

Documents

Transcript of The purpose of RMOs