Katalog Fluke Temperature Calibrator - PT Siwali Swantika (021-45850618 || 031-/8421264)

Temperature Calibration Equipment and Services

2

See page 63 for moreinformation.


See page 171 formore information.


3

Primary standards selection guide . . . . . . . . 4Why buy primary standards from Hart? . . . . . . . . . . . . . . . 6

Quartz-sheath SPRTs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Working standard SPRT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10Extended range metal-sheath SPRT . . . . . . . . . . . . . . . . . . . 11Glass capsule SPRTs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12Annealing furnace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13Triple point of water cells . . . . . . . . . . . . . . . . . . . . . . . . . . . 14TPW maintenance bath . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Use TPW and ratio method to improve SPRTstability and accuracy. . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

DC bridge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20Why use Fluke metal freeze-point cells? . . . . . . . . . . . . . 21

ITS-90 fixed-point cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23ITS-90 fixed-point cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Traceability and thermometric fixed-point cells . . . . . . . . 26Freeze-point furnaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28Mini fixed-point cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30Mini TPW maintenance apparatus . . . . . . . . . . . . . . . . . . . . . 32Gallium cell maintenance apparatus . . . . . . . . . . . . . . . . . . . 33Mini fixed-point cell furnace . . . . . . . . . . . . . . . . . . . . . . . . . 34LN2 comparison calibrators. . . . . . . . . . . . . . . . . . . . . . . . . . . 35DC resistance standards. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36Standard AC/DC resistors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Thermometer readout selection guide . . . . 38Super-Thermometer readouts . . . . . . . . . . . . . . . . . . . . . . . . 40

Evaluating calibration system accuracy . . . . . . . . . . . . . . 44International Temperature Scale of 1990 (ITS-90)quick reference. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45PRT temperature vs. resistance table . . . . . . . . . . . . . . . . 46Thermocouple EMF and sensitivity chart . . . . . . . . . . . . . 47

The Black Stack thermometer readout . . . . . . . . . . . . . . . . . . 48Tweener thermometer readouts. . . . . . . . . . . . . . . . . . . . . . . 56Handheld thermometer readouts . . . . . . . . . . . . . . . . . . . . . . 58Reference multimeter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62The DewK thermo-hygrometer . . . . . . . . . . . . . . . . . . . . . . . 63

Thermometer probe selection guide . . . . . 66How accurate is that probe?. . . . . . . . . . . . . . . . . . . . . . . 67

Secondary standard PRTs . . . . . . . . . . . . . . . . . . . . . . . . . . . 68Ultra high-temp PRT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69Secondary reference temperature standards . . . . . . . . . . . . . 70Secondary PRT with calibration options . . . . . . . . . . . . . . . . 725616 secondary reference PRT . . . . . . . . . . . . . . . . . . . . . . . 74

Sensible temperature specifications . . . . . . . . . . . . . . . . . 76Small diameter industrial PRT . . . . . . . . . . . . . . . . . . . . . . . . 77Precision freezer PRT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78Precision industrial PRTs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79Fast response PRTs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

What are stem conduction errors and how can theycreate errors in calibration? . . . . . . . . . . . . . . . . . . . . . . . 81

Thermistor standards probes . . . . . . . . . . . . . . . . . . . . . . . . . 82Secondary referencethermistor probes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

Thermistors: the under appreciated temperaturestandards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86Thermocouples 101... or, maybe... 401! . . . . . . . . . . . . . . 88

Type R and S thermocouple standards . . . . . . . . . . . . . . . . . 90Reference table: letter-designated thermocoupletolerances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91Temperature conversion table . . . . . . . . . . . . . . . . . . . . . 92Why did my temperature sensor fail calibration?. . . . . . . 93

Software selection guide . . . . . . . . . . . . . . 96Interface-it . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96MET/TEMP II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97TableWare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100LogWare and LogWare II . . . . . . . . . . . . . . . . . . . . . . . . . . 101LogWare III . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

Establishing traceability . . . . . . . . . . . . . . . . . . . . . . . . . 103

Bath selection guide . . . . . . . . . . . . . . . . . . 104Buying the right bath . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

Deep-well compact baths . . . . . . . . . . . . . . . . . . . . . . . . . . 108Compact baths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

Why a Hart bath? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112Really cold baths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114Cold baths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116Hot baths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118Really hot bath . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120Bath accessories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121Deep-well baths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122Resistor baths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124Constant temperature ice bath. . . . . . . . . . . . . . . . . . . . . . . 126

Avoid water problems in cold baths . . . . . . . . . . . . . . . . 127Bath fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128Controller for Rosemount-designed baths . . . . . . . . . . . . . . 132Benchtop controllers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

Industrial calibrator selection guide . . . . . 134Selecting an industrial temperature calibrator . . . . . . . . 136

Metrology Well calibrators . . . . . . . . . . . . . . . . . . . . . . . . . . 139Field Metrology Wells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

Understanding the uncertainties associatedwith Metrology Wells . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

Micro-Baths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152Eliminating sensor errors in loop calibrations . . . . . . . . 154

Field dry-wells. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156A few dry-well dos and donts.... . . . . . . . . . . . . . . . . . . 158

Industrial dual-block calibrator . . . . . . . . . . . . . . . . . . . . . . 160High-accuracy dual-well calibrator . . . . . . . . . . . . . . . . . . . 162Handheld dry-wells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164Portable lab dry-well . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166Thermocouple furnace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167Thermocouple calibration furnace . . . . . . . . . . . . . . . . . . . . 168Zero-point dry-well . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1704180 series precision infrared calibrators . . . . . . . . . . . . . . 171Portable IR calibrators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174Surface calibrator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

Other neat stuff selection guide . . . . . . . . 177External battery pack. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177Benchtoptemperature/humiditygenerator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

On rutabagas and their origins. . . . . . . . . . . . . . . . . . 180Hart Scientific temperature seminars . . . . . . . . . . . . . . . . . . 181

NVLAP accreditation at Hart . . . . . . . . . . . . . . . . . . . . . . 184Calibration services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

Guidelines for Hart product specifications . . . . . . . . . . . 194Hart Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195How to order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

Table of Contents

4 Primary Standards

SPRTsModel RTPW Range Page5681 25.5 200 C to 670 C 85683 25.5 200 C to 480 C5684 0.25 0 C to 1070 C5685 2.5 0 C to 1070 C5698 25.5 200 C to 670 C 105699 25.5 200 C to 670 C 115686 25.5 260 C to 232 C 125695 25.5 200 C to 500 C

Fixed-point cellsModel Description Temperature Page5901A-G TPW Cell, 12 mm ID with handle, glass shell 0.01 C 145901A-Q TPW Cell, 12 mm ID with handle, quartz shell 0.01 C5901C-G TPW Cell, 13.6 mm ID with handle, glass shell 0.01 C5901C-Q TPW Cell, 14.4 mm ID with handle, quartz shell 0.01 C5901D-G TPW Cell, 12 mm ID, glass shell 0.01 C5901D-Q TPW Cell, 12 mm ID, quartz shell 0.01 C5901B-G TPW Cell, mini, glass shell 0.01 C5900 TP Mercury, SST 38.8344 C 235904 Freezing Point of Indium 156.5985 C5905 Freezing Point of Tin 231.928 C5906 Freezing Point of Zinc 419.527 C5907 Freezing Point of Aluminum 660.323 C5908 Freezing Point of Silver 961.78 C5909 Freezing Point of Copper 1084.62 C5924 Open Freezing Point of Indium 156.5985 C5925 Open Freezing Point of Tin 231.928 C5926 Open Freezing Point of Zinc 419.527 C5927A-L Open Freezing Point of Aluminum, Long 660.323 C5927A-S Open Freezing Point of Aluminum, Short 660.323 C5928 Open Freezing Point of Silver 961.78 C5929 Open Freezing Point of Copper 1084.62 C5943 Melting Point of Gallium, SST 29.7646 C5914A Mini Freezing Point of Indium 156.5985 C 305915A Mini Freezing Point of Tin 231.928 C5916A Mini Freezing Point of Zinc 419.527 C5917A Mini Freezing Point of Aluminum 660.323 C5918A Mini Freezing Point of Silver 961.78 C5919A Mini Freezing Point of Copper 1084.62 C5944 Mini Freezing Point of Indium 156.5985 C5945 Mini Freezing Point of Tin 231.928 C5946 Mini Freezing Point of Zinc 419.527 C5947 Mini Freezing Point of Aluminum 660.323 C

Primary standards selectionguide

Temperature Calibration Equipment 5

ApparatusModel Features/Use Page7012 Maintains: triple point of water and gallium cells. Comparisons: 10 C to 110 C. 1167037 Maintains: triple point of water and gallium cells. Comparisons: 40 C to 110 C.7312 Maintains: two TPW cells. Compact size, runs quietly. Comparisons: 5 C to 110 C. 177341 Maintains: triple point of mercury cell. Comparisons: 45 C to 150 C. 1089210 Maintains: mini triple point of water and mini gallium cells. Comparisons: 10 C to 125 C. 329230 Maintains: stainless steel gallium cell. Comparisons: 15 C to 35 C. 339260 Maintains: indium, tin, zinc, and aluminum cells. Comparisons: 50 C to 680 C. 349114 Maintains: indium, tin, zinc, and aluminum cells. Comparisons: 100 C to 680 C. 289115A Maintains: aluminum and silver cells. Comparisons: 550 C to 1000 C.9116A Maintains: aluminum, silver, gold, and copper cells. Comparisons: 400 C to 1100 C.9117 Anneals SPRTs, HTPRTs, and thermocouples to 1100 C. Protects them against contamination from metal ions. 13

Boiling point of liquid nitrogen7196 Affordable substitute for a triple point of argon system. Provides for low-temperature comparison calibrations at

approximately 196 C with uncertainties of 2 mK.35

Resistance bridge5581 0.1 ppm accuracy for calibration of standard resistors and SPRTs. 13:1 measurement ratio allows resolution

to 0.001 mK.18

1590 1 ppm accuracy for calibration of SPRTs and thermistors. 40

Standard resistors742A Excellent performance without oil or air baths. Values from 10 ohm to 100 megohm. 365430 Highest stability oil-filled resistors (< 2 ppm/year drift). AC cal uncertainty to 3 ppm. 37

Primary standards selectionguide

Setting up a primary temperature stan-dards lab is no small project. Decisionsmust be made about temperature range,uncertainty requirements, the types ofstandards you need, and the companiesthat can supply your standards. Whoseproducts are reliable? Which companybacks up its performance claims? Whoprovides after-sale support and training?Who really demonstrates the most integ-rity throughout your ownership experi-ence? After all, substantial investmentsare being made, and in many cases thecredibility of your lab can be affected bythe outcome.

So why does Hart Scientific claim to bethe worlds best supplier of primary tem-perature standards? Because our productshave been tested over and over again bynational labs around the world andproven to outperform their specs. Becausethe people who design and build primarytemperature standards at Hart have beendesigning and building primary tempera-ture standards longer than any othersupplier in the world. We not only manu-facture primary standards, we perform ba-sic research and innovate with newprimary standards designs. No one else of-fers the high-quality training and post-sale support that we do. No one!

Metal fixed-point cellsFor realizing the ITS-90 temperature scale,Harts metal fixed-point cells provide per-formance you can trust, and we supply thedata with each cell to prove it. Hartsfixed-point cells benefit from more than20 years of experience in research, de-sign, and manufacturing. Three types ofcells are available: traditional size cells,mini quartz cells, and new mini metal-cased cells. All three provide outstandingperformance.

Each Hart cell is carefully assembled,tested, and supplied with an assay ofmetal-sample purity. Every traditional-sizecell further undergoes more rigorous test-ing to a CCT-based procedure in ourNVLAP-accredited lab, where we realize atleast three freezing curves and perform adetailed slope analysis to confirm cellpurity. If youd like this more thoroughslope analysis for a mini quartz cell ornew mini metal-cased cell, we offer thatas an option. And if you still want more,we can also supply comparison data withour own reference cells that have been in-dependently tested at NIST.

No other commercial company has asmuch experience in the development offixed-point cells as Hart does. Harts ownXumo Li was a key contributor to the

development of the ITS-90 scale. Thatsone reason youll find Hart cells in many ofthe national metrology institutes aroundthe world.

Water triple point cellsLike our metal fixed-point cells, Harts tri-ple point of water cells come in traditionaland mini quartz sizes, as well as smallstainless steel, which can be realized in adry-well calibrator. Our traditional cellshave been tested at NIST (see chart onfacing page) and are within a few micro-kelvin of NISTs cells.

If youre new to primary temperaturestandards and are considering a water tri-ple point cell, one of our cells is sure tomeet your requirements. We offer trainingthrough our seminars, insurance for ourglass cells, and our stainless steel cell justcant be broken!

Maintenance apparatusMaintaining fixed-point cells requireshigh-stability apparatus with tight

gradient control so plateaus last longerand your work is more productive. EveryHart maintenance apparatus, including ourmetrology furnaces and fluid baths, usestemperature controllers designed andmanufactured by Hart. These controllersare widely recognized for their unmatchedstability and uniformity control.

For metal fixed-point cells, choose fromone-zone, three-zone, or heat-pipe fur-naces for regular or mini cells. Optionalequilibration blocks fit into the furnacesfor annealing and comparison calibrations.Dont let the competition try to tell youthat a furnace fitted with process control-lers can provide the same performance asa furnace fitted with controllers designedspecifically for high-stability temperaturecontrol. With a Hart furnace youll get lon-ger cell plateaus with smaller gradientsthan you will from any other furnace onthe market.

6 Primary Standards

Why buy primary standardsfrom Hart?

SPRTsSPRTs are the only acceptable ITS-90 in-terpolation devices from the triple point ofhydrogen (13.8033 K) to the freezingpoint of silver (961.78 C). While mostSPRT manufacturers lost their design ca-pabilities years ago, Hart continues to de-velop new innovative designs with thelowest drift rates.

Hart manufactures quartz SPRTs in fourdifferent temperature ranges, includingcapsule SPRTs for low temperatures, anultra-stable SPRT for the range to 480 C,and a new working-standard SPRT forthe range to 660 C. Our metal-sheathSPRTs include a 25.5-ohm, contamina-tion-resistant SPRT. Hart SPRTs are thestandards of choice for many national me-trology institutes around the world.

ThermometryTraditionally, SPRT measurements havebeen made using expensive, difficult-to-use bridges. If you need 1 ppm accuracy,theres nothing that provides a betterprice/performance ratio than Harts 1590Super-Thermometer. The 1590 Super-Thermometer provides bridge accuracy ata fraction of the cost and provides a multi-tude of features that improve your produc-tivity. With a Super-Thermometer, there isvirtually no learning curve. Its so easy touse that youll be making measurementswithin minutes after switching it on.

If you truly need 0.1 ppm performance,the 5581 MIL Bridge offers conventionalDC measurement for a wide range. Itsperfect for temperature metrology workand for labs looking to combine tempera-ture with electrical resistance standards.

TrainingOnce youve determined which primarystandards products you need and youvemade a major investment, what abouttraining and after-sales support? Hartstemperature school offers a fun andunique seminar that provides all the an-swers to your toughest questions. Our 2-day Realizing and Approximating ITS-90seminar provides all the theory and somefirst-rate practical, hands-on experience toget you started. Youll learn some key tem-perature theory from former national labscientists and take part in practical experi-ments with our lab staff. If you want more,talk to us about individual ITS-90 training,where you can work alongside our cal labstaff in Harts primary standards lab, per-forming practical realizations on the cellsthat you just purchased.

Weve been making and using primarytemperature standards for many years,and we understand the issues you face inyour lab. Our own lab is accredited (NVLAPlab code 200348-0) and our uncertaintiesare among the best in the world. Whenyou buy primary standards, dont compro-mise the quality of the products, the repu-tation of your supplier, or the level ofservice and training they can provide.


Why buy primary standardsfrom Hart?

Comparison of Hart TPW Cells with NIST TPW Cell s/n A-13-1286NIST Cell 3 is used as a check and NIST Cell 4 is used as a reference

-0.03

-0.02

-0.01

0.00

0.01

0.02

0.03

0 1 2 3 4 5 6

days of measurements

{R[c

ell X

] /R

[cel

l 4(1

)] -

1}

(dT

/dW

) / m

K

Hart TPW Cells/n 5901A-7-1022

Hart TPW Cells/n 5901A-7-1024

NIST TPW Cell 3A-13-1291

NIST TPW Cell 4A-13-1286

Mingjian Zhao, Harts director of Primary StandardsEngineering.

Choosing the right platinum thermometeras your primary standard may be the mostcritical purchase decision in your lab. Un-fortunately, other manufacturers are prettysecretive about how their SPRTs are made.They wont tell you much more than youcan already see by looking at one. Manyof the leaders of a few decades ago havelost their original craftsmen and designscientists. Hart Scientific has one of only afew active SPRT design groups in theworld today.

So how do you know youre makingthe best purchase? Self-proclaimed exper-tise shouldnt convince you. You shouldexpect some sound evidence that thecompany is qualified in the ongoing sci-ence of SPRT development. At Hart, welltell you how we make an SPRT. Well letyou talk to the people here who design,build, and calibrate SPRTs. Finally, whenyou buy one, if you dont like it, well takeit back and return your money.

Hart has four quartz-sheath SPRTs,covering the ITS-90 range of 200 C to1070 C. The 5681 is used from 200 Cto the aluminum point at 660.323 C. The5683 is used from 200 C to 480 C withgreater long term stability. The 5684 and

the 5685 cover higher temperatures up to1070 C and can be calibrated at the silverpoint.

Yes, they have all the features youwould expect in a world-class SPRT. Theyhave gold-plated spade lugs, a strain-re-lieved four-wire cable, convection preven-tion disks, the finest quartz glassavailable, delustered stems, and the purestplatinum wire available.

The purity of a thermometers platinumwire is critical to meeting ITS-90 require-ments. Platinum resistance is measured bythe resistance ratio W at specified ITS-90 fixed points. Maintaining that purityover the life of the thermometer impactslong-term stability. The quartz glass tubeof the SPRT should be properly sealed toprevent contamination of the platinumwire. Others use mechanical assembliesand epoxy seals. These introduce addi-tional materials to the thermometers in-ternal environment and can be prone tomechanical failure, risking exposure of theplatinum to impurities.

Theoretically, the best seal would be adirect seal between the quartz glass andthe platinum wire. However, the quartzglass used in thermometer sheaths has a

very small coefficient of expansion whileplatinum has a much larger coefficient ofexpansion. If you simply sealed thesheaths glass to the platinum wire, thesedifferent rates of expansion would resultin a poor seal as the assembly is exposedto changing temperatures.

Weve figured out a way to match theexpansion coefficients of the glass sheathand the platinum wires. We do it by creat-ing a graduating seal thats made of 18separate pieces of glass, each with a dif-ferent coefficient of expansion. The expan-sion and contraction rate of the innermostpiece of glass matches that of the plati-num, resulting in an overall seal that pre-vents gas leakage and impuritypenetration for at least 20 years.

Fusing each piece of glass to the next isa painstaking process. Sure it costs us extra,but the results are worth it!

Theres more!We use only pure quartz glass materials

for the cross frames, disks, and tubes. Wedont use mica or ceramic materials. Addi-tionally, we have a special glass-treatingprocess to increase the resistance of thequartz to devitrification and remove moreimpurities than the typical cleaning process.

Weve done some research to find thebest-performing balance of argon to oxygenin the tube. Some oxygen in the sheath isnecessary to minimize the danger of theplatinum being poisoned by foreign metalsat high temperatures, but too much oxygenat temperatures below 500 C acceleratesthe oxidation process affecting the integrityof the platinum. Weve got a balance thatprovides exactly the right protection for theplatinum.

Each of these seemingly small things addsup to better uncertainties and less drift.

5681: 200 C to 670 CThis 25-ohm thermometer is the work-horse of the ITS-90 ranges. It can be cali-brated for any of the subranges from thetriple point of argon to the freezing pointof aluminum. The 5681 meets the ITS-90requirements for resistance ratios asfollows:

W(302.9146 K) 1.11807andW(234.3156 K) 0.844235

5683: 200 C to 480 CWhile SPRTs traditionally cover tempera-tures to the aluminum point (660 C), mostmeasurements occur between 100 Cand 420 C. The 5683 SPRT covers thisrange and more, from 200 C to 480 C,and does so with long-term stabilities that

8 Primary Standards

Drift rates as low as 0.0005 K

Proprietary gas mixture ensures high stability

Most experienced SPRT design team in the business

Quartz-sheath SPRTs

extended range SPRTs cant match. Typi-cal drift is less than 0.5 mK after 100hours at 480 C.5684 and 5685: 0 C to 1070 CITS-90 extended the use of the platinumthermometer from 630 C to 962 C. The0.25-ohm HTPRT sensor uses a strip-shaped support made from high-purityquartz glass. The 2.5-ohm model uses aquartz glass cross frame. Stability afterthermal cycling is excellent, and the de-sign is reasonably tolerant of vibration.Choose from 0.25-ohm or 2.5-ohm

nominal RTPW values. In addition to meet-ing the resistance ratio requirementsshown above, these thermometers meetthe following additional criterion:

W(1234.93 K) 4.2844

These glass probes really are a notchabove the rest!

Ordering Information

5681-S SPRT 25.5 , 670 C

5683-S SPRT 25.5 , 480 C, Ultrastable5684-S SPRT 0.25 , 1070 C

5685-S SPRT 2.5 , 1070 CMaple carrying case includedSee page 186 for SPRT calibration options.


Specifications 5681 5683 5684 5685

Temperature Range 200 C to 670 C 200 C to 480 C 0 C to 1070 C 0 C to 1070 C

Nominal RTPW 25.5 0.25 2.5 Current 1 mA 10 mA or 14.14 mA 3 or 5 mA

Resistance Ratio W(302.9146 K) 1.11807 andW(234.3156 K) 0.844235

W(302.9146 K) 1.11807 andW(1234.93 K) 4.2844

Sensitivity 0.1 / C 0.001 / C 0.01 / CDrift Rate < 0.002 C/100 hours at

661 C(typically < 0.001 C)

< 0.001 C/100 hours at480 C (0.0005 C typical)

< 0.003 C/100 hours at 1070 C(typically < 0.001 C)

Self-heating at TPW < 0.002 C under 1 mA current < 0.002 C under 10 mAcurrent

< 0.002 C under 3 mAcurrent

Reproducibility 0.001 C or better 0.00075 C or better 0.0015 C or better

RTPW drift after ThermalCycling

< 0.00075 C < 0.0005 C < 0.001 C

Sensor Support Quartz glass cross Quartz glass strip withnotches

Quartz glass cross

Diameter of Sensor PtWire

0.07 mm (0.003 in) 0.4 mm (0.016 in) 0.2 mm (0.008 in)

Protective Sheath Quartz glass, Diameter: 7 mm (0.28 in),Length: 520 mm (20.5 in)

Quartz glass, Diameter: 7 mm (0.28 in),Length: 680 mm (26.8 in)

25.52110

25.52115

25.52120

25.52125

25.52130

25.52135

25.52140

25.52145

25.52150

25.52155

25.52160

0 250 500 750 1000 1250Total Heated Time at Different Temperatures (hours)

Rtp

(oh

ms)

At 720 C At 675 C At 250 C At 450 C At 675 C

0.5 mK

A typical stability graph of a 5681 SPRT (#71122). Units are calibrated or shipped to customers after about 250hours of annealing.

Quartz-sheath SPRTs

Maximize your SPRTsperformanceAmazingly high accuracies can be ob-tained from a good SPRT if it is handledcorrectly. Expanded uncertainties as lowas a few tenths of a millikelvin at 0 Care possible provided you do thefollowing:

Avoid physical shock or vibration toyour SPRT. An SPRT is a delicateinstrument, highly susceptible tomishandling.

Make a measurement at the triplepoint of water after eachmeasurement. Use the resistance ratio(W) rather than the absoluteresistance to calculate thetemperature.

Measure at two different inputcurrents and extrapolate the results todetermine the value at zero power.This will eliminate the often-ignoredeffects of self-heating.

SPRTs. Art or science? It takes, in fact,quite a bit of both. The one thing not in-volved is mystery. Thats why Hart SPRTsalways include detailed published specifi-cations, including drift rates.

Our newest SPRT is no exception. Itwas designed by the same Hart metrol-ogists who have created a dozen differentSPRT designs used in national labs aroundthe world. And it performs just like we sayit does.

The 5698 Working Standard SPRT is atrue SPRT. It meets the ITS-90 ratio re-quirements for SPRTs and includes a Hart-designed, completely strain-free platinumsensor. With a 485-mm quartz sheath, this25-ohm SPRT covers a temperature rangefrom 200 C to 670 C. Long-term drift,which we define as the change in outputresistance at the triple point of water after100 hours at 670 C, is (after converting totemperature) less than 6 mKtypicallyless than 3 mK.

The 5698 is the perfect companion to aHart Super-Thermometer such as the1590, which reads 25-ohm SPRTs towithin 1 mK at 0 C and includes a num-ber of convenient features for workingwith SPRTs. Requiring 1 mA of excitationcurrent, the 5698 can also be used easilywith a Hart Black Stack, or even a Chub-E4 Thermometer.

If you need your SPRT calibrated by areputable calibration lab, we offer appro-priate calibration options by fixed-point in

our NVLAP-accredited lab. Our calibrationprices are as reasonable as our instrumentprices, so you get maximum value fromyour SPRT.

Why buy critical temperature standardsfrom companies unwilling to publish com-plete specifications? At Hart, we not onlyprovide excellent post-purchase supportso you have the best possible ownershipexperience, we also provide you all the in-formation we can before you purchaseincluding detailed performancespecifications.

Maybe there is some art mixed withour science. But that doesnt mean wekeep secrets. Trust your lab standards toHart Scientific.

10 Primary Standards

Fully conforms to ITS-90 SPRT guidelines

Drift rate typically less than 0.003 C

Multiple calibration options by fixed point

Unmatched performance-to-price-ratio

Specifications

TemperatureRange

200 C to 670 C

Nominal RTPW 25.5 ( 0.5)

Current 1.0 mA

ResistanceRatios

W(234.315K) 0.844235W(302.9146K) 1.11807

Sensitivity 0.1/ C

Drift Rate < 0.006 C/100 hours atmax temperature (typically 100 M at 661 C> 1000 M at 20 C


5699-S Extended Range Metal-SheathSPRT

Maple carrying case includedSee page 186 for SPRT calibration options.See page 38 for optional readouts.

Extended range metal-sheathSPRT

Sometimes you would like to make SPRTmeasurements but traditional SPRTs are toolong or awkward for a particular application.Our miniature glass capsule SPRTs are per-fect for cryogenics, calorimetry, and othermetrology work requiring small SPRTs.

These are true SPRTs. The high-purityplatinum wire is hand-wound on a glasscross frame in a strain-free design. Theglass capsule is designed to match thethermal expansion of the platinum wire toensure a true seal at all operating temper-atures. The capsules are pressure sealedand come protected in their own maplecases. Both models comply completelywith ITS-90 requirements for platinum

purity including the following resistanceratio:

W(302.9146K) 1.11807and

W(234.3156K) 0.844235The 5686 covers temperatures from

260 C to 232 C, so its perfect for cryo-genic applications. It is 5.8 mm (.23inches) in diameter and 56 mm (2.2inches) long.

These SPRTs are small but meetcustomary SPRT performance forreproducibility, reliability, and stability.


Temperatures from 260 C (13K) to 232 C

Stability typically 0.001 C over a 100 C range

Miniature capsule package eliminates stem conduction

Glass Sheath

Platinum Helix Platinum Lead Wires

Glass-Platinum SealSensor Support

Specifications

TemperatureRange

260 C to 232 C(13 K to 505 K)

Nominal RTPW 25.5

ResistanceRatio

W(302.9146 K) 1.11807W(234.3156 K) 0.844235

Drift Rate < 0.005 C per year overthe entire range

Self-heating atTPW

< 0.002 C under 1 mAcurrent

Reproducibility 0.001 C or better

RTPW Drift AfterThermalCycling

< 0.001 C

Filling Gas helium

Lead Wires Four platinum wires, 30 mmlong (1.18 in)

Size 5.8 mm dia. x 56 mm long(0.23 x 2.2 in)


5686-B Glass Capsule SPRT, 260 C to232 C

Maple carrying case includedSee page 186 for SPRT calibration options.

Glass capsule SPRTs

Youve spent some serious money to equipyour lab with some of the finest SPRTs inthe world because theyre the most accu-rate temperature measurement instru-ments you can buy. Now that youve gotthem, part of your job is to keep them per-forming at their highest levels. You can dothat with a Hart 9117 Annealing Furnace.

All HTPRTs and SPRTs are subject tomechanical shock no matter how carefullyyou handle them. This shock changes theresistance characteristics of the platinumand shows up as temperature measure-ment errors. Annealing relieves the stresson the platinum sensor caused by me-chanical shock and is recommended priorto any calibration of an SPRT.

In addition to removing mechanicalstrain, annealing also removes the oxida-tion from sensors that have been used for

long periods at temperatures between200 C and 500 C. Oxidation impacts thepurity of the element and therefore the ac-curacy of temperature readings. Oxide iseasily removed by annealing at 670 C forone or two hours.

During the annealing process, contami-nation must be controlled. At temperaturesabove 500 C, the lattice structure of aquartz sheath is transparent to metal ions.The thermometer must be cleaned and allcontaminating materials removed from itssheath. Annealing should only be done ina furnace thats designed to avoid emittingmetal ions during its heating cycle. Hartsolves this problem in its 9117 furnace byusing an alumina block that is speciallydesigned to guard against contamination.

The furnace also has a programmablecontroller specifically designed for the an-nealing process.

As a manufacturer of SPRTs, Hartmetrologists understand every aspect ofSPRT use and calibration procedures, in-cluding the annealing process. We usethis furnace in our own lab, so we knowexactly how well it works.

Specifications

TemperatureRange

300 C to 1100 C

Stability 0.5 C

Uniformity 0.5 C at 670 C (overbottom 76 mm [3 in])

Power 230 V ac ( 10 %), 50/60Hz, 12 A, 2500 W

DisplayResolution

0.1 C below 1000 C1 C above 1000 C

Display Accuracy 5 C

Thermal Wells Five: 8 mm diameter x430 mm long (0.31 x 16.9in)

Controller PID, ramp and soak pro-grammable, thermocouplesensor

Over-TempProtection

Separate circuit protectsfurnace from exceedingrated temperature limit

ExteriorDimensions(HxWxD )

863 x 343 x 343 mm(34 x 13.5 x 13.5 in)

Weight 28 kg (61 lb)

Communications RS-232


9117 Annealing Furnace(includes 2129 Alumina Block)

2129 Spare Alumina Block, 5 wells2125-C IEEE-488 Interface (RS-232 to IEEE-

488 converter box)


Guards against contamination

Anneals both SPRTs and HTPRTs

Fully programmable

Annealing furnace

The triple point of water (TPW) is not onlythe most accurate and fundamental tem-perature standard available, its also oneof the least expensive and simplest to use.

Water cells are essential!Triple point of water cells fill four criticalpurposes. First, they provide the most reli-able way to identify unacceptable ther-mometer drift between calibrationsincluding immediately after a calibration ifthe thermometer has been shipped. In-terim checks are critical for maintainingconfidence in thermometer readings be-tween calibrations. Second, they provide acritical calibration point with unequaleduncertainties.

Third, for users who characterizeprobes using ratios (that is, they use the

ratios of the resistances at various ITS-90fixed points to the resistance of the ther-mometer at the triple point of water, indi-cated by W), interim checks at the triplepoint of water allow for quick and easyupdates to the characterizations of criticalthermometer standards, which can beused to extend calibration intervals.

And lastly, the triple point of water iswhere the practical temperature scale(ITS-90) and the thermodynamic tempera-ture scale meet, since the triple point ofwater is assigned the value 273.16 K(0.01 C) by the ITS-90 and the Kelvin isdefined as 1/273.16 of the thermody-namic temperature of the triple point ofwater.

Good triple point of water cells containonly pure water and pure water vapor.

(There is almost no residual air left inthem.) When a portion of the water isfrozen correctly and water coexists withinthe cell in its three phases, the triplepoint of water is realized. Hart water cellsachieve this temperature with expandeduncertainties of less than 0.0001 C andreproducibilities within 0.00002 C.

In simple terms, water cells are madefrom just glass and water, but theresmuch more to it than that! For starters,thats not just any water in there.

Heavy waterHart cells contain carefully and repetitivelydistilled ocean water and are meticulouslyevacuated and sealed to maintain an iso-topic composition nearly identical to theinternational standard, Vienna StandardMean Ocean Water, or VSMOW.

The oxygen atoms found in most waterare predominantly comprised of eight pro-tons and eight neutrons (16O). Some oxy-gen atoms, however, have an extraneutron (17O) or two (18O). Similarly, thehydrogen atoms in water normally containonly a single proton (1H), but sometimescontain a neutron also (2H), resulting inheavy water. These isotopes coexist invarying proportions in ocean water, polarwater, and continental water, with oceanwater being the heaviest.

The ITS-90 recommends that watercells be made from water with substan-tially the isotopic composition of oceanwater. Research has shown that TPW er-rors associated with isotopic compositioncan be as large as 0.00025 C. The uncer-tainty contribution due to the effect of de-viation from VSMOW in Hart cells is lessthan 0.000007 C. Thats sevenmicrokelvin!

Hart offers two options for verifying theisotopic composition of any purchasedwater cell, both at nominal costs. We cansubmit a sample of water taken from yourown cell to a testing laboratory (after itwas completely manufactured, so you geta valid comparison) and give you the testreport. Or, we can send that water sampleto you in a sealed ampoule for you to con-duct your own tests. We can even providemultiple samples from the same cell (virtu-ally as many as youd like) so you cancheck for changes over time.

ImpuritiesFurther, the potential for errors due to wa-ter impurity is even greater than the errorsfrom isotopic composition. Hart cells un-dergo multiple distillation processes andutilize special techniques to retain waterpurity. Among other things, our primary


Easy-to-use, inexpensive standard with uncertainty better than 0.0001 C

Four sizes and two shells (glass and quartz) to choose from

Isotopic composition of Vienna Standard Mean Ocean Water

Triple point of water cells

standards scientists are able to connectquartz cells directly to the glass distillationsystem without using coupling hardwarethat may invite contamination.

Glass vs. quartzMost Hart water cells may either be pur-chased with borosilicate glass or withfused silica (quartz) housings. Whats thedifference? Glass is less expensive thanquartz, but its also more porous, allowingimpurities to pass through it over time. Re-search indicates that glass cells generallydrift about 0.000006 C per year whilequartz cells drift less.

Many sizesHart cells come in four general sizes. Mod-els 5901A, 5901C, and 5901D each comein either quartz or glass shells and include265 mm of thermometer immersion depth.The primary difference between thesemodels (other than the arm on the 5901A)is the inside diameter of the probe well.(See chart on page 16 and note that theinside diameter of the 5901C cells varieswith the shell material). A variety of bathsis available, which can maintain the triplepoint within these cells for many weeks.Accredited (NVLAP) test certificates areavailable with any cell under our Model1904-TPW.

5901A cells include an arm that can beused as a handle, a hook, or a McLeodgauge to demonstrate how much residualair is trapped in the cell. Carefully devel-oped manufacturing processes at Hartkeep the air bubble in a quartz cell assmall as the air bubble in glass cells.

A fourth size, the 5901B cell, comes ina glass version and is significantly smallerthan the other cells. It is designed for usein our Model 9210 Maintenance Appara-tus, which automates the realization andmaintenance of the TPW. The 9210-5901B combination is perfect for both cal-ibrating thermometers and providing peri-odic checks of sensor drift.

AccessoriesFor simplest realization of the TPW in ourlarger cells, the Model 2031A Quick StickImmersion Freezer uses dry ice and alco-hol to facilitate rapid formation of an icemantle within the cell without requiringconstant intervention while the mantleforms.

For best results, use a 3901 bushingwith your triple point of water cell. Abushing is used to improve the thermalcontact between your SPRT and the icemantle of your water triple point cell. Besure to choose a bushing that matches theinner diameter of the reentrant well of thecell and the outer diameter of the SPRT.Additionally, a small piece of foam (


Triple point of water cells

5901B-G5901A-G/Q

450

mm

50 mm

12 mm

265

mm

5901C-G

420

mm

265

mm

60 mm

13.6 mm

5901D-G/Q

420

mm

265

mm

60 mm

5901C-Q

420

mm

265

mm

60 mm

12 mm14.4 mm

8 mm

30 mm18

0 m

m

118

mm

Specifications

5901A-G 5901A-Q 5901C-G 5901C-Q 5901D-G 5901D-Q 5901B-G

Expanded Uncertainty (k=2) < 0.0001 C < 0.0002 C

Reproducibility 0.00002 C 0.00005 C

Dimensions 50 mm OD12 mm ID

450 mm long

60 mm OD13.6 mm ID

420 mm long

60 mm OD14.4 mm ID

420 mm long

60 mm OD12 mm ID

420 mm long

30 mm OD8 mm ID

180 mm long

Immersion Depth (watersurface to well bottom)

265 mm 118 mm

Material BorosilicateGlass

Fused Silica(Quartz)

BorosilicateGlass


BorosilicateGlass


BorosilicateGlass

Water Source Ocean

DVSMOW 10 ( 1 %) 20 18OVSMOW 1.5 ( 0.15 %) 3

Effect of Deviation fromVSMOW

7 K 14 K

For frequent use of traditional-size triplepoint of water cells, nothing helps saveyou time and hassle like a good mainte-nance bath. The 7312 Triple Point of Wa-ter Maintenance Bath keeps your cells upand running reliably for weeks at a timeeven during heavy usageand comes at aprice youll love.

The 7312 accommodates two TPW cellsand includes three pre-cool wells for prop-erly cooling probes prior to measurementswithin the cells. Stability and uniformity areeach better than 0.006 C, so your cellsstay usable for up to eight weeks. Whatevermethod you use for building your ice man-tles, you can be assured theyll last in a7312 bath.

An independent safety circuit protectsyour water cells from freezing and break-ing by monitoring the temperature of thebath and shutting down its refrigerationsystem should the bath controller fail.Noise-reduction techniques in the manu-facturing process ensure your bath doesntadd excessive noise to your lab.

With a temperature range from 5 C to110 C, this bath can also be used forcomparison calibrationsparticularly oflong-stem probesor maintenance of gal-lium cells. An optional gallium cell holdingfixture fits two cells, which, in a 7312bath, can maintain their melting plateausfor up to two weeks.

In fact, the 7312 is available with a time-saving 2031A Quick Stick ImmersionFreezer so you can build your ice mantlesquickly and hands-free. Just fill the 2031Ascondensing reservoir with dry-ice and alco-hol, insert it into the cell, and get some otherwork done while your ice mantle forms inless than an hour. (Alternatively, LN2 may beused.)

If youre using traditional-size TPWcells, dont take the time to create an icemantle only to watch it melt quickly as itsits in a bucket of ice. Maintain your cellsthe right way in a Hart 7312 TPW Mainte-nance Bath.


Harts 2031A Quick Stick Immersion Freezer offersunmatched convenience and simplicity in formingthe triple point of water ice mantle.

Specifications

Range 5 C to 110 C

Stability 0.001 C at 0 C(alcohol-water mix) 0.004 C at 30 C(alcohol-water mix)

Uniformity 0.003 C at 0 C(alcohol-water mix) 0.006 C at 30 C(alcohol-water mix)

TPW Duration Six weeks, typical(assumes correctly formedice mantle)

Set-PointAccuracy

0.05 C at 0 C

Set-PointRepeatability

0.01 C

DisplayResolution

0.01 C

Set-PointResolution

0.002 C; 0.00003 C inhigh-resolution mode

Access Opening 121 x 97 mm(4.75 x 3.8 in)

Immersion Depth 496 mm (19.5 in)

Volume 19 liters (5 gallons)

Communications RS-232 included

Power 115 V ac ( 10 %), 60 Hzor 230 V ac ( 10 %), 50Hz, specify

Size ( HxWxD ) 819 x 305 x 622 mm(12 x 24.5 x 32.25 in)

Weight 34 kg (75 lb)


7312 TPW Maintenance Bath (includesTPW Holding Fixture, MPGaHolding Fixture, and RS-232Interface)

2001-IEEE Interface, IEEE-4882031A Quick Stick Immersion Freezer

Maintains TPW cells for up to two months

Optional immersion freezer for simple cell freezing

Independent cutout circuit protects cells from breaking

TPW maintenance bath

Reprinted from Random News

IntroductionThe Standard Platinum Resistance Ther-mometer (SPRT) is the most accurate ther-mometer in the extended temperaturerange from 259 C to 962 C. The uncer-tainty of an SPRT can be as low as a fewtenths of a millikelvin (mK).

More and more metrologists are usingSPRTs as reference standards to calibrateother types of thermometers or to achievea high level of accuracy for any reason.However, the handling and use of anSPRT is as important to achieving a highlevel of accuracy as the design and perfor-mance of the SPRT itself. Several types oferrors can corrupt SPRT measurements.

Sometimes absolute resistance is usedto calculate temperature instead of the re-sistance ratio. When absolute resistance issubstituted for the resistance ratio, errorsof more than 10 mK at 660 C are com-mon. In addition, even when the correctmeasurement and calculations are made,the resistance of the SPRT in the triplepoint of water should be determined im-mediately after a high accuracy measure-ment is made with the thermometer.

The triple point of water measurementis often overlooked but is vital to accuracy.The relationship of the triple point of wa-ter measurement to SPRT accuracy is ex-plained with a few key points.

TPW and accuracyIn general, SPRTs have excellent stability;however, a small drift in resistance mighthappen now and then, especially aftertransportation, thermal cycling, or acci-dental rough handling. A change as lowas 1 ppm in resistance at about 660 C(the freezing point of aluminum) will beequivalent to a change of 1.1 mk in tem-perature. The stability required of a high-quality standard resistor is about 1 ppm.The working and environmental condi-tions normally associated with a standardresistor are much better than the condi-tions usually found when working with anSPRT. So a few ppm of stability might bethe best we can expect for most SPRTs.

The ratio of two resistances of an SPRTbased on two temperatures is much morestable than the stability expected when anabsolute resistance at a single fixed tem-perature is used. For example, using onlythe freezing point of silver as a referencepoint over a six year time frame, an SPRTexhibited a change of 5 ppm in its resis-tance [1]; this is equivalent to a change of7.5 mk in temperature. On the other hand,the change in the resistance ratio,

W(961.78 C) = R(961.78 C)/R(0.01 C)

was within 1 ppm (a change of 2 mk intemperature) across the same six-year pe-riod. This explains why the resistance ra-tio W(t) is specified by the InternationalTemperature Scales since 1960 instead ofthe absolute resistance R(t).

The best method for accomplishing thisratio is to use the Triple Point of Water asthe second temperature because of its ex-cellent stability and simplicity. It has beenspecified as a reference point for SPRTssince 1960 [2]. Thus, the highest SPRT ac-curacy possible is achieved when the re-sistance of an SPRT at the triple point of

water (Rtp) is made immediately after ameasurement at any other temperature.

Use of the ratio method also reduces sys-tem error introduced by any electronic read-out. This reduction in system error isimportant because as little as 0.7 PPM of er-ror in resistance will cause an error of 1 mkin temperature at 962 C (see table below).

Frequency of Rtp measurementWhen accuracy requirements dont extendto the highest levels, Rtp may need mea-suring only once a day, every few days, orat some other suitable interval. How fre-quently Rtp needs measuring depends onseveral factors, such as acceptable


PrecoolingAccess Hole

SPRT

ThermometerGuide Tube

Water Vapor

BorosilicateGlass

ReentrantThermometer Well

Water from Ice Bathor Alcohol

Water Solid (Ice)

Water Liquid

Metal Bushing

Soft Pad

Ice Bath orMaintenance Bath

Cushion

Use TPW and ratio method to improve

SPRT stability and accuracy

uncertainty, the stability of the SPRT, themeasuring temperature range, and theworking conditions. If the required uncer-tainty is 1 mk or so, Rtp measurementshould follow each Rt measurement. If ac-curacy requirements are 20 mk or more ina temperature range lower than 420 Cand the SPRT used is quite stable, the Rtpmight be measured once a week. The sta-bility over time of each SPRT must bemeasured, even when using SPRTs manu-factured in the same lot from the samesupplier.

When temperature measurements arehigher than 800 C, it is better to measurethe Rtp as soon as the SPRT cools down toroom temperature. Whenever possible, anSPRT should cool down to at least 500 Cwith a low cooling rate (about 100 C perhour). Otherwise, the SPRT should be an-nealed before making a measurement atthe triple point of water.

A suitable annealing procedure is atwo-hour anneal at 700 C at the end ofwhich the SPRT is allowed to cool to450 C over a period of about two and onehalf hours. After this initial cooling period,the SPRT can cool quickly to room temper-ature. Fast cooling from high temperaturesabove 500 C may cause significant in-creases in Rtp because of the quenching-ineffect on lattice defects found in platinumwire. This increase of Rtp could be as largeas 30 mk.

Can the Rtp given in the NISTCalibration Report be used tocalculate the ratio?Some metrologists may feel the Rtp mea-sured by NIST is more accurate than thatmeasured in their own lab, so they preferto use the value for Rtp given in the NISTCalibration Report to calculate the resis-tance ratio in the interpolation equation.While its true that the accuracy of NISTsmeasurements are generally much betterthan those done in other labs, the Rtp ofyour SPRT may have changed during

transportation, so it should be measuredagain in your own lab. Furthermore, theRtp should be measured using the sameinstrument and time frame as the Rt to re-duce system error with the readout in-cluded in the measuring procedure. It isimportant to always use the same readoutinstrument to measure both Rt and Rtp.

Avoiding mechanical strain and theannealing procedureAn SPRT is a delicate instrument. Shock,vibration, or any other form of accelerationmay cause strains that change its temper-ature-resistance characteristics. Even alight tap, which can easily happen whenan SPRT is put into or taken out of a fur-nace or a triple point of water cell, maycause a change in Rtp as high as 1 mk.Careless handling of an SPRT over thecourse of a year has resulted in Rtp in-creases equivalent to 0.1 C.

Annealing at 660 C for an hour willrelieve most of the strains caused by mi-nor shocks and nearly restore the Rtp to itsoriginal value. If the maximum tempera-ture limit for an SPRT is lower than660 C, it should be annealed at its maxi-mum temperature. Such an annealing pro-cedure is always advisable after any typeof transportation.

The annealing furnace should be veryclean and free of metals, such as copper,iron, and nickel. SPRTs are contaminatedwhen they are annealed in furnaces con-taining a nickel block, even when theSPRTs were separated from the nickelblock by quartz sheaths [ 3 ]. Well de-signed, clean annealing furnaces are im-portant for quality measurements withSPRTs.

ConclusionsSPRTs are among the finest temperaturemeasuring devices known. However, highaccuracy comes at a price and not just interms of money. Patience, care and proper

procedures are major factors in producinghigh quality measurements.

Support instruments such as triplepoint of water cells are inexpensive andsimple to use. Annealing is a well under-stood process. Uncompromised measure-ments are possible in almost everylaboratory situation.

References1, Li, Xumo et al, Realization of the In-

ternational Temperature Scale of 1990 be-tween 0 C and 961.78 C at NIM,Temperature, Its Measurement and Con-trol in Science and Industry, Volume 6,Part 1, p. 193 (1992).

2, CGPM (1960): Comptes Rendus desSeances de la Onzieme Conference Gener-ale des Poids et Mesures, pp. 124-133.

3, Li, Xumo et al, A New Type of HighTemperature Platinum Resistance Ther-mometer, Metrologie, 18 (1982), p. 203.


Temperature(C)

The temperature error caused byan error of 1 PPM in resistance

measurement (mK)

The resistance error equivalentto an error of 1 mK in

temperature (PPM)

200 0.04 25.4

100 0.14 6.9

0.01 0.25 4.0

232 0.51 2.0

420 0.74 1.4

660 1.1 0.9

962 1.5 0.7

Use TPW and ratio method to improve

SPRT stability and accuracy

Several companies manufacture highquality resistance bridges for both AC andDC applications that can take measure-ments at the 0.1 ppm level. Research hasshown that all of them compensate wellfor any theoretical inaccuracies predictedin their design.

We like the MI bridge because we feelconfident about its measurements, and itssoftware gives us more information thanwe can get from the other instruments.While its true we do use the other bridgesfor certain functions we undertake in ourlab, including some experimental testing,we use the MI bridge every day for fixed-point calibrations of SPRTs.

The 5581 Bridge performs a true auto-balancing procedure to nine significantdigits. As the check proceeds, the bridgesteps through an internal comparison ofthe transformers windings, the results ofwhich are recorded to track its perfor-mance over time.

Another function of this bridge is itsreal-time uncertainty analysis program. Inthis mode, you enter external uncertaintyfactors such as the uncertainty of your re-sistor, and the 5581 combines your infor-mation with its own uncertainties to

compute a system uncertainty for yourmeasurement.

The optional Windows compatiblecontrol software offers history logging andregression analysis, along with uncer-tainty analysis and the auto-self-checkfeature. The program also calculates stan-dard deviations if you need them. You canenter coefficients for your SPRT and readtemperature rather than resistance.

Of course, if you prefer, you can operatethe bridge manually. The choice is yours,but either way youll find this to be a greatbridge to use.


Measurement uncertainty to 0.025 mK Uses conventional standard resistors

Specifications

Bridge

Range/Accuracy

0.001 to 0.01 : < 5 ppm0.01 to 0.1 : < 0.5 ppm0.1 to 1 : < 0.1 ppm0.1 to 10 K: < 0.1 ppm10 K to 10 K: < 0.2 ppm

Linearity 0.01 ppm

Max Ratio 13:1

TestCurrents

10 A to 150 mA, 30-Voltcompliance

CurrentReversal

Automatic 4 to 1000 seconds

Power 100, 120, 220, and 240 V( 10 %), 4763 Hz, 180 VA

Weight 60 lb (27.3 kg)

Dimensions(WxHxD )

432 x 279 x 381 mm(17 x 11 x 15 in)

Scanner

Inputs 20/10

Operation Matrix

ThermalEMFs

< 500 nanovolts

ErrorContribution

< 20 nanovolts

ContactRatings

Relay2-coil latching

MaxCarryingCurrent

2 A (ac/dc) (optional 30 A)

ContactResistance

1012

Inputs andOutputs

Tellurium Copper(rear panel)

IEEE-488 24-pin IEEE-488

Weight 5313-001: 18 kg (40 lb)5313-002: 9 kg (20 lb)

Dimensions(WxHxD )

5313-001: 432 x 279 x381 mm (17 x 11 x 15 in)5313-002: 127 mm H (5 in)


5581 MI Bridge5313-001 Scanner, 20 channels5313-002 Scanner, 10 channels5313-003 IOTech 488 Interface Card5313-004 Windows Software

The DC AdvantageAC bridges are more susceptible to elec-trical interference than DC bridges.Therefore, when AC furnaces are used,DC bridges are preferred. The likelihoodof electrical interference increases attemperatures above the freezing point ofAluminum (660.323 C), because the in-sulation resistance of the furnace and theSPRT decline significantly.

DC bridge

By now youve probably been preparingbudgets for the new primary standardsyou need in your lab. Well, youve come tothe right place because weve got a com-plete line of the best metal freeze-pointcells you can buy.

Fluke scientists have designed andtested metal freeze-point cells for manyyears. Not only do we manufacture all themajor freeze-points, our metrologists havewritten extensively on the theory and useof cells and have created new designscovering a range of applications no othercompany can match. They can answer ev-ery question you have on freeze-pointsand explain how we have handled the ra-diation losses along the well, minimizedpossible stem error and made pressurecorrections for the highest possible mea-surement accuracy.

Pure metals melt and freeze at a uniquetemperature through a process involvingthe absorption, or liberation, of the latentheat of fusion. If 100% pure metals wereavailable, each cell of a particular typewould melt and freeze at its exact theoret-ical temperature, but since 100% puremetal is not possible, cells vary slightlyfrom their theoretical absolute.

The best freeze-point cells are the onesthat get very close to their theoreticalfreezing temperature and provide a tem-perature plateau thats stable and longlasting for calibration work. Changes assmall as 0.01 mK (0.00001C) are measur-able, so cell uncertainty is definable. Flukecells come very, very close to the absolutetheoretical limits possible with todaysmaterial science. We use 99.9999% puremetal. Six 9s purity is as good as it gets.Some manufacturers may use five 9s.Make sure they specify purity on theirquotations.

The shape of the curve generated byfreezing a fixed-point cell tells a lot aboutthe purity of the metal. Unlike most manu-facturers we provide certificates for eachmetal fixed-point cell at no extra charge.The certificates we provide include threegraphs of full plateaus, as measured by abridge and SPRT, and a calculated purityfor the metal sample. Exceptions includegallium cells (extremely long plateaus),gold and copper cells (SPRT temperaturerange exceeded).

We test each cell three times so thatwe can evaluate the stability of the pla-teau. This provides evidence that the cellis good and that the metal sample is notbeing contaminated over time. Some ex-amples of freezing curves from Flukefixed-point cells are pictured below. The

graphs also demonstrate the length ofFluke fixed-point cell freezing plateaus.

Freezing plateau length provides an-other quality measure for comparing fixed-point cells. The longer the plateau, thebetter the cell and the more cost-effectiveit is to use. As indicated in the table, Flukemetal fixed-points have plateaus thatrange from 20 hours to 14 days depend-ing on the type of cell. No other producerof fixed-points beats the performance of aFluke fixed-point cell.

Further proof of cell performance isavailable through our accredited cell cer-tification service (NVLAP Lab Code200348-0). This service is a rigorous testthat involves repeatedly testing the melt-ing and freezing plateaus of the fixed-point cell and then comparing the thermo-dynamic temperature to that of a NIST-certified reference cell. The result is ahighly-qualified primary standard

instrument with a certificate traceable tonational and international standards.

All freeze-point cells have to be usedand maintained in specially-designedfreeze-point furnaces. Furnaces come in avariety of temperature ranges, and eachone must have a very uniform temperatureregion for the freezing and melting pro-cess. High accuracy is achieved by the for-mation of two liquid-solid interfacesduring freezing. One interface should beadjacent to the inner surface of the cruci-ble, and the other should be on the outersurface of the central well, which is theclosest point to the thermometer.

A very thin solid shell should form onthe central well, and the concentration ofimpurity in this thin shell should be muchlower than the average impurity in thecore metal. The temperature on the sec-ond liquid-solid interface should be veryclose to that found in a 100% pure metalunder ideal conditions.


Why use Fluke metal freeze-point cells?

Substance Impurity LevelDeviation from Pure

Liquidus PointAcheivable

Plateau LengthMercury 99.999999 % 0.002 mK 30 hoursGallium 99.99999 % 0.014 mK 14 daysIndium 99.9999 % 0.24 mK 25 hoursTin 99.9999 % 0.30 mK 25 hoursZinc 99.9999 % 0.54 mK 25 hoursAluminum 99.9999 % 0.67 mK 20 hoursSilver 99.9999 % 1.12 mK 20 hours

Sample freezing curve for 5904 indium cell

Flukes freeze-point furnaces providethe proper temperature, stability and uni-formity to create the longest, most stabletemperature plateau possible.

The immersion depth of a thermometerin the liquid metal (the distance from thesurface of the liquid metal to the middlepoint of the thermometers sensor) is ap-proximately 180 mm in cells made byFluke. The outside diameter of each cell is48 mm, the thermometer well is 8 mmand cell length is 290 mm.

Cells are delicate and must be handledwith extreme care. Fluke provides specialstorage cases, gloves for handling andcomplete care instructions. Cells are handcarried rather than shipped to ensurecomplete integrity of measurements.

The performance of every Fluke cell isguaranteed. When you call, ask about ourtemperature calibration school where youcan learn more about the theory and oper-ation of all metal freeze-point cells. Wellanswer every question you have and helpsolve any problem you encounter. We usemetal freeze-point cells in our lab everyday, and our metrologists have decades ofexperience with them.


Why use Fluke metal freeze-point cells?

Sample freezing curve for 5905 tin cell.

Sample freezing curve for 5908 silver cell.

Hart scientists have designed and testedITS-90 fixed-point cells for many years.Not only do we manufacture all the majorfixed points, our metrologists have writtenextensively on the theory and use of cellsand have created new designs covering arange of applications no other companycan match.

Our testing of fixed-point cells is alsounmatched. The scope of our accreditationincludes the testing of ITS-90 fixed-pointcells. Each cell may be purchased with thisintercomparison option, which includescomparing the equilibrium value of yourcell against that of a reference Hart cell.

Traditional freeze-point cellsIf you want true primary temperature stan-dards capability, you want metal freeze-point cells that are very close to the theoret-ical freezing temperature and provide pla-teaus that are both stable and long lasting.

Harts metal freeze-point cells are theculmination of more than 20 years of pri-mary standards experience. No other com-pany has as much experience in thedevelopment of metal fixed-point cells asHart. Thats why youll find Hart cells in

many national metrology institutes aroundthe world.

Each Hart cell is carefully constructedin our ultra-clean, state-of-the-art lab, us-ing high-density, high-purity graphitecrucibles containing metal samples withpurity of at least 99.9999 % (six 9s) and,in many cases, 99.99999 % (seven 9s).The crucible is enclosed within a sealedquartz glass envelope that is evacuatedand back-filled with high-purity argongas. A special sealing technique is used toseal the cell at the freezing point. Wemeasure and record for you the precisepressure of the argon gas to ensure themost accurate corrections for pressure.

Once manufactured, all Hart cells aretested and supplied with an assay ofmetal-sample purity. Every traditional sizeITS-90 cell further undergoes more rigor-ous testing in our primary standards labwhere we realize melt-freeze curves andperform a detailed slope analysis to con-firm cell purity. If you want more data,well give you an optional intercomparisonwith our own reference cells.

Gallium cellsGallium cells are a great reference for vali-dation of instruments subject to drift (likeSPRTs), and theyre important for calibrat-ing sensors used near room or body tem-peratures, in environmental monitoring,and in life sciences applications.

Harts 5943 Gallium Cell is sealed in astainless steel envelope. High purity gal-lium (99.99999 %) is enclosed in a plasticand metal shell. The stainless steel con-tainer is then filled with pure argon gas atone standard atmosphere at the melting-point temperature.

Gallium expands by 3.1 % when itfreezes requiring the cell to have flexiblewalls. Unlike some manufacturers cells,which are made from PTFE enclosure ma-terials, our cells dont need pumping andrefilling because theyre not gas perme-able. In fact, we guarantee our cells willmaintain their uncertainty of < 0.1 mK forat least five years. Realization and mainte-nance of the cell is automated with our9230 Maintenance Apparatus (see page33). This apparatus will provide meltingplateaus up to eight days and a conve-nient control to automatically achieve anew melt plateau each week with an in-vestment of just five minutes. Never hasthe maintenance of a world-class galliumcell been easier.

Water cellsWhile simple ice baths are often used as acalibration point at 0 C, their limitationsinclude gradients, purity problems, repeat-ability issues, and variances in construc-tion and measurement techniques. Triplepoint of water cells not only solve theseproblems; they represent the most usedtemperature on the ITS-90, and theyre in-expensive to own and use.

Hart makes three traditional-size TPWcells (see page 14) that have been repeat-edly proven in national labs to surpasstheir published uncertainty specification of 0.0001 C. Ice mantles may be formedusing dry ice, LN2, or immersion freezersand can last for up to two months whenmaintained in our 7012 or 7312 baths.

Open metal cellsMade from the same materials and withthe same manufacturing techniques astheir sealed counterparts, Harts new se-ries of open metal fixed-point cells in-clude a high quality valve for connectingto a precision pressure-handling systemwithin your lab. Using such a system, thecell can be evacuated, charged, andpurged several times with a pure inert


Best cell uncertainties commercially available

Every ITS-90 fixed point available from mercury to copper

Plateaus last days (gallium for weeks and TPW for months)

Manufactured and tested by Harts primary standards scientists

ITS-90 fixed-point cells

gas, then charged again to a regulatedpressure level while measurements aremade with the cell.

Once assembled and tested, each HartITS-90 open cell further undergoes morerigorous testing in our lab, unlike cellsfrom some manufacturers who providetheir open cells as a kit of parts, withoutany test data.

Because open cells allow users to mea-sure the pressure within the cell, uncer-tainties due to pressure corrections maybe minimized. Use of open cells is nowbeing suggested by the CCT, and opencells can be used for demanding tempera-ture-versus-pressure applications, as wellas precision SPRT calibrations.

The height of these cells has been ex-tended to allow easy access to the gasvalve while the cells are in use. Purequartz-wool insulation and four high-pu-rity graphite discs prevent heat loss fromthe metal sample to the pressure regula-tion system while optimizing vertical tem-perature gradients within the cell. Eachcell has an outside diameter of 50 mm (2

inches) and a height of 600 mm (23.5inches)(silver and copper cells are 700mm [27.6 inches] tall).

When it comes to primary temperaturestandards, Hart supplies more equipmentthan all of our competitors combined. Ifyour goal is to reduce uncertainty, start by

buying from the company that supports itsproducts better than any other metrologycompany in the world. Why trust your pri-mary standards to any other company?



Specifications

ModelFixedPoint Style

AssignedValue ( C)

OutsideDiameter

InsideDiameter

Total OutsideCell Height Depth

CellUncertainty(mK, k=2)

Certification(mK, k=2)

5900 Mercury Stainless Steel 38.8344 31.8 mm 8.0 mm 417.6 mm 208 mm 0.2 0.25

5904 Indium Traditional QuartzGlass

156.5985 48 mm 8 mm 282 mm 195 mm 0.7 0.7

5905 Tin Traditional QuartzGlass

231.928 48 mm 8 mm 282 mm 195 mm 0.5 0.8

5906 Zinc Traditional QuartzGlass

419.527 48 mm 8 mm 282 mm 195 mm 0.9 1.0

5907 Aluminum Traditional QuartzGlass

660.323 48 mm 8 mm 282 mm 195 mm 1.3 1.8

5908 Silver Traditional QuartzGlass

961.78 48 mm 8 mm 282 mm 195 mm 2.4 4.5

5909 Copper Traditional QuartzGlass

1084.62 48 mm 8 mm 282 mm 195 mm 10.1 12.0

5924 Indium Open Quartz Glass 156.5985 50 mm 8 mm 596 mm 195 mm 0.7 0.7

5925 Tin Open Quartz Glass 231.928 50 mm 8 mm 596 mm 195 mm 0.5 0.8

5926 Zinc Open Quartz Glass 419.527 50 mm 8 mm 596 mm 195 mm 0.9 1.0

5927A-L Aluminum Open Quartz Glass(long)

660.323 50 mm 8 mm 696 mm 195 mm 1.3 1.8

5927A-S Aluminum Open Quartz Glass(short)

660.323 50 mm 8 mm 596 mm 195 mm 1.3 1.8

5928 Silver Open Quartz Glass 961.78 50 mm 8 mm 696 mm 195 mm 2.4 4.5

5929 Copper Open Quartz Glass 1084.62 50 mm 8 mm 696 mm 195 mm 10 12.0

5943 Gallium Stainless Steel 29.7646 38.1 mm 8.2 mm 250 mm 168 mm 0.1 0.1Certifications at lower uncertainties are available for national laboratories.Depth is measured from the bottom of the thermometer well to the top of the pure reference material.


5900 Mercury Cell, Stainless Steel5904 Indium Cell, Traditional Quartz

Glass5905 Tin Cell, Traditional Quartz

Glass5906 Zinc Cell, Traditional Quartz

Glass5907 Aluminum Cell, Traditional

Quartz Glass5908 Silver Cell, Traditional Quartz

Glass5909 Copper Cell, Traditional Quartz

Glass

5924 Indium Cell, Open Quartz Glass5925 Tin Cell, Open Quartz Glass5926 Zinc Cell, Open Quartz Glass5927A-L Aluminum Cell, Open Quartz

Glass, Long5927A-S Aluminum Cell, Open Quartz

Glass, Short5928 Silver Cell, Open Quartz Glass5929 Copper Cell, Open Quartz Glass5943 Gallium Cell, Metal Cased2068-D Stand, Fixed-Point Cell, Black

Delron



Direct Comparison of Hart Scientific Ga Cell (s/n Ga-7010)with NIST Reference Ga Cell (Ga 98-1)

1.118 100 5

1.118 100 7

1.118 100 9

1.118 101 1

1.118 101 3

1.118 101 5

0 1 2 3 4 5 6 7

number of readings

W(G

a)

NIST Ga 98-1,set #1

Hart Ga-7010,set #1, -0.07 mK

NIST Ga 98-1,set #2

Hart Ga-7010,set #2, -0.05 mK

0.10 mK

Direct Comparison of Hart Scientific Sn Cell (s/n Sn-8014)with NIST Reference Sn Cell (Sn 88A)

1.892 589 8

1.892 590 0

1.892 590 2

1.892 590 4

1.892 590 6

1.892 590 8

1.892 591 0

1.892 591 2

0 1 2 3 4 5 6 7

number of readings

W(S

n)

NIST Sn 88A,set #1

Hart Sn 8014,set #1, -0.25 mK

NIST Sn 88A,set #2

Hart Sn 8014,set #2, -0.24 mK

0.27 mK

What is the uncertainty of my cell?Fixed-point cells are standards which em-body reproducible physical phenomena.The uncertainty associated with these stan-dards can be viewed in two ways.

The first way is based on the purity ofthe constituent components only. Unfortu-nately, most of the assays provided by man-ufacturers of pure metals are not ofsufficient quality to make a determination ofthe purity of the supplied metal to the levelof uncertainty required. To be used in real-izing the ITS-90, it would typically be nec-essary to have a high quality traceableassay capable of verifying 99.9999% purityor better. Even with an assay, additionalevidence of the purity is necessary. This

evidence includes an analysis of the slopeof the freezing and melting curves and acomparison against another cell whichmakes similar or better claims of purity.Finally, because the temperature of a fixedpoint cell is defined at one atmosphere,pressure traceability is required as well.

The second approach to fixed-point un-certainties is similar but shifts the emphasisaway from the traceable assay and derivesits traceability through inter-comparisonwith another traceable fixed-point cell. Inthis case, the assay and the slope analysisbecome the supporting evidence for theobserved difference against the traceablecell. This second approach represents actual

observed performance in a laboratory ratherthan unproved claims and weakly justifiedhopes. This approach is particularly impor-tant with sealed cells because there is noway to verify the pressure within a sealedcell after it has been sealed.

Harts published specs are guaranteedand can be verified through an optional ac-credited certification in our primary temper-ature lab. Are the values assigned to yourfixed-point cells traceable?

Open cells allow users to minimize the uncertaintyfrom pressure corrections by regulating cell pressuresthemselves.

Reprinted from Random News

What is an intrinsic standard?Intrinsic standards are defined by theNCSL as standard(s) recognized as havingor realizing, under prescribed conditions ofuse and intended application, an assignedvalue the basis of which is an inherentphysical constant or an inherent and sta-ble physical property. Thermometricfixed-point cells are included in the NCSLCatalog of Intrinsic and Derived Stan-dards. Some other well-known intrinsicstandards include the Josephson-arrayvoltage standard, the Quantum Hall resis-tance standard, and the Cesium atomicfrequency standard.

The definitions themselves do not di-rectly address the issues of uncertainty,traceability, or accreditation. However, inthe case of thermometric fixed points,these issues are covered in the notes tothe definition. The notes indicate that thevalue is assigned by consensus and neednot be established through calibration.The uncertainty is said to have two funda-mental components: (a) that associatedwith its consensus value, and (b) that as-sociated with its construction and applica-tion. Traceability and stability are said tobe established through verification at ap-propriate intervals. Verification can eitherbe based upon application of a consensusapproved test method or throughintercomparison. Furthermore, theintercomparison may be accomplishedwith standards in a local quality controlsystem or external standards includingnational and international standards.

Do fixed-point cells fit thedefinition?The basic parameter of the cell, the phasetransition, is believed to be an inherentand stable physical property of the cellwhen used under prescribed conditions.The generally accepted values for thetemperatures of the phase transitionsalong with corrections due to pressure and

hydrostatic head are assigned by the ITS-90, the current temperature scale adoptedby the BIPM. From the values given and bytaking a few measurements, the theoreti-cal temperature of the phase transitioncan be calculated. Also defined by the ITS-90 is the intended application, namely, asdefining thermometric fixed points to beused in conjunction with an appropriateinterpolation instrument and associatedequations. Finally, the conditions of useare described by supplementary informa-tion to the ITS-90 as well as a significantbody of literature. Although not everyoneagrees on the exact procedures, for themost part, they are quite well understoodand accepted. It appears, then, that fixed-

point cells can indeed be considered in-trinsic standards.

However, several issues arise: First, theITS-90 discusses fixed points based onphase transitions of pure substances. Anideal substance behaves differently thanthe real materials that we are able to ob-tain. The departure depends on the impu-rity content of the sample once it isassembled into a cell. For very highly purematerials, the slope of the plateau can beused to approximately determine the pu-rity, but the absolute temperature remainsdifficult to predict. Second, the ITS-90does not directly specify an optimum celldesign, furnace or cryostat design, or min-imum purity requirements. Quite to thecontrary, many designs and options are


IssueVerification via

SPRTVerification via Industry

Intercomparison NIST MAP Cell Certification

Uncertainty for Cells maybe no maybe yes

Traceability for Cells maybe no maybe yes

Laboratory Apparatus yes yes yes maybe

Laboratory Equipment yes yes yes maybe

SPRT Calibration no maybe yes no

Procedure Evaluation maybe maybe yes no

Computation Evaluation no maybe yes no

Sealed metal freeze-point cells

Traceability and thermometricfixed-point cells

presented in the literature. Although ex-perimental results may suggest one designover another, the conclusions regardinguncertainty are not always clear. Third,the measurement results obtained from acell are highly dependent upon experi-mental conditions. Having a good cell isonly part of the exercise. Fourth, since athermometer always measures its owntemperature, the thermometer must beable to come to thermal equilibrium withthe cell. This is affected by the cell, its ap-paratus, the thermometer, and the tech-nique used to realize the phase transition.Finally, since we wish to perform trace-able calibrations, knowing only the theo-retical temperature is not adequate.

To certify or not to certify?So, how do we demonstrate that our cellsembody the ITS-90 definitions and howdo we establish traceability? Before wetackle those points, there are three issuesthat we must consider. First, whatevermethod we choose, we must perform a ro-bust uncertainty analysis on our measure-ments. The uncertainty associated withthe temperature of the phase transition isonly one component among many thatshould be considered. Second, statisticalprocess control (SPC) is critical wheneverthe measurement relies upon physicalprocesses (such as the realization of aphase transition). Through SPC, we canquantify the repeatability of our processand show that the test experiment repre-sents the process. Third, although SPRTsand sealed cells can be used as transferstandards, intercomparison of fixed-pointcells over long distances is problematic.

That having been said, the simplestmethod is to perform measurements inyour cells using a calibrated SPRT. If theuncertainty of the measurement is suffi-ciently small, the temperature can then beshown to be within the estimated uncer-tainty based upon the theoretical consid-erations of the cell construction. In thecase of low purity cells (five 9s or lower),it may be appropriate to assign a tem-perature and uncertainty to the realizationobtained from the cell. These methodsmay be considered the least robust andwill typically result in the largest uncer-tainties, but they can be shown to betraceable determinations.

A similar but more complicated methodis to intercompare calibration results withpeer laboratories or a reference laboratoryusing a suitable transfer standard. Al-though this type of analysis cannot di-rectly certify the performance of a fixed-point cell, it will show your laboratorys

capability to calibrate SPRTs using them.In many cases, this is what you are at-tempting to illustrate. A well-designedintercomparison will evaluate the resultsof the calibration, the raw and intermedi-ate data, and the computations. Much in-sight can be obtained from such scrutiny.NIST offers a measurement assurance pro-gram (MAP) to satisfy this need.

Finally, the cells can be tested by anexperienced laboratory that has the capa-bility to provide traceable results with un-certainties in the neighborhood of yourrequirements. If the laboratory performingthe test is using its own equipment andapparatus, this type of test will show theperformance of the cell only. Additionalexperiment and uncertainty evaluationmay be required for use in your laboratory.Also, this method will not illustrate yourlaboratorys ability to use the cells prop-erly. The major advantage of this methodis that it can provide the lowest overalluncertainty. Often, this is referred to asdirect comparison.

So, what should we do?At one time, it was considered acceptableto use a fixed-point cell with plateau anal-ysis to show that the cell was behaving it-self. This approach has proven to beinadequate as our understanding evolvesand we try to improve our laboratories.Moreover, laboratory accreditation re-quires that we follow rigorous proceduresin evaluating our uncertainties. If our un-certainties approach National MetrologyInstitute level, our data and analysis mustjustify this. And, the fixed-point cell is acritical component in the uncertainty eval-uation. The intrinsic standard argumentprovided by the NCSL does state that somelevel of intercomparison is necessary.

Presumably, the NCSL expects theintercomparison to be appropriate to theuncertainty claimed and based on themost current practices.

We must choose the method thatmakes economic sense and that satisfiesour requirements. For example, if we arecalibrating secondary PRTs using minifixed-point cells, we may be justified inusing a calibrated SPRT to verify the per-formance of our cells. Many laboratories(several accredited) use this method withsuccess. Traceability can easily be demon-strated and the uncertainty analysis isstraightforward. On the other hand, if wehave spent tens of thousands of dollars ona system to calibrate SPRTs and theseSPRTs are used for critical measurements,the NIST MAP program is a very good op-tion, provided we qualify. Finally, if wewish to provide cell certifications, we willrequire a set of certified reference cellsalong with a robust uncertainty evalua-tion. At Hart, we use a combination of allthree methods.

Conclusion and recommendationsSo, are fixed point cells intrinsic standardsor certified artifacts? It really doesnt mat-ter. Both viewpoints require testing andtraceability. Both approaches require rig-orous uncertainty analysis that must sat-isfy the scrutiny of our accreditationassessors, our customers, and our peers.And each perspective can be logically jus-tified. At Hart, we treat some as intrinsicstandards and others as certified artifacts.Our uncertainty analyses are as rigorousas we can make them and we welcomecomment from our peers. Additionally, theNIST thermometry staff is available to as-sist in the development of uncertaintybudgets, meeting traceability and accredi-tation requirements, as well as uniquetesting requirements. Finally, we use theapproach that will result in the lowest un-certainties for a given set of equipmentand techniques. After all, isnt that whatits all about?


Traceability and thermometricfixed-point cells

Water triple point cells

Several companies manufacture freeze-pointfurnaces. Most of these furnaces are of ade-quate theoretical design and of reasonablequality. Most are priced similarly. However,there is a difference that cant be seen fromspecifications or price, and thats how wellthe furnace performs with the freeze-pointcells theyre designed to maintain.

Establishing and maintaining a freeze-point plateau is what these furnaces aresupposed to be about. Nothing they do ismore important than this performance issue.

Hart Scientific makes three freeze-pointfurnaces that, when combined with Hartfreeze-point cells, produce the longestplateaus in the industry. A Hart furnaceand cell can establish plateaus that rangefrom 24 to 40 hours or more.

Fixed-point furnaces can also be usedfor comparison calibrations and for an-nealing. In these processes, stability anduniformity are very important, and nothingspeaks more about stability and uniformitythan the length of the plateaus produced

by the furnace. No other furnace beats aHart furnace where it counts.

All three of these furnaces have exter-nal cooling coils for circulation of tapwater at less than 60 PSIG and approxi-mately 0.4 GPM to reduce heat load to thelab. They also come with RS-232 portsand have equilibration blocks available forcomparison calibrations. IEEE-488 inter-face packages are also available, if thatsyour preference.

One of Harts three fixed-point furnacemodels will meet your needs. Remember,the length of the plateau is the best mea-sure of a furnaces performance. Call us forperformance data on actual cell freezes andtest data on furnace gradients.

9114This furnace has a range of 100 C to680 C, which includes the indium, tin,zinc, and aluminum fixed points, all inone furnace.

The 9114 furnace has an inlet for useof clean dry air or inert gas to initiate the

supercool of a tin cell. Other furnacesrequire the user to remove the hot andfragile tin cell from the furnace by handbefore cooling. In a Hart furnace, you sim-ply turn on your gas, monitor your cellduring its supercool, and turn the gas offwhen the freeze begins.

The 9114 is a three-zone furnace withthe best in Hart digital controller technol-ogy. Hart designs and builds proprietarycontrollers that have a reputation of beingthe best in the business. All of our fixed-point furnaces use them to achieve excel-lent stability and uniformity.

For easy access and visibility, all threezones are controlled from the top of theunit. The primary controller can be set in0.01 C increments, and actual tempera-ture is readable to two decimal places.

The freezing and melting process can beautomated using eight preset, user-programmable temperature settings. Thetop and bottom zones are slaved to the pri-mary zone using differential thermocouples.A high-temperature PRT acts as the maincontrol sensor for the best accuracy, sensi-tivity, and repeatability.

9115AThe 9115A Sodium Heat Pipe Furnace isspecifically designed for maintenance ofaluminum and silver freeze-point cells.

It has a temperature range of 550 Cto 1000 C with gradients of less than 0.1 C throughout. The sodium heat-pipe design provides a simple, yet uni-form, single heating zone that ensuresvery uniform changes in states duringheating and cooling.

Melting, freeze initiation, and plateaucontrol for a variety of freeze-point cells arepossible by entering up to eight set-points,ramp rates, and soak times. The controllerdisplays temperature in degrees C or F, andtemperature feedback is done via a ther-mocouple. Freeze-point plateaus of 8 to 10hours