Self heating in resistive components causes a temperature effect and a corresponding resistance change.
Use low test signals or pulsed measurement.
Heat sink the DUT.
Allow for effects of power coefficient of resistance.
Use low power coefficient components.
High voltage applied to DUT
Some high resistance resistors have a significant voltage coefficient of resistance.
Use low voltage coefficient components.
Measure at low voltages.
Allow for the effects of the voltage coefficient of resistance.
Ambient temperature
Temperature coefficient effects; possible permanent retrace effects may result from large shipping or storage temperature shifts.
Maintain stable temperature and minimize exposure to temperature extremes.
Humidity
Humidity may cause leakage effects on high resistance components.
Maintain relative humidity under 50%.
Thermal emf
The thermal emf, i.e. the voltage generated at contacts of dissimilar metals at temperature gradients, can cause erroneous voltage and resistance measurements.
Use Cu. to Cu. contacts and leads wherever possible; silver contacts and solder are acceptable.
Avoid using steel and brass.
Minimize temperature gradients or drafts.
Use switched or "true ohm" measurement instruments.
Alternate leads to determine the degree of the problem.
Low resistance
Lead resistance and thermal emf may introduce errors.
Use 4-wire measurement, Kelvin leads.
See thermal emf (above).
High resistance
Leakage through lead insulation and bench top, resulting from humidity, may cause errors.
Use low leakage insulation such as Teflon
Set DUT on high insulation sub-plate.
Maintain all terminals clean.
Shield and avoid high voltage and movement nearby.
Use 5 or 6 terminal guard circuit.
Ground loops
Ground currents can introduce noise and offset voltage.
Use radial grounds to only one reference point.
Test conditions
Most resistors, capacitors, and inductors are non-ideal; wire wound resistors are both inductive and capacitive; capacitors have losses, and inductors can be very resistive. Test conditions of voltage, frequency and model (parallel or series) may be significant to the measurement.
Apply the instrument test conditions that are the most representative model of the DUT.