International comparison of measurement of 200 kV DC, Results obtained of SP.

(1)

Anders Bergman

Measurement Technology SP Report 2008:26

(2)

(3)

International comparison of

measurement of 200 kV DC, Results

obtained by SP

(4)

Abstract

International comparison of measurement of 200 kV DC,

Results obtained by SP

SP has participated in international comparison EURAMET.EM- S29 of measurement of DC voltage up to 200 kV. The availability of another measuring system for high voltage DC has made it possible to investigate the measuring systems at SP in more detail and to identify refinements in evaluations of the measurements. It is expected that result of the intercomparison will permit new and reduced uncertainty levels for SP high voltage DC measurements.

Key words: High voltage DC

SP Sveriges Tekniska Forskningsinstitut SP Technical Research Institute of Sweden SP Report 2008:26

ISBN 978-91-85829-43-9 ISSN 0284-5172

(5)

Innehållsförteckning / Contents

Abstract 4

Innehållsförteckning / Contents

5 Preface 7

Summary 8

1 Instrumentation used at SP

9

2 Traceability 10

3 Travelling Measuring System

11

3.1 General 11

3.2 Description 11

4 Measurement setup

11

5 Uncertainty budget

12

5.1 Model function 12

5.2 Uncertainty contributions to measurement of divider output 13

5.3 Uncertainty contributions to scale factor evaluation 14

6 Results 14

6.1 General 14

6.2 Test object definition 15

6.3 Short term stability 15

6.3.1 Procedure 15

6.3.2 Short term stability of Vishay 50 kV divider 17

6.3.3 Short term stability of Vishay 150 kV divider 18

6.3.4 Short term stability of Vishay 200 kV 21

6.3.5 Short term stability of LCOE Travelling Measuring System 22

6.4 Summary of intercomparison results 24

6.5 Results at 1 kV 24 6.6 Results at 10 kV 26 6.7 Results at 50 kV 28 6.8 Results at 100 kV 30 6.9 Results at 150 kV 32 6.10 Results at 200 kV 34

7 Investigations of SP reference dividers

36

7.1 General 36

7.2 Humidity dependence of Vishay resistors 37

7.3 Long term stability SP 50 kV system 39

7.6 Input resistance of DVMs 40

8 Conclusions 42

(6)

(7)

Preface

Measurement of high DC voltages is important in several areas of society. Energy levels of medical X-ray equipment depend on applied DC voltage in the range 50 to 100 kV, and the energy level in its turn governs the resolution and contrast of X-ray records. Proper calibration of DC voltage is important in this area.

Energy transmission with direct voltage has been used since 1954 with the inauguration of the first Mainland Sweden - Gotland inter-tie. Today DC inter-ties are being built world-wide where large-scale bulk transmission of electrical energy is desired. Calibration of high DC voltage is needed in this area to ensure that protection systems operate with proper measurement of dangerous voltage. There is also a growing interest in energy billing based on DC side measurements, with subsequent need for proper calibration.

To support international trade, the European organisation EURAMET has decided to recognise a supplementary international comparison EURAMET.EM- S29, with comparison of DC voltage up to 200 kV. The pilot laboratory is Laboratorio Central Oficial de Eletrotecnica (LCOE), Spain. A total of 6 laboratories participate in the intercomparison.

(8)

Summary

Measurement of high DC voltages is usually made with a combination of resistive voltage dividers and voltage measuring instruments. In the present work, a comparison of

complete measuring systems has been made. In each comparison, both dividers have been connected to the same voltage source and their output voltages have been measured using precision multimeters. Foremost in the work has been attention to all extraneous effects that could influence the measurement result in order to reduce uncertainty limits as far as reasonably possible. The dominating uncertainty contribution is in most case the observed long term stability of SP’s reference dividers. The cause has been identified and work is in progress to rectify the situation and it is expected to be possible to lower uncertainty estimates in the foreseeable future.

(9)

1 Instrumentation used at SP

For high voltage DC measurements, SP has three reference dividers of type Vishay-Mann. Their rated operating voltage is 50, 150 and 200 kV respectively. Output voltage of the divider is measured with one of several available precision multimeters type HP3458A, all with options 001 and 002 (long memory and high stability). Measurement is normally done with software, to avoid misreading.

Vishay 50 and 100 kV Vishay 150 kV Vishay 200 kV

Figure 1. SP voltage dividers

The 50 kV divider was acquired 1992, the 150 kV divider in 1995 as a 100 kV device and augmented with new section in order to reach 150 kV in 1997. The 200 kV divider was acquired from NPL after their cessation of high voltage work and it arrived at SP in 2006. Unfortunately, the low voltage arm of the 200 kV divider proved to be unstable and a new one was built in 2006 and taken into use beginning of 2007.

The HP3458A’s have been at SP for more than 10 years. High voltage used in the tests was generated with

• a dual polarity DC supply made by Hipotronics, with a maximum thermal limit of 10 mA at 500 kV. The output of the generator is electronically stabilised with a typical short-term stability of 0,02%. Ripple is on the order of 100 V regardless of applied voltage;

• Rack-mounted dual polarity DC supply made by FUG for 35 kV stabilised voltage; • Calibrator FLUKE 5500 for DC voltage up to 1 kV.

Instrument list:

SP501616, High voltage divider Vishay S9909, nr 91039, 50 kV SP603267, High voltage divider Vishay S9986, 150 kV

SP603187, High voltage divider Vishay-Mann SN 94055, 200 kV SP502401, Multimeter HP 3458A

SP502935, Multimeter HP 3458A

Table 1. Scale factor of Vishay 50 kV divider as given in last calibration certificate (2008) Output Nominal ratio Assigned scale factor Expanded uncertainty

/ 10-6 A 500:1 500.0161 - B 1 000:1 1000.006 7 C 1 500:1 1500.024 7 D 2 500:1 2500.001 7 E 3 000:1 3000.010 7 F 5 000:1 5000.050 7

(10)

Output Nominal ratio Assigned scale factor Expanded uncertainty / 10-6 G 10 000:1 10000.10 7 H 15 000:1 15000.06 7 J 25 000:1 25000.26 7 K 30 000:1 30000.49 7 L 50 000:1 50000.67 7

/ 10-6

A 30 000:1 29999.43 10

B 45 000:1 44999.14 9

/ 10-6 B 1 000:1 1000.016 15 C 2 000:1 2000.012 15 D 10 000:1 10000.14 15 E 20 000:1 20000.10 15 F 200 000:1 200009.2 15

2 Traceability

SP maintains the unit for voltage based on realisation with the Josephson effect. The unit for resistance is realised with the von Klitzing effect. The DC dividers have been

calibrated annually with traceability to the realised standards for voltage and resistance. Uncertainty (k=2) of divider ratios has been between 7 and 15 ppm. The traceability chain is graphically depicted in Annex A, Traceability chart.

The calibrations of the HP3458A show that they fulfil their specifications for the high stability option throughout the years.

SP has also participated in other DC intercomparisons, DC voltage up to 100 kV

organised by Cigre 33-03 in 1995 [1], [2], [3] and DC voltage up to 100 kV organised by Euromet, EUROMET.EM-S14 in 1999 [4].

The intercomparison in 1995 is of special interest since the travelling standard used was a zener type reference that could be used to obtain a stand-alone realisation of 100 kV DC. The result of the intercomparison showed a small change in the scale factor with voltage, a change smaller than the estimated uncertainty. The change is on the order of 8 ppm from low voltage to 100 kV, i.e. approximately the same magnitude as the short term

(11)

stability obtained during this intercomparison, 6.3.3. No separate voltage dependence could be deduced.

3 Travelling Measuring System

3.1 General

The Travelling Measuring System consists of a shielded resistive divider with fixed input and grounding leads, a coaxial cable with a 1 MΩ terminating resistor, a digital

multimeter and a computer with a printer. Computer and printer were not utilised for the intercomparison at SP

Estimated uncertainty of the TRMS is ± 100 ppm.

3.2 Description

H.V lead: Description: Copper tube (length 2 m;

Ǿ = 28 mm).

Not used at SP Divider: Description: Resistive divider.

Manufacturer: ROSS ENGINEERING.

Type:

VD240-6Y-CBD-KC-BBC.

Serial Nº : 930729-5.

LCOE’s Reference: III-1-DT-003.

Nominal DC voltage: 240 kV.

Nominal ratio: 10 000 below 100 kV 10047 above 100 kV

Measuring cable: Description: Coaxial cable.

Type: RG-59/U.

LCOE’s Reference: III-3-CABL-052.

Length: 10 m.

Resistor: Description: Terminating resistor.

Nominal value: 1 MΩ.

LCOE’s Reference: III-3-CONC-002.

Measuring cable: Description: Coaxial cable. Not used at SP

Type: RG-59/U.

LCOE’s Reference: III-3-CABL-067.

Length: 0.5 m

Voltmeter: Manufacturer: Hewlett-Packard.

Type: 3458A.

Serial Nº : 2823 A 18964.

LCOE’s Reference: III-1-MD-013.

4 Measurement setup

All equipment was erected in SP high voltage laboratory. The laboratory is 12 m by 12 m with a ceiling at 7 m. The area is climatised with heating and cooling systems that maintain a temperature of 22 ±1 C. No humidity control is implemented.

The concrete floor is screened by sturdy reinforcement nets, welded together to form a continuous screen. At the walls this screen is connected to the sheet metal screen on the wall, which in turn also connects to the sheet metal screen in the ceiling.

For safety reasons, all high voltage equipment is interlocked with the entrance doors to prevent unauthorised entrance to the hall while high voltage supplies are energised. A by-pass feature is available, where an operator can enter the energised hall, if supervised by a colleague operating a dead-man switch.

Further in the interest in safety, all measurement cables are earthed at the laboratory screen wall before entering the control room. SP’s DC dividers are arranged such that the measurement is always taken directly across the low voltage arm. For the Travelling

(12)

Measuring System from LCOE, this is not possible, but separate measurements have been performed to investigate the influence of voltage difference between the voltage divider base and the local earth in the control room. For all high voltage work with the Travelling Measuring System, this has then been included as a type B uncertainty component. All equipment was in place in the laboratory at least 2 days before commencement of measurements.

All HP3458A were powered 24 hours per day and were subjected to Autocal DC every morning before measurements. Function Math Null was replaced by measurement of actual offset voltages and numerical corrections in evaluation spreadsheets.

5 Uncertainty budget

5.1 Model function

The scale factor SF2 of the device under test is calculated from

1 2 2 1 1 2 SF_DIV Offset DVM Offset DVM SF ⋅ − − = ,

where symbols are defined below.

A more complete model equation including corrections as needed for uncertainty evaluation is

(

)

(

)

1 2 _ 1 _ 1 _ 1 _ 1 _ 1 2 2 1 1 1 1 1 1 2 2 2 2 1 2 1 1 1 1 1 1 1 1 2 DIV imp input Out rated st DIV amb DIV volt DIV lt DIV cal DIV gnd drift Offset Offset range range rdg Out Bias drift Offset Offset range range rdg SF DVM DIV V V Offset DVM DVM DVM DVM DIV DVM Offset DVM DVM DVM DVM SF ⋅ ⎟⎟ ⎟ ⎠ ⎞ ⎜⎜ ⎜ ⎝ ⎛ + + ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ⋅ + + + + + ⋅ ⋅ + + + − ⋅ + + ⋅ ⋅ + + + − ⋅ + + ⋅ = δ δ δ δ δ δ δ δ δ δ δ δ δ δ δ

where index 1 refers to the reference measuring system and index 2 to the Travelling Measuring System and

• DVM1, DVM2, is the voltage reading from the multimeter;

• δDVM1rdg , δDVM2rdg , is the voltage correction due to DVM stability in terms of % of reading;

• DVM1range , DVM2range ,is the voltage range that the multimeter is set to;

• δDVM1range , δDVM2range , is the voltage correction due to DVM stability in terms of % of voltage range;

• Offset1, Offset2, is the average offset voltage determined with the multimeter connected to the respective divider;

• δoffset1 , δoffset2 , is the voltage correction due to drift of offset voltage during the measurements;

• DVM1bias is the input bias current of the multimeter connected to the reference divider;

• DIV1out is the output resistance of the reference voltage divider low voltage tap; • δgnd is the voltage correction due to potential difference between two earth points for

the Travelling Reference Divider; • V is the applied high voltage;

• V1rated is the rated voltage of the reference divider;

• DVM1input is the input resistance of the multimeter connected to the reference voltage divider, see also 7.6;

• DVM input impedance on SP divider ratio;

(13)

• δDIV1_cal is the relative scale factor correction of SP divider obtained from calibration and due to calibration uncertainty;

• δDIV1_lt is the relative scale factor correction of SP divider due to long term stability of the divider;

• δDIV1_volt is the relative scale factor correction of SP divider due to voltage induced effects on the divider, excluding heating effects;

• δDIV1_amb is the relative scale factor correction of SP divider due to ambient temperature;

• δDIV1_st is the relative scale factor correction of SP divider due to temperature increase caused by applied voltage, see also evaluation of short term stability in 6.3;

• δimp is the correction due to uncertainty in calculated influence from non-infinite. Influence of DVM bias current DVM2bias and input impedance DVM2input was disregarded for the Travelling Measuring System. See also 6.2

An uncertainty budget has been set up for each measurement and is shown in result tables. A general background on the estimation of the contributions is given here. Up to three DC voltage measuring systems could be measured simultaneously with automatic logging of measurements. Before each series of measurements, one or more sets of measurement of offset voltage was performed. A similar measurement of offset was performed after completion of the series of measurements. The offset value used was the mean of the offset before and the offset after the measurements.

5.2 Uncertainty contributions to measurement of

divider output

For each voltage channel the following contributions to uncertainty have been considered. • δDVM1range , δDVM2range . Stability of measurement estimated from DVM

specification, as % of range used and with an expectation value of zero. The

specification was considered as a rectangular distribution with half-width equal to the specification;

• δDVM1rdg , δDVM2rdg . The stability of measurement estimated from DVM specification, as % of reading and with an expectation value of zero. The

specification was considered as a rectangular distribution with half-width equal to the specification;

• DVM1bias. Influence of parasitic bias current in DVM input on voltage measurement. The uncertainty is taken as the assumed bias current DVM1bias multiplied by the output resistance DIV1out of the voltage divider. The estimate was considered having an expectation value of zero and a rectangular distribution;

• δoffset1 , δoffset2 . Mean value of standard deviations of the offset determinations before and after the measurements The estimate was considered having a normal

distribution;

• δoffset1drift , δoffset2 drift , Drift in offset voltage during measurements and estimated from measurement of offset before and after high voltage measurements, respectively. The estimate was considered having an expectation value of zero and a rectangular distribution ;

• δgnd. For the Travelling Measuring System, a contribution was also considered based on the observed voltage difference δgnd between earth point in the high voltage laboratory and the instrument earth point. The investigation was performed initially and showed that the difference was within ± 1 µV for any applied high voltage. This value was taken as the uncertainty estimate with an expectation value of zero and with a rectangular distribution.

(14)

The evaluation of voltage measurement uncertainty is first done absolute – i.e. the uncertainty has dimension [V]. This is then converted to relative uncertainty – i.e [V/V] – for further calculations of the uncertainty of the high voltage comparison.

5.3 Uncertainty contributions to scale factor

evaluation

Evaluation of scale factor was performed for each individual measurement in each series in order to ensure that the stability of voltage source does not contribute unnecessarily to uncertainty. The average value of scale factor was then calculated from this set of scale factors. For this reason the uncertainty budgets given later do not exhibit the conventional calculation of scale factor but instead provides the average value directly in the line “Exp std dev of the mean”.

For calculation of scale factor of the Travelling Measuring System the following contributions to uncertainty have been considered.

• δDIV1_cal. SP reference divider calibration uncertainty δDIV1_cal as stated in the latest calibration certificate, see 9, with expectation value zero and normal distribution; • δDIV1_lt. SP reference divider long term stability δDIV1_lt. Expectation value and

distribution are given in 7.3, 7.4 and 7.5;

• δDIV1_volt. SP reference divider voltage dependence δDIV1_volt. The divider resistors are wire-wound resistors with very high stability versus applied voltage. An estimate with expectation value zero and a rectangular distribution with a half-width of 1 ppm, has been chosen and is also supported by results of intercomparison in 1995; • δDIV1_amb. SP reference divider ambient temperature dependence δDIV1_amb. The

resistors of the divider have been carefully matched to provide a tracking of better than 1.5 ppm/K. An estimate of no more than ± 1 K temperature change leads to an estimate with expectation value zero and rectangular distribution with a half-width of 2 ppm;

• δDIV1_st. SP reference divider short term stability δDIV1_st. Expectation value and distribution are given in 6.3.2, 6.3.3 and 6.3.4;

• δimp. Influence of non-infinite input resistance DVM1input of DVM on ratio of SP Reference Divider. The assigned scale factor of the divider was corrected by an amount equal to the output resistance DIV1out of the divider divided by the estimated input resistance DVM1input of the DVM. The uncertainty estimate δimp was

considered to have expectation value zero and having a rectangular distribution with half-width equal to same value as the correction. Input resistance of DVM is based on measured data, see 7.6.

• Relative uncertainty of voltage measurement on the divider output of SP Reference Divider, as described in 5.2;

• Relative uncertainty of voltage measurement on the divider output of Travelling Measuring System, as described in 5.2;

• Experimental standard deviation of the mean of the scale factor.

Influence of non-infinite input resistance of DVM on the Travelling Divider is not considered, because divider and DVM are viewed as the Device Under Test;

6 Results

6.1 General

Test protocol demanded that the scale factor of the Travelling Measuring System be established at 1, 10, 50, 100, 150 and 200 kV. The short term stability should be established both for the Travelling Measuring System and for the local measuring systems. All measurements were suggested to be made in groups of 10 measurements.

(15)

Early investigation of the properties of the software provided by the pilot laboratory for the intercomparison gave an unnecessarily large standard deviation of the measured scale factor. Using SP software “Read 3 DVM” with NPLC set to 40 and flag LFREQ=LINE, proved that the standard deviation could be reduced to practically zero. All measurements by SP have therefore been performed with Read 3 DVM. The functionality of the

program makes it possible to log measurements continuously and simultaneously. Therefore the grouping of measurements into groups of 10 was dispensed with. This feature is especially useful for the short-term stability test.

In all measurements, instrument offset before and after each set of measurements was determined by taking a set of 50 measurements spaced 3 to 5 seconds apart.

Tests were performed during February, 2008.

Time separation between measurements at different voltage levels was at least several hours, except for 1 and 10 kV levels where no significant heating effects were

anticipated.

6.2 Test object definition

The measuring instrument Agilent3458A is part of the Travelling Measuring System, and its stated uncertainties according to Agilent manual could be regarded as out of scope for the uncertainty analysis. SP has however chosen to include the instrument uncertainties related to specification of % of range and % of reading in the uncertainty budget, since it cannot be assumed that the corresponding measurement errors would be constant over a time span relevant for the circulation.

6.3 Short term stability

6.3.1 Procedure

Short term stability is defined in the present draft for the revision of IEC 60060-2 as the change in scale factor from the time instant of application of the high voltage and for a duration valid for the measuring system, usually 1 hour. For the present intercomparison, SP has decided to interpret this as the time until stable readings are achieved. Other “short term stability” effects of random nature are disregarded here since they will show up in the experimental standard deviation.

For a resistive voltage divider, the main mechanism of short-term stability is recognised as the heating of the resistor column. The effect is quadratic with applied voltage, but may also have a rather large time constant.

In practice the evaluation of the short time stability has been performed such that one divider has been connected to high voltage for an extended period, long enough to ensure that it has reached equilibrium, before the test object has been connected. This pre-heating time has been chosen to be in the order of hours. The other divider had been disconnected form high voltage for at least several hours before the test.

To ensure that the reference divider was under stable conditions, care was taken to avoid voltage interruption when connecting the test object. When thermal equilibrium had been reached for the reference divider, a pre-prepared high voltage bus was lifted from the divider under voltage and connected to the divider to be tested. This operation was performed with an insulated stick and under full voltage. The output voltage of both dividers were logged every 5 seconds starting at the instant of connection. A decimation of the number of data points has been performed by calculating the average value of each

(16)

10 measurements in a series, thus ensuring compatibility with measurements performed by other participants.

All voltage dividers, including the Travelling Reference Divider, were in the laboratory environment for several days before measurements were performed.

In the investigation of short-term stability, nominal scale factors were used for the dividers, without attempts to further correct for known deviations. This is acceptable since only the relative change between the stabilised divider and the divider under test is of interest in this test.

The result of the short term stability investigation can be interpreted different ways, the first and most practical is to understand the change from cold to thermal equilibrium as a measurement uncertainty. This is often the case when the time duration of voltage application to the divider before measurement is not clearly specified. The other interpretation is the non-stability after thermal equilibrium has been reached. This interpretation is of interest in this intercomparison where the fully stabilised scale factors of the measuring systems can be compared.

In this intercomparison all measurements have been performed after thermal equilibrium had been reached.

Lifting HV bus from divider under voltage.

Connecting cold divider to high voltage

(17)

6.3.2 Short term stability of Vishay 50 kV divider

Short term stability at 50 kV

10000.19 10000.20 10000.21 10000.22 10000.23 10000.24 10000.25 10000.26 10000.27 0:00:00 0:14:24 0:28:48 0:43:12 0:57:36 1:12:00 1:26:24 1:40:48 1:55:12 2:09:36 2:24:00 S cale f acto r 0 1 2 3 4 5 St anda rd deviat ion [pp m ] Vishay050 +3 ppm -3 ppm Stdev

Figure 3. Scale factor of Vishay 50 kV divider over a period of 2 hours, at 50 kV. Reference is Vishay 200 kV divider. Each data point shown is the mean of 10 consecutive

measurements. The standard deviation of the mean for each data point is given by the lower curve. The straight lines show ± 3 ppm deviation from the average scale factor after thermal equilibrium had been reached.

The Vishay 200 kV divider had been allowed to stabilise at 50 kV for a period of 1 hour 15 min before the 50 kV divider was connected.

The step change observed in Figure 3 at approximately 1 h 30 min, has later been traced to the low voltage arm of Vishay 200 kV (rebuilt low voltage arm with resistors type Caddock USF240/270). It is thought that this phenomenon is a random noise originating in the resistive material of the Caddock resistors. No such behaviour has been noted for the Vishay resistors. Compare with Figure 6 showing results for Vishay 150 kV divider, which was measured simultaneously.

The reference divider has been used as stable comparison device only, and its scale factor has been taken as the nominal scale factor given by its calibration certificate and therefore the cited values for scale factor of the tested divider are valid only for the changes

occurring during the test.

Short term stability evaluation for Vishay 50 kV divider Time [min] Variation half-width at 50 kV [ppm] 0-120 5 10-120 1

Drift from start to steady state At 50 kV

[ppm] 4

(18)

For cases when voltage application time is not controlled, the scale factor SFcal is taken

from the calibration certificate, and the uncertainty contribution ust due to short-term

stability is estimated to be a rectangular distribution with 5 ppm half-width at rated voltage. The uncertainty contribution is scaled with the square of the applied voltage V divided by the rated voltage,

3 1 10 50 10 5 2 3 6 _⋅ ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ⋅ ⋅ ⋅ = − V u_st .

For cases when voltage application time is more than 10 minutes before measurements, the scale factor SF1V of Vishay 50 kV divider at voltage V is is calculated as

(

)

2 3 6 10 50 10 5 1 1 1 _⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ⋅ ⋅ ⋅ + ⋅ =SF − V

SF V cal , where SF1cal is the scale factor obtained at low

voltage calibration. Uncertainty contribution ust due to short-term stability is in this case

estimated to 3 1 10 50 3 10 5 2 3 6 ⋅ ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ⋅ ⋅ ⋅ = − V

u_st . The distribution is taken as rectangular.

6.3.3 Short term stability of Vishay 150 kV divider

Short term stability Vishay150 at 200 kV

29998.80 29999.00 29999.20 29999.40 29999.60 29999.80 30000.00 0:00:00 1:12:00 2:24:00 3:36:00 4:48:00 6:00:00 7:12:00 8:24:00 Scale f acto r 0 1 2 3 4 5 Stan dard deviatio n [p p m ] vs SP 200 +3 ppm -3 ppm Stdev

Figure 4. Scale factor of Vishay 150 kV at 200 kV for a period of 7 hours. Reference is Vishay 200 kV. Each data point shown is the mean of 10 consecutive measurements. The standard deviation of the mean for each data point is given by the lower curve. The straight lines show ± 3 ppm deviation from the average scale factor after thermal equilibrium had been reached.

The Vishay 200 kV had been allowed to stabilise at 200 kV for a period of 9 hours before the 150 kV divider was connected.

The Vishay 150 kV divider was intentionally driven at 133 % of rated voltage to investigate the possibility to extend its useful range. The results prove that it can be operated at this voltage with only very slight degradation of stability.

(19)

Short term stability at 100 kV 29998.90 29998.95 29999.00 29999.05 29999.10 29999.15 29999.20 29999.25 29999.30 29999.35 29999.40 0:00:00 0:28:48 0:57:36 1:26:24 1:55:12 2:24:00 2:52:48 Scale f acto r 0 1 2 3 4 5 Stan dard deviatio n [p p m ] Vishay150 +3 ppm -3 ppm Stdev

Figure 5. Scale factor of Vishay 150 kV at 100 kV for a period of 2 hours 30 min. Reference is Vishay 200 kV. Each data point shown is the mean of 10 consecutive measurements. The standard deviation of the mean for each data point is given by the lower curve. The straight lines show ± 3 ppm deviation from the average scale factor after thermal equilibrium had been reached.

The Vishay 200 kV had been allowed to stabilise at 100 kV for a period of 1 hour 45 minutes before the 150 kV divider was connected.

Short term stability at 50 kV

29999.02 29999.04 29999.06 29999.08 29999.10 29999.12 29999.14 29999.16 29999.18 29999.20 29999.22 29999.24 0:00:00 0:14:24 0:28:48 0:43:12 0:57:36 1:12:00 1:26:24 1:40:48 1:55:12 2:09:36 2:24:00 S cale f acto r 0 1 2 3 4 5 St anda rd deviat ion [pp m ] Vishay150 +3 ppm -3 ppm Stdev

Figure 6. Scale factor of Vishay 150 kV divider over a period of 2 hours, at 50 kV. Reference is Vishay 200 kV. Each data point shown is the mean of 10 consecutive measurements. The standard deviation of the mean for each data point is given by the lower curve. The straight lines show ± 3 ppm deviation from the average scale factor after thermal equilibrium had been reached.

The Vishay 200 kV had been allowed to stabilise at 50 kV for a period of 1 hour 15 min before the 150 kV divider was connected.

(20)

The step change observed in Figure 6, has later been traced to the low voltage arm of Vishay 200 kV (rebuilt low voltage arm with resistors type Caddock USF240/270). It is thought that this phenomenon is a random noise originating in the resistive material of the Caddock resistors. No such behaviour has been noted for the Vishay resistors.

The reference divider has been used as stable comparison device only, and the its scale factor has been taken as the nominal scale factor given by its calibration certificate and therefore the cited values for scale factor of the tested divider is valid only for the change occurring during the test.

Short term stability evaluation for Vishay 150 kV divider Time [min] Variation half-width at 50 kV [ppm] Variation half-width at 100 kV [ppm] Variation half-width at 200 kV [ppm] 0-420 2 10 28 10-420 1 4 17 30-420 1 1.5 3

[ppm] At 100 kV [ppm] At 200 kV [ppm]

2 10 28

stability is estimated to be a rectangular distribution with 17 ppm half-width at rated voltage (150 kV). The contribution is scaled with the square of the applied voltage V divided by the rated voltage,

3 1 10 150 10 17 2 3 6 _⋅ ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ⋅ ⋅ ⋅ = − V u_st .

For cases when voltage application time is more than 30 minutes before measurements, the scale factor SF1V of Vishay 150 kV divider at voltage V is calculated as

(

)

2 3 6 10 150 10 17 1 1 1 _⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ⋅ ⋅ ⋅ + ⋅ =SF − V

SF _V _cal , where SF1cal is the scale factor obtained at

low voltage calibration. Uncertainty contribution ust due to short-term stability is in this

case estimated to 3 1 10 150 3 10 17 2 3 6 ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ⋅ ⋅ ⋅ = − V u_st .

(21)

6.3.4 Short term stability of Vishay 200 kV

Short term stability Vishay 200 at 200 kV

19999.70 19999.75 19999.80 19999.85 19999.90 19999.95 20000.00 0:00:00 0:28:48 0:57:36 1:26:24 1:55:12 2:24:00 2:52:48 3:21:36 3:50:24 4:19:12 Scale f acto r 0 1 2 3 4 5 Stan dard deviatio n [p p m ] vs SP 150 kV +3 ppm -3 ppm Stdev

Figure 7. Scale factor of Vishay 200 kV at 200 kV for a period of 7 hours. Reference is Vishay 150 kV. Each data point shown is the mean of 10 consecutive measurements. The standard deviation of the mean for each data point is given by the lower curve. The straight lines show ± 3 ppm deviation from the average scale factor after thermal equilibrium had been reached.

The Vishay 150 kV had been allowed to stabilise at 200 kV for a period of 3 hours min before the 200 kV divider was connected.

The reference divider has been used as stable comparison device only, and the its scale factor has been taken as the nominal scale factor given by its calibration certificate and therefore the cited values for scale factor of the tested divider are valid only for the changes occurring during the test.

Short term stability evaluation for Vishay 200 kV divider Time [min] Variation half-width at 200 kV [ppm] 0-240 10 30-240 3 45-420 2

[ppm] 10

(22)

(200 kV). The contribution is scaled with the square of the applied voltage V divided by the rated voltage,

3 1 103 200 10 10 2 6 _⎟ ⎠ ⎞ ⎜ ⎝ ⎛ ⋅ ⋅ ⋅ = − V u_st .

For cases when application time is more than 10 minutes before measurements, the scale factor SF1V of Vishay 200 kV divider at voltage V is calculated as

(

)

2 3 6 10 200 10 10 1 1 1 _⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ⋅ ⋅ ⋅ + ⋅ =SF − V

SF V cal , where SF1cal is the scale factor obtained at

low voltage calibration. Uncertainty contribution ust due to short-term stability is in this

case estimated to 3 1 10 200 3 10 10 2 3 6 ⋅ ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ ⋅ ⋅ ⋅ = − V u_st .

6.3.5 Short term stability of LCOE Travelling Measuring

System

LCOE short term stability at 200 kV

10047.60 10047.70 10047.80 10047.90 10048.00 10048.10 10048.20 10048.30 10048.40 10048.50 10048.60 10048.70 0:00:00 0:28:48 0:57:36 1:26:24 1:55:12 2:24:00 2:52:48 3:21:36 3:50:24 4:19:12 Scale f acto r 0 1 2 3 4 5 Stan dard deviatio n [p p m ] vs SP200 +3 ppm -3 ppm Stdev

Figure 8. Scale factor of LCOE Travelling Measuring System at 200 kV for a period of 4 hours. Reference is Vishay 200 kV divider. Each data point shown is the mean of 10 consecutive measurements. The standard deviation of the mean for each data point is given by the lower curve. The straight lines show ± 3 ppm deviation from the average scale factor after thermal equilibrium had been reached.

The Vishay 200 kV had been allowed to stabilise at 200 kV for a period of 1 hour 30 min before the LCOE Travelling Measuring System was connected.

(23)

Short term stability at 100 kV 9999.90 10000.00 10000.10 10000.20 10000.30 10000.40 10000.50 10000.60 0:00:00 0:28:48 0:57:36 1:26:24 1:55:12 2:24:00 2:52:48 Scale f acto r 0 1 2 3 4 5 Stan dard deviatio n [p p m ] LCOE +3 ppm -3 ppm Stdev

Figure 9. Scale factor of LCOE Travelling Measuring System at 100 kV for a period of 2 hours 30 min. Reference is Vishay 200 kV divider. Each data point shown is the mean of 10 consecutive measurements. The standard deviation of the mean for each data point is given by the lower curve. The straight lines show ± 3 ppm deviation from the average scale factor after thermal equilibrium had been reached.

The Vishay 200 kV had been allowed to stabilise at 100 kV for a period of 1 hour 45 min before the LCOE Travelling Measuring System was connected.

The reference divider has been used as stable comparison device only, and the its scale factor has been taken as the nominal scale factor given by its calibration certificate and therefore the cited values for scale factor of the tested divider is valid only for the change occurring during the test.

Short term stability evaluation for LCOE Travelling Measuring System Time [min] Variation half-width at 100 kV [ppm] Variation half-width at 200 kV [ppm] 0-240 48 85 10-240 16 9 30-240 4 7

[ppm]

At 200 kV [ppm]

(24)

6.4 Summary of intercomparison results

Table 4. Results from intercomparison between Travelling Measuring System and SP reference measuring system and using measurement software SP Read3dvm. SP Measuring System used as reference system.

Voltage level [kV] Obtained Scale Factor of the TRMS combined standard uncertainty u[ppm] Effective degrees of

freedom νeff SP reference divider

+ 1 10000.05 11.5 65738 HP3458A + 10 10000.27 11 18 Vishay 150 + 50 10000.34 10.5 18 Vishay 150 + 100 10000.48 11 18 Vishay 150 + 150 10048.73 11 24 Vishay 150 + 200 10048.78 12.5 33 Vishay 150 - 1 9999.99 11.5 65733 HP3458A - 10 10000.29 11 18 Vishay 150 - 50 10000.35 10.5 18 Vishay 150 - 100 10000.49 11 19 Vishay 150 - 150 10048.69 11 24 Vishay 150 - 200 10048.71 12.5 33 Vishay 150

6.5 Results at 1 kV

The Travelling Measuring System was supplied by a Fluke 5500 as source and the applied voltage measured with a HP3458. No waiting time was applied before measurements.

In total 3 sets of 50 measurements were taken on each polarity. All three sets were combined for each polarity to obtain the scale factor.

The dominating contribution to uncertainty stems from uncertainty in Travelling Measuring System low voltage arm ground reference point, see above, followed by uncertainty in 1000 V measurement.

The recorded ambient temperature was between 21.7 and 22.4 ºC and the relative humidity was 28 %.

Table 5. Scale factor of travelling reference (LCOE) versus SP HP3458 at 1 kV Polarity Measured scale factor Expanded uncertainty

/ 10-6

Positive 10000.05 23

Negative 9999.99 23

The difference between positive and negative polarity is ¼ of the uncertainty and is therefore of doubtful relevance but it is hypothesised that this effect is due to the input bias current of the multimeter. The bias current of Agilent 3458 can be estimated to a probable value of 5 pA. Combined with the output resistance of the travelling reference

(25)

divider – 48 kΩ – an influence of 2.5 ppm would be present. This difference is additive for one polarity and subtractive for the other. Thus the difference between polarities could be caused by the bias current. Since no measurement of bias current was made on the LCOE Agilent 3458A, there is at present no basis for vindication or falsification of the hypothesis.

Identification of measurement 1 kV 22.4 C, 28% RH

Uncertainty budget DVM1 voltage measurement on Direct

Quantity Value std uncertainty

degrees of

freedom sens. Coeff

contribution to std uncert DVM1 range 5.7735E-08 100000 1000.0 V 5.77E-05 V DVM1 rdg 1000.000 621 V 6.93E-03 V 100000 1 6.93E-03 V DVM1 bias current into Direct LV

arm 5.77E-12 A 100000 0 ohm 0.00E+00 V DVM1 offset determination -0.000 015 V 2.58E-06 V 98 1 2.58E-06 V DVM1 offset drift 7.01E-07 V 100000 1 7.01E-07 V Direct voltage absolute 1000.000 635 V 100014 6.93E-03 V Direct volt relative uncertainty 1 6.92844E-06

Uncertainty budget DVM2 voltage measurement on LCOE

DVM2 range 1.73205E-06 100000 0.1 V 1.73E-07 V DVM2 rdg 0.100 000 V 2.89E-07 V 100000 1 2.89E-07 V DVM1 bias current into LCOE

LV arm 5.77E-12 A 100000 0 ohm 0.00E+00 V DVM2 offset determination -0.000 001 V 2.26E-08 V 98 1 2.26E-08 V DVM2 offset drift 4.98E-07 V 100000 1 4.98E-07 V DVM2 Gnd potential difference 7.00E-07 V 10000 1 7.00E-07 V LCOE voltage absolute 0.100 000 V 29369 9.23E-07 V LCOE volt relative uncertainty 1 9.22936E-06

Relative uncertainty of scale factor of LCOE vs Direct

SP divider certificate 0 10000 1 0.000 000 0 SP divider long term stability 0 200 1 0.000 000 0 SP divider short term stability,

(voltage squared) 0 100000 1.000001241 0.000 000 0 SP divider voltage dependence 0 100000 1 0.000 000 0 SP divider ambient temp dep 0 100000 1 0.000 000 0 DVM1 input impedance

influence on Direct 1.000 000 0 5.77E-10 S 100000 0 ohm 0.000 000 0 DVM2 input impedance

inflluence on LCOE 2.89E-11 S 100000 0 ohm 0.000 000 0 DVM1 voltage 6.92844E-06 100014 1 0.000 006 9 DVM2 voltage 9.22936E-06 29369 1 0.000 009 2 Exp std dev of the mean 10000.05305 0.002737875 149 1 0.000 000 3 LCOE vs Direct 10000.05 ± 0.23 65738 1.15438E-05

Scale factor abs unc. rel unc. k

10000.05 ± 0.23 ± 0.0023 % 2.00

(26)

Identification of measurement negative 1 kV 21.7 C, 28% RH

Uncertainty budget DVM1 voltage measurement on Direct

degrees of

freedom sens. Coeff

contribution to std uncert DVM1 range 5.7735E-08 100000 1000.0 V 5.77E-05 V DVM1 rdg -1000.005 462 V -6.93E-03 V 100000 1 -6.93E-03 V DVM1 bias current into Direct LV

arm 5.77E-12 A 100000 0 ohm 0.00E+00 V DVM1 offset determination -0.000 015 V 2.58E-06 V 98 1 2.58E-06 V DVM1 offset drift 7.01E-07 V 100000 1 7.01E-07 V Direct voltage absolute -1000.005 447 V 100014 6.93E-03 V Direct volt relative uncertainty 1 6.92844E-06

DVM2 range 1.73205E-06 100000 0.1 V 1.73E-07 V DVM2 rdg -0.100 001 V -2.89E-07 V 100000 1 -2.89E-07 V DVM1 bias current into LCOE

LV arm 5.77E-12 A 100000 0 ohm 0.00E+00 V DVM2 offset determination -0.000 001 V 2.26E-08 V 98 1 2.26E-08 V DVM2 offset drift 4.98E-07 V 100000 1 4.98E-07 V DVM2 Gnd potential difference 7.00E-07 V 10000 1 7.00E-07 V LCOE voltage absolute -0.100 000 V 29369 9.23E-07 V LCOE volt relative uncertainty 1 9.22942E-06

Relative uncertainty of scale factor of LCOE vs Direct

SP divider certificate 0 10000 1 0.000 000 0 SP divider long term stability 0 200 1 0.000 000 0 SP divider short term stability,

(voltage squared) 0 100000 1.000010924 0.000 000 0 SP divider voltage dependence 0 100000 1 0.000 000 0 SP divider ambient temp dep 0 100000 1 0.000 000 0 DVM1 input impedance

influence on Direct 1.000 000 0 5.77E-10 S 100000 0 ohm 0.000 000 0 DVM2 input impedance

inflluence on LCOE 2.89E-11 S 100000 0 ohm 0.000 000 0 DVM1 voltage 6.92844E-06 100014 1 0.000 006 9 DVM2 voltage 9.22942E-06 29369 1 0.000 009 2 Exp std dev of the mean 9999.993758 0.002615231 149 1 0.000 000 3 LCOE vs Direct 9999.99 ± 0.23 65733 1.15436E-05

9999.99 ± 0.23 ± 0.0023 % 2.00

Figure 11. Example of uncertainty budget for 1 kV measurement, negative polarity.

6.6 Results at 10 kV

The Travelling Measuring System was connected in parallel to SP reference dividers Vishy200 and Vishay150 and they were supplied from a rack-size 35 kV stabilised DC supply. High voltage was applied for 1 hour 30 min at positive voltage before

measurements, followed by polarity change and another 25 minutes of stabilisation before measurements.

Due to problems with long term stability of Vishay 200 low voltage arm, results obtained with Vishay 200 had about 1.5 times larger uncertainty compared to those obtained with Vishay 150. Result summary is based on results obtained with Vishay 150 .

The scale factors of Vishay 150 and Vishay 200 have been corrected for the scale factor change due to self-heating. The background is given in 6.3

Measurements were performed using all available voltage taps on Vishay 200 kV divider. Evaluation of uncertainty indicates that using voltage taps with higher output voltage – and correspondingly higher output resistance – can lead to higher uncertainty due to influence of non-infinite input resistance (or non-zero conductance) of the multimeter. The recorded ambient temperature was 22.2 ºC and the relative humidity was 29 %.

(27)

Table 6. Scale factor of LCOE Travelling Measuring System with SP system as reference at 10 kV

Reference divider and

output Polarity Measured scale factor Expanded uncertainty

/ 10-6 Vishay 200 E Positive 10000.23 30 Vishay 200 D Positive 10000.23 30 Vishay 200 C Positive 10000.25 31 Vishay 200 B Positive 10000.28 37 Vishay 150 A Positive 10000.27 22 Vishay 200 E Negative 10000.08 30 Vishay 200 D Negative 10000.18 30 Vishay 200 C Negative 10000.29 31 Vishay 200 B Negative 10000.32 37 Vishay 150 A Negative 10000.29 22

Uncertainty budget DVM1 voltage measurement on Vishay200

degrees of

freedom sens. Coeff

contribution std uncert

DVM1 range 1.73205E-07 100000 1.0 V 1.73E-07 V DVM1 rdg 0.500 904 V 1.16E-06 V 100000 1 1.16E-06 V DVM1 bias current into

Vishay200 LV arm 5.77E-12 A 100000 100 000 ohm 5.77E-07 V DVM1 offset determination -0.000 004 V 6.60E-08 V 98 1 6.60E-08 V DVM1 offset drift 1.94E-06 V 100000 1 1.94E-06 V Vishay200 voltage absolute 0.500 908 V 186101 2.34E-06 V Vishay200 volt relative uncertainty 1 4.66573E-06

LV arm 5.77E-12 A 100000 0 ohm 0.00E+00 V DVM2 offset determination 0.000 000 V 2.69E-08 V 98 1 2.69E-08 V DVM2 offset drift 1.92E-07 V 100000 1 1.92E-07 V DVM2 Gnd potential difference 7.00E-07 V 10000 1 7.00E-07 V LCOE voltage absolute 1.001 790 V 112475 2.43E-06 V LCOE volt relative uncertainty 1 2.42667E-06

Relative uncertainty of scale factor of LCOE vs Vishay200

SP divider certificate 0.0000075 3254 1 0.000 007 5 SP divider long term stability 1.1547E-05 200 1 0.000 011 5 SP divider short term stability,

(voltage squared) 1.9245E-06 100000 0.002509072 0.000 000 0 SP divider voltage dependence 5.7735E-07 100000 1 0.000 000 6 SP divider ambient temp dep 1.1547E-06 100000 1 0.000 001 2 DVM1 input impedance

influence on Vishay200 1.000 000 6 5.77E-12 S 100000 100 000 ohm 0.000 000 6 DVM2 input impedance

influence on LCOE 5.77E-12 S 100000 0 ohm 0.000 000 0 DVM1 voltage 4.66573E-06 186101 1 0.000 004 7 DVM2 voltage 2.42667E-06 112475 1 0.000 002 4 Exp std dev of the mean 10000.22824 0.000351112 49 1 0.000 000 0 LCOE vs Vishay200 10000.23 ± 0.30 535 1.48068E-05

10000.23 ± 0.30 ± 0.0030 % 2.00

Figure 12. Example of uncertainty budget at 10 kV positive polarity, Travelling Measuring System versus Vishay 200 kV divider, output E.

(28)

Quantity Value std uncertainty degrees of freedom sens. Coeff

contribution std uncert DVM3 range 1.73205E-07 100000 1.0 V 1.73E-07 V DVM3 rdg 0.333 945 V 7.71E-07 V 100000 1 7.71E-07 V DVM1 bias current into

Vishay150 LV arm 5.77E-12 A 100000 50 001 ohm 2.89E-07 V DVM3 offset determination -0.000 002 V 1.17E-08 V 98 1 1.17E-08 V DVM3 offset drift 1.44E-08 V 100000 1 1.44E-08 V Vishay150 voltage absolute 0.333 947 V 138794 8.42E-07 V Vishay150 volt relative uncertai 1 2.5204E-06

LV arm 5.77E-12 A 100000 0 ohm 0.00E+00 V DVM2 offset determination 0.000 000 V 2.69E-08 V 98 1 2.69E-08 V DVM2 offset drift 1.92E-07 V 100000 1 1.92E-07 V DVM2 Gnd potential difference 7.00E-07 V 10000 1 7.00E-07 V LCOE voltage absolute 1.001 790 V 112475 2.43E-06 V LCOE volt relative uncertainty 1 2.4267E-06

SP divider certificate 0.000004575 2114 1 0.000 004 6 SP divider long term stability 0.000009 9 1 0.000 009 0 SP divider short term stability,

10000.27 ± 0.22 ± 0.0022 % 2.15

Figure 13. Example of uncertainty budget at 10 kV positive polarity, Travelling Measuring System versus Vishay 150 kV divider, output A.

6.7 Results at 50 kV

The Travelling Measuring System was connected in parallel to SP reference dividers Vishy200 and Vishay150 and they were supplied from a 500 kV stabilised DC supply. High voltage was applied for 1 hour 10 min at positive voltage before measurements, followed by polarity change and another 1 hour 45 minutes of stabilisation before measurements.

Due to problems with stability of Vishay 200 low voltage arm, results obtained with Vishay 200 had about 1.5 times larger uncertainty compared to those obtained with Vishay 150. Result summary is based on measurements using Vishay 150.

(29)

Table 7. Scale factor of LCOE Travelling Measuring System, with SP system as reference, at 50 kV

/ 10-6 Vishay 200 E Positive 10000.23 28 Vishay 200 D Positive 10000.27 29 Vishay 150 A Positive 10000.34 21 Vishay 200 E Negative 10000.23 28 Vishay 200 D Negative 10000.28 29 Vishay 150 A Negative 10000.35 21

degrees of

freedom sens. Coeff

Vishay200 LV arm 5.77E-12 A 100000 100 000 ohm 5.77E-07 V DVM1 offset determination -0.000 001 V 2.71E-08 V 98 1 2.71E-08 V DVM1 offset drift 2.93E-07 V 100000 1 2.93E-07 V Vishay200 voltage absolute 2.524 552 V 102971 5.87E-06 V Vishay200 volt relative uncertainty 1 2.32643E-06

10000.23 ± 0.28 ± 0.0028 % 2.01

Figure 14. Example of uncertainty budget at 50 kV positive polarity, Travelling Measuring System versus Vishay 200 kV divider, output E

(30)

10000.33 ± 0.21 ± 0.0021 % 2.15

Figure 15. Example of uncertainty budget at 50 kV positive polarity, Travelling Measuring System versus Vishay 150 kV divider, output A

6.8 Results at 100 kV

The Travelling Measuring System was connected in parallel to SP reference dividers Vishy200 and Vishay150 and they were supplied from a 500 kV stabilised DC supply. High voltage was applied for 1 hour 10 min at positive voltage before measurements, followed by polarity change and another 1 hour 45 minutes of stabilisation before measurements.

Due to problems with stability of Vishay 200 low voltage arm, results obtained with Vishay 200 had larger uncertainty compared to those obtained with Vishay 150. Result summary is based on results obtained with Vishay 150.

(31)

Table 8. Scale factor of LCOE Travelling Measuring system with SP systems as reference at 100 kV

/ 10-6 Vishay 200 E Positive 10000.39 28 Vishay 200 D Positive 10000.44 29 Vishay 150 A Positive 10000.48 22 Vishay 200 E Negative 10000.42 28 Vishay 200 D Negative 10000.45 29 Vishay 150 A Negative 10000.49 22

degrees of

freedom sens. Coeff

Vishay200 LV arm 5.77E-12 A 100000 100 000 ohm 5.77E-07 V DVM1 offset determination 0.000 000 V 3.13E-08 V 98 1 3.13E-08 V DVM1 offset drift 1.46E-07 V 100000 1 1.46E-07 V Vishay200 voltage absolute 4.996 128 V 100660 1.16E-05 V Vishay200 volt relative uncertainty 1 2.31321E-06

10000.39 ± 0.28 ± 0.0028 % 2.01

(32)

contribution std uncert DVM3 range 2.88675E-08 100000 10.0 V 2.89E-07 V DVM3 rdg -3.330 846 V -7.69E-06 V 100000 1 -7.69E-06 V DVM1 bias current into

Vishay150 LV arm 5.77E-12 A 100000 50 001 ohm 2.89E-07 V DVM3 offset determination -0.000 001 V 1.74E-07 V 98 1 1.74E-07 V DVM3 offset drift 7.83E-08 V 100000 1 7.83E-08 V Vishay150 voltage absolute -3.330 845 V 100660 7.71E-06 V Vishay150 volt relative uncertai 1 2.3134E-06

10000.49 ± 0.22 ± 0.0022 % 2.15

6.9 Results at 150 kV

The Travelling Measuring System was connected in parallel to SP reference dividers Vishy200 and Vishay150 and they were supplied from a 500 kV stabilised DC supply. High voltage was applied for 50 min at positive voltage before measurements, followed by polarity change and another 30 minutes of stabilisation before measurements. Due to problems with stability of Vishay 200 low voltage arm, results obtained with Vishay 200 had larger uncertainty compared to those obtained with Vishay 150. Result summary is based on results obtained with Vishay 150 kV.

The ambient temperature was between 21.6 and 22.4 ºC and the relative humidity was 23 %.

(33)

Table 9. Scale factor of LCOE Travelling Measuring System with SP systems at as reference, 150 kV

/ 10-6

Vishay 200 E Positive 10048.64 29

Vishay 150 A Positive 10048.73 23

Vishay 200 E Negative 10000.62 29

Vishay 150 A Negative 10000.69 23

degrees of

freedom sens. Coeff

10048.64 ± 0.29 ± 0.0029 % 2.01

(34)

10048.73 ± 0.23 ± 0.0023 % 2.11

6.10 Results at 200 kV

The Travelling Measuring System was connected in parallel to SP reference dividers Vishy200 and Vishay150 and they were supplied from a 500 kV stabilised DC supply. High voltage was applied for 45 min at positive voltage before measurements, followed by polarity change and another 30 minutes of stabilisation before measurements. Due to problems with stability of Vishay 200 low voltage arm, results obtained with Vishay 200 had larger uncertainty compared to those obtained with Vishay 150. Result summary is based on results obtained with Vishay 150 kV.

The ambient temperature was between 22.1 and 22.6 ºC and the relative humidity was 26 %.

(35)

Table 10. Scale factor of LCOE Travelling Measuring System with SP systems as reference, at 200 kV

/ 10-6

Vishay 200 E Positive 10048.68 29

Vishay 150 A Positive 10048.78 25

Vishay 200 E Negative 10048.61 29

Vishay 150 A Negative 10048.71 25

degrees of

freedom sens. Coeff

10048.68 ± 0.29 ± 0.0029 % 2.00