An instrument gives readings tightly clustered together but consistently offset from the true value. This describes

high precision, low accuracy — the signature of a systematic error

high precision and high accuracy simultaneously

low precision caused entirely by random error

A quantity Z = X/Y is computed from two independently measured quantities X and Y. The error in Z combines the errors in X and Y as

the error in Z is always smaller than in either X or Y

The Kelvin double bridge is used specifically because

its extra ratio arm cancels lead and contact resistance in low-resistance measurement

it measures capacitance more precisely than a Schering bridge

its extra ratio arm cancels lead and contact resistance in low-resistance measurement

it needs no standard for comparison

it operates only at very high frequency

A Schering bridge is the standard choice for

capacitance and dielectric loss measurement, especially in high-voltage insulation testing

very low resistance measurement

capacitance and dielectric loss measurement, especially in high-voltage insulation testing

measuring the Q factor of a coil directly

Opening the secondary of a current transformer while its primary carries current is dangerous because

the demagnetising ampere-turns are removed, the core saturates and a high voltage can appear

the primary current drops to zero suddenly

the demagnetising ampere-turns are removed, the core saturates and a high voltage can appear

the transformer immediately short-circuits itself

it has no effect since the transformer is designed for this

The two-wattmeter method for three-phase power gives the power factor angle from readings W1 and W2 as

phi cannot be found from W1 and W2 alone

A true-RMS meter differs from an average-responding meter in that it

computes the actual root-mean-square value regardless of waveform shape

reads correctly only for a pure sinusoid

computes the actual root-mean-square value regardless of waveform shape

is always less accurate than an average-responding meter

A digital storage oscilloscope's key advantage over a purely analog scope is that it can

capture, freeze and measure a single transient event by sampling and storing it digitally

display frequency-domain content directly with no processing

capture, freeze and measure a single transient event by sampling and storing it digitally

measure voltage with no analog-to-digital conversion at all

operate without any triggering circuitry

Errors, Bridges and Electrical Measurement

Section 8 of the GATE Instrumentation Engineering (IN) paper, and the discipline's own name in miniature: how a reading is actually obtained, how confident one can be in it, and what an error in it is really worth. Start with how errors are classified and combined, since that classification decides which errors can be calibrated away and which cannot; then the bridge circuits that measure resistance, inductance and capacitance by NULLING a detector rather than by reading a deflection, which is what makes them precise; then the everyday measurement of voltage, current and power, DC and AC, single- and three-phase; and finally the digital instruments — the DVM, the DSO, the spectrum analyzer — that have replaced the analog meter on almost every bench.

1. Errors, uncertainty and regression

A systematic error repeats the same way every time under the same conditions — a miscalibrated instrument, a fixed offset — and it CAN be corrected once identified, by calibration or a correction factor. A random error scatters differently on every repeated reading, from causes too small and numerous to track individually, and it can only be reduced statistically, by averaging more readings, never eliminated by a one-time correction. Accuracy is closeness to the true value; precision is closeness among repeated readings of the same quantity — a instrument can be precise (tightly repeatable) and inaccurate (consistently offset from truth) at the same time, which is exactly what a systematic error alone produces.

🧠 Propagation of error, by operation

For a quantity computed from measured ones: if it is a SUM or DIFFERENCE, absolute errors add (in quadrature for independent random errors); if it is a PRODUCT or QUOTIENT, relative (fractional) errors add instead. This single rule is why a quantity formed by dividing two poorly-known numbers can have a much larger relative error than either number alone, and why choosing to compute a difference of two large, nearly equal readings is a classically bad idea — the absolute errors stay the same size while the result shrinks, so the relative error explodes.

Linear regression fits a straight line to a scatter of points by minimising the sum of squared vertical deviations (least squares); weighted regression extends this by giving points known to be measured more precisely a larger weight in that sum, so a handful of noisy points cannot drag the fitted line as far as they would in an unweighted fit. Both are the standard way an instrument's own calibration curve is built from a set of reference measurements.

2. Bridges: measuring by nulling, not deflecting

Every bridge in this list works the same way: an unknown impedance is placed in one arm of a four-arm network, known standards in the others, and the arms are adjusted until a detector across the bridge's diagonal reads zero — at balance, the unknown is a simple ratio of the known standards and the detector's own sensitivity and calibration drop out of the answer entirely. That is the whole reason a null method beats a deflection method for precision work: the accuracy of the result depends on the accuracy of the standards, not on the accuracy of the meter.

Which bridge measures what, and its one distinguishing feature
Bridge	Measures	Distinguishing feature
Wheatstone	medium resistance	the base case: four arms, DC or AC excitation
Kelvin (double bridge)	very low resistance	a second ratio arm cancels lead and contact resistance
Maxwell	inductance (medium Q)	balances an unknown inductor against a known capacitor
Anderson	inductance (wide Q range)	an extra branch makes balance independent of frequency
Schering	capacitance and dielectric loss	the standard bridge for high-voltage insulation testing
Wien	frequency (or capacitance)	balance depends on frequency; also the RC network behind the Wien-bridge oscillator

🎯 Why the Kelvin bridge exists at all

An ordinary Wheatstone bridge balances by comparing ratios of resistance, and at very low resistance values the lead and contact resistance in the connections becomes comparable to the unknown itself, corrupting the result. The Kelvin double bridge adds a second, auxiliary ratio arm specifically so that the lead resistance between the unknown and its adjacent standard cancels out of the balance equation exactly — which is precisely the case a plain Wheatstone bridge cannot handle.

A Q-meter measures the quality factor of a coil directly, by exciting a series LC circuit at resonance and reading the voltage magnification across the capacitor; megohm (insulation-resistance) measurement uses a dedicated high-voltage meter (the megger) precisely because an ordinary ohmmeter's low test voltage cannot reliably stress and reveal a failing insulation path the way the working voltage would.

3. Voltage, current and power measurement

A meter that reads true RMS computes the actual root-mean-square value of a waveform regardless of its shape, by squaring, averaging and taking the root; a cheaper average-responding meter is calibrated to read correctly ONLY for a pure sinusoid, and reads a genuinely wrong value on any distorted or non-sinusoidal waveform, which is why a true-RMS meter matters wherever harmonics are expected. Voltage and current scaling for ranges beyond a meter's native span uses multipliers (a series resistor extends a voltmeter's range) and shunts (a parallel low resistance extends an ammeter's range) at low voltage and current, and instrument transformers — potential transformers and current transformers — at power-system voltages and currents, which additionally provide galvanic isolation between the high-power circuit and the measuring instrument.

⚠️ A current transformer's secondary must never be left open

A current transformer's primary is driven by the line current regardless of what the secondary is doing, so opening the secondary while the primary carries current removes the demagnetising ampere-turns that normally hold the core flux down — the core saturates and the open secondary can develop a dangerously high voltage. A potential transformer has no such rule; its secondary is normally left open when unused and is instead never SHORTED, the opposite hazard.

Single-phase power is read by a single wattmeter (voltage coil and current coil, output proportional to VIcos(phi)); three-phase power on a balanced or unbalanced system is read with the two-wattmeter method, whose two readings W1 and W2 give total power as their sum and, separately, the power factor angle from tan(phi) = sqrt(3)*(W1-W2)/(W1+W2) — a result that also reveals the load's power factor from readings that individually mean little on their own.

4. Time, frequency and digital instruments

A digital timer/counter measures time interval or frequency by gating a known-frequency clock (for time interval, count clock pulses during the event; for frequency, count the unknown signal's cycles during a known gate time) — the same hardware answers both questions depending only on which signal is gated by which. Phase measurement between two signals of the same frequency reduces to a time-interval measurement between corresponding zero-crossings, scaled by the period.

A digital voltmeter (DVM) reads a single quantity, DC or AC voltage, converted through an internal ADC and displayed numerically.
A digital multimeter (DMM) extends this to current, resistance and often frequency and capacitance, by adding a signal-conditioning front end ahead of essentially the same ADC.
A digital storage oscilloscope (DSO) captures a waveform's shape over time by sampling and storing it digitally, which is what lets it freeze, zoom into and measure a single transient event a purely analog scope's phosphor persistence could never hold still.
A spectrum analyzer displays a signal's content by FREQUENCY rather than by time, revealing harmonics and interference an oscilloscope's time-domain trace hides in plain sight.

ℹ️ A digital energy meter's real job

A digital energy meter multiplies sampled voltage and current instantaneously to get instantaneous power, then integrates that product over time — it is not merely 'a digital voltmeter and ammeter combined', because energy is the time-integral of power, not a static product of two readings, and getting this integration right under distorted, non-sinusoidal waveforms is exactly what distinguishes a good energy meter from a poor one.

Key takeaways

A systematic error can be calibrated away once identified; a random error can only be reduced statistically, by averaging more readings.
Errors add absolutely for sums/differences and relatively for products/quotients — computing a difference of two large near-equal readings inflates relative error badly.
Every bridge in this section balances a detector to null, so the result depends on the accuracy of the standards, not the meter — that is the whole reason a null method is precise.
The Kelvin double bridge's extra ratio arm exists specifically to cancel lead and contact resistance, which corrupts an ordinary Wheatstone bridge at very low resistance.
A current transformer's secondary must never be opened while the primary carries current; a potential transformer's secondary must never be shorted — the two hazards are opposite.
A digital energy meter integrates instantaneous power over time; it is not simply a combined voltmeter and ammeter reading.

Practice questions (8)

Attempt each one before opening the answer. Every explanation names the tempting wrong option as well as the right one, because that is where marks are lost.

An instrument gives readings tightly clustered together but consistently offset from the true value. This describes
1. high accuracy, low precision
2. high precision, low accuracy — the signature of a systematic error
3. high precision and high accuracy simultaneously
4. low precision caused entirely by random error
Show answer
Answer: B — high precision, low accuracy — the signature of a systematic error
Tight clustering (repeatability) is precision; a consistent offset from truth is a lack of accuracy — and a consistent offset is exactly what a systematic error produces, since it repeats the same way every time rather than scattering. Random error would scatter the readings and reduce precision, which is not what is described here.
A quantity Z = X/Y is computed from two independently measured quantities X and Y. The error in Z combines the errors in X and Y as
1. absolute errors simply add
2. relative (fractional) errors add
3. the error in Z is always smaller than in either X or Y
4. the errors cancel exactly
Show answer
Answer: B — relative (fractional) errors add
For a product or quotient, it is the RELATIVE errors that add (in quadrature, for independent random errors) — this is the standard error-propagation rule, and it means the fractional uncertainty in a ratio of two poorly known quantities can be substantial even if each is reasonably well measured. Absolute-error addition applies to sums and differences, not products or quotients.
The Kelvin double bridge is used specifically because
1. it measures capacitance more precisely than a Schering bridge
2. its extra ratio arm cancels lead and contact resistance in low-resistance measurement
3. it needs no standard for comparison
4. it operates only at very high frequency
Show answer
Answer: B — its extra ratio arm cancels lead and contact resistance in low-resistance measurement
At very low resistance values, the lead and contact resistance in the connecting wires becomes comparable to the unknown itself; the Kelvin bridge's extra auxiliary ratio arm is specifically designed to make the balance condition independent of exactly that stray resistance. It has nothing to do with capacitance, still needs known standards like any bridge, and is not restricted to very high frequency.
A Schering bridge is the standard choice for
1. very low resistance measurement
2. capacitance and dielectric loss measurement, especially in high-voltage insulation testing
3. measuring the Q factor of a coil directly
4. frequency measurement alone
Show answer
Answer: B — capacitance and dielectric loss measurement, especially in high-voltage insulation testing
The Schering bridge is built around measuring capacitance and, critically, the dielectric loss angle of that capacitance, which is why it is the standard tool for insulation and cable testing at high voltage. Kelvin covers very low resistance, the Q-meter measures coil Q directly by resonance, and the Wien bridge is the one primarily associated with frequency.
Opening the secondary of a current transformer while its primary carries current is dangerous because
1. the primary current drops to zero suddenly
2. the demagnetising ampere-turns are removed, the core saturates and a high voltage can appear
3. the transformer immediately short-circuits itself
4. it has no effect since the transformer is designed for this
Show answer
Answer: B — the demagnetising ampere-turns are removed, the core saturates and a high voltage can appear
The primary current is set by the line, not by the secondary, so it does not drop. With the secondary open, the demagnetising ampere-turns that secondary current normally provides vanish, the core flux and hence the induced secondary voltage rise sharply (the core saturates), and a dangerously high voltage can appear across the open terminals — the opposite hazard from a potential transformer, whose secondary should never be shorted.
The two-wattmeter method for three-phase power gives the power factor angle from readings W1 and W2 as
1. tan(phi) = W1 + W2
2. tan(phi) = sqrt(3) * (W1 - W2) / (W1 + W2)
3. tan(phi) = W1 * W2
4. phi cannot be found from W1 and W2 alone
Show answer
Answer: B — tan(phi) = sqrt(3) * (W1 - W2) / (W1 + W2)
The standard two-wattmeter relation gives total power as W1 + W2 and, separately, the power factor angle from tan(phi) = sqrt(3)*(W1 - W2)/(W1 + W2) — the same two readings answer both questions, which is exactly what makes the method useful rather than merely a way to add two numbers.
A true-RMS meter differs from an average-responding meter in that it
1. reads correctly only for a pure sinusoid
2. computes the actual root-mean-square value regardless of waveform shape
3. can only measure DC quantities
4. is always less accurate than an average-responding meter
Show answer
Answer: B — computes the actual root-mean-square value regardless of waveform shape
A true-RMS meter squares, averages and roots the actual waveform, giving the correct RMS value whatever its shape; it is the average-responding meter that is calibrated correctly only for a pure sinusoid and reads wrongly on distorted waveforms. Neither statement about DC-only measurement or lower accuracy is correct — the true-RMS meter is the MORE accurate one on non-sinusoidal signals.
A digital storage oscilloscope's key advantage over a purely analog scope is that it can
1. display frequency-domain content directly with no processing
2. capture, freeze and measure a single transient event by sampling and storing it digitally
3. measure voltage with no analog-to-digital conversion at all
4. operate without any triggering circuitry
Show answer
Answer: B — capture, freeze and measure a single transient event by sampling and storing it digitally
Digital storage lets a DSO sample and hold a waveform's data, so a single non-repeating transient can be frozen, zoomed and measured after the fact — an analog scope's phosphor persistence fades and cannot do this. Frequency-domain display is a spectrum analyzer's job, not automatic on a DSO; it still requires an ADC internally; and triggering circuitry is still very much needed to decide when to capture.