Transcript Transmille Training - Making Good Measurements

Training
Making accurate Electrical
Measurements
Introduction

Sources of errors in electrical
measurements

Understanding instrument specification

Getting the best from a Digital multimeter
and Calibrator
Making good connection

Large errors in measurements can be easily
caused by using the incorrect leads.

The key measurements areas where leads effect
accuracy are:Low DC voltage
High Value Resistance
Low AC current
Low resistance 4 wire connections
Measuring Low DC Voltage

When measuring DC voltage below 1volt the
main source’s of error are thermal generated
voltages (emf) in the lead connections. These
voltages which can easily be tens of micro volts
constantly change as temperature gradients
change.

Thermo electric voltages occur when ever 2
different metals are in contact with a temperature
gradient across the contact. This is how a
thermocouple works.
Reducing Thermal EMF 1

There are three ways to reduce EMF:-
Firstly and most important
Use connectors made with metals which produce low
thermal voltages, like gold, silver.
Avoid connectors made, or plated with materials like nickel,
which is often use to plate ‘banana’ type plugs. Note
nickel is used in ‘K type’ thermocouple and can produce
up to 33uV/C.
The terminals in precision instruments are usually made
with oxygen free copper, and gold plated and if
connected to a similar material the EMF voltage will be
less than 1uV/C.
The Transmille precision lead set contains one pair of safety
4mm to 4mm low thermal leads.
Reducing Thermal EMF - 2

There are three ways to reduce EMF:-
Second
Avoid large temperature gradients across the measurement
area. These can be caused by strong drafts of cold or
warm air coming from air condition units, the air from
such can be several degrees different. Also take care with
sun light through a window or a instrument or soldering
iron causing local heating.
Try to obtain a stable environment and leave connection to
thermally stabilise before zeroing the system. A calibrator
or DMM which was last used on high current may take
many minutes to settle down.
Reducing Thermal EMF - 3

There are three ways to reduce EMF:-
Thirdly
Make sure there is the minimum number of
connection between source and measurement
instrument. Do not extended leads by plugging
together. Try to keep leads short and close so
they will be at same temperature.
Reducing Thermal EMF - 4

After you have done all you can to reduce the EMF
repeating the measurement several times can give a better
measurement as the effects of thermals are usually
random. With some measurements it is possible to
reverse leads and take the difference, but note that
holding the connectors will warm them up, even the
friction caused by plugging in causes heat.

A practical limit is around 0.5uV and this is a common
value to used in uncertainty calculations.
Some manufactures to help when calculating uncertainties including
Transmille try to give a figure for this as in part it is the interaction
between the instrument terminal and the lead and in practice it will
always be present to some degree.
Making Accurate High Value
Resistance Measurements - 1

There are 3 problems measuring resistance above 100kohm.
Firstly is lead insulation resistance. Most test leads are made out of PVC
type plastic which is not a good insulator. A typical figure for insulation
for this type of test leads could be between 50 and 100Gohms. Even dirt
on the leads will reduce insulation.
Test leads with an insulation resistance of 100Gohm will give an error of
1% when measuring 1Gohm, and 0.1% on 100Mohm.
Note that Insulation resistance is not connected with voltage/current rating
of the lead.
Screened leads made from PTFE are the best, however this is brittle and
brakes very easily. Almost as good is 50 Ohm Coax with a polythene
insulating material.
Making Accurate High Value
Resistance Measurements - 2
Second problem is noise pick – up in the test leads and the resistor
itself. There is always present a strong field from the Power
line and this can easily swamp out the true signal.
As this pick up is AC it presence may not be noticed, all that may
be seen is a more noisy reading. This pick up will not be a nice
sine wave but a distorted signal, which when averaged out by
the DMM’s filter will give an offset which will also result in a
reading error
A 1Gohm resistor measured with a DMM at 10 volts will only
have a current of 10nA flowing in the test leads, a noise pick up
of just 10 Pico amps will be 0.1% noise on the reading. It is
therefore very important that screened leads are used, and the
case of the resistor is earthed.
Making Accurate High Value
Resistance Measurements - 3
To overcome, or at least reduce the lead and pick up problem it is
very common to measure high value resistors at higher voltages
to obtain higher currents to measure. The voltage is generated
by a calibrator and the current is measured by a Pico amp meter
or electrometer. Transmille 2000 series can also be used for
this measurement.
This however introduces the effect of ‘voltage coefficient’ in the
resistor where the actual value of the resistor changes with the
applied voltage, this is not due to self heating as in a power
resistor but due to ‘voltage stress’ and is an instantaneous
effect. It may be necessary to characterise the resistor for its
voltage coefficient.
Low AC Current Measurements

The main problem with this is current leakage to earth
through capacitive coupling through mains transformers,
capacitance to case etc, in fact anywhere there is an un
guarded path to earth.
Note this is not a DC path but an AC path which is
frequency dependant.
Again Pick – up is also a problem, use screened leads but
keep the distance short as the capacitance of screened
cable is high and this will again leak some current away.
Low value Resistance

To measure low ohms a 4wire Kelvin connection must be used, otherwise
the resistance of the connections & leads used will also be measured.

Kelvin connection work by using separate leads for the measurement
current and for measuring the voltage drop across the resistor. Any
voltage dropped along the leads carrying the current is not measured, and
there is no current flowing in the voltage sense leads the resistance of the
leads will make no difference.
Thermal EMF in the voltage sense leads will make a difference and the
same precautions as used for any low DC voltage measurement must be
taken.

Note some DMM’s can automatically compensate for thermal EMFs by
turning the measurement current on and off and measuring the voltage
difference.
Instrument Specifications.

Different manufacturers present specifications in many
different ways. Often the different word is term is used for
the same thing.

Basically however all specification come down to the same
3 terms and a temperature coefficient
1: A percentage of the actual Value
Sometime called % or ppm of reading or output
2: A percentage of the Full scale of that range
Sometime called % or ppm of range or full scale
3: A fixed allowance – or Floor
Often expressed as a number of digits or counts,
Or directly in units, example 3uV
Instrument Specifications - 1

To work out the allowance at any point simple work out
each value and add them up. Note that the ‘counts’ or
‘Digits’ refers to the last digit in the display – the resolution

Example
Measuring 1 volt on a DMM with a spec of .05% of reading +3counts



1 Volt at 0.05%
= 5mV
3 counts on 1.999V display = 3mV
Total error allowed
= 8mV
Instrument Specifications - 2




PPM and Percentage
These terms work in the same way, but percentage is based
on 100, ppm is based on one million,
(ppm = parts per million)
0.01% = 100ppm
Instrument Specifications - 3
An instrument accuracy specification is based on 3 factors
1)
2)
3)
The instruments change with time
The Design which may limit initial accuracy
The uncertainties to which the laboratory calibration can be
performed
The specification normally provided by the manufacturer are 1 and 2
combined, this of course requires that the manufacture supplies also
the time, often one year that the specification applies for.
This specification is relative to calibration standards, which means no
account has been taken of the uncertainties to which the instrument
has been calibrated to.
Absolute accuracy can be calculated by adding this figure to the
uncertainty to which the calibrator is certified. Transmille belive it
is unhelpful for the manufacture to give an absolute figure based on
the manufactures own laboratory, as this limits the choice of the
customer to select the calibration source.
Getting the best out of
modern high performance
DMMs and calibrators






Leave on in a stable environment
Run any auto-cal / self cal before use
Connect using the best leads for the measurement
Select the best range and settings
Null / zero any offset errors first
Allow time to stabilise after use on high power
ranges