Measurement systems:

● Measurands are sensed/measured

● Measurements systems contains:
○ Sensing element (something that responds to change)
○ Signal modification (converts signal to measurement)
○ Indicator or recorder (displays measurement)
● eg. mercury thermometer
○ Measurand: temperature
○ Sensing element: volume of mercury (V∝T)
○ Signal modification: narrow stem of tube
○ indicator/recorder: graduated markings
Error and uncertainty:
● Types of error
○ Gross error
■ due to mistake
○ Systematic error (also called bias error) denoted: B
■ Consistent and repeatable
■ Fixed offset (added value)
■ Calibration error
■ Loading error (measuring device effects measurand)
■ Spatial error (measurement point doesn’t reflect whole system)
○ Random error (also called precision error) denoted: P
■ Inconsistent and lacks repeatability
■ ↗scatter → ↗random error
● 𝑒𝑟𝑟𝑜𝑟(ε) = 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑑 𝑣𝑎𝑙𝑢𝑒 (𝑥) − 𝑡𝑟𝑢𝑒 𝑣𝑎𝑙𝑢𝑒 (𝑋)
○ True value is unknown ⇒ error is inability to accurately measure
● Range and span
○ Measurement devices are rated for specific range (if no lower bound, assume 0)
■ 𝑠𝑝𝑎𝑛 = 𝑢𝑝𝑝𝑒𝑟 − 𝑙𝑜𝑤𝑒𝑟
● Resolution
○ Smallest increment detectable (by measuring device)
Sensitivity:
∆𝑜𝑢𝑡𝑝𝑢𝑡
● 𝑠𝑒𝑛𝑠𝑖𝑡𝑖𝑣𝑖𝑡𝑦 = ∆𝑖𝑛𝑝𝑢𝑡
● Want to maximize sensitivity
○ ∴small change in input → large change in output
Accuracy:
● State as percentage (%)
○ Full scale (FS) → span of measurement device
○ Reading (rdg) → current reading of device
○ +3 dgts → 3 units of smallest increment
● Generally, want small range that still encampases measurement
● Accuracy
 ○ Closeness of data to true value
● Precision
○ Spread of data (random error)
Uncertainty:
𝑎𝑏𝑠𝑜𝑙𝑢𝑡𝑒 𝑢𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦
● 𝑟𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑢𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 = 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑚𝑒𝑛𝑡 𝑣𝑎𝑙𝑢𝑒

Calibration:
● Static calibration (inputs independent of time)
● Dynamic calibration (inputs vary with time)
● Performed with known values
● Linearity error
○ Incorrect assumption of linear relationship
● Calibration range → +/- 50% beyond expected measurements range
● Hysteresis error
○ Loading vs unloading readings are different
○ Due to friction, electrical capacitance, etc
Rounding:
● 0.1 to 0.4 → 0
● 0.5 → closest even number
● 0.6 to 0.9 → 1
Significant figures:
● Measure of precision of value
● Calculated value is as precise as least precise measurement
● Fractions are infinite sig figs
● Operations with sig figs
○ Adding/subtracting
■ Same #of decimal points as least precise value
○ Multiplying/dividing
■ Same sig figs as least precise value
Electricity basics:
● Almost all measurement devices are electronics
● Physical properties
○ Electric charge
■ Denoted: q
■ Units: [C] (coulomb)
○ Current
■ Denoted: I
■ Units: [A] (ampere)
𝑑𝑞
■ 𝐼= 𝑑𝑡
○ Voltage (also called electric potential difference)
■ Denoted: V
■ Units: [V]
 ∆𝑃𝐸
■ 𝑉= 𝑞
○ Resistance
■ Denoted: R
■ Units: [Ω]
■ 𝑉 = 𝐼𝑅
○ Electrical resistivity
■ Denoted: ρ
■ Units: [Ω·m]
■ Property of material
ρ𝑙
■ 𝑅= 𝐴
(pouillet’s law), where A is cross-sectional area, l is length of wire
○ Electrical power
■ Denoted: P
■ Units: [W]
𝑑𝑊
■ 𝑃= 𝑑𝑡
= 𝑉𝐼
○ Electrical capacitance
■ Denoted: C
■ Units: [F] (farad)
■ Property of capacitor
𝑑𝑞
■ 𝐶= 𝑑𝑉
𝑑𝑉
■ 𝐼= 𝐶 𝑑𝑡
○ Magnetic flux
■ Denoted: Ψ
■ Units: [Wb] (weber)
■ Magnetic field through area/surface
○ Electrical inductance
■ Denoted: L
■ Units: [H] (henry)
𝑑Ψ
■ 𝐿= 𝑑𝐼
𝑑𝐼
■ 𝑉= 𝐿 𝑑𝑡
● Equations
𝑞1 𝑞2
○ 𝐹 = 𝑘𝑒 2 (coulomb’s law)
𝑟
1 2
○ 𝐸= 2
𝐶𝑉
● Electrical components
○ Resistors
○ Capacitors
■ Stores charge (q)
○ Inductor
■ Resists change to current (I)
■ They are coiled wire
 ● Coil+current → Induced magnetic field → induced current
● Kirchhoff’s law
○ Σ𝐼𝑖𝑛 = Σ𝐼𝑜𝑢𝑡
○ Σ𝑉 = 0

Measurement devices circuit:

● Often measurement device measure a change in resistance
Voltage dividers:
𝑉1 𝑉2
● 𝑅1
= 𝑅2
(= 𝐼)
𝑉−𝑉2 𝑉2
𝑅1
= 𝑅2
𝑅1
𝑉 − 𝑉2 = 𝑅2
𝑉2
𝑅1 𝑅1+𝑅2
𝑉= 𝑅2
𝑉2 + 𝑉2 = 𝑅2
𝑉2
𝑅2
𝑉2 = 𝑅1+𝑅2
𝑉
𝑅1
● Loading ratio: 𝑘 = 𝑅2, 0
where R2, o is resistance of R2 with no applied load

● 𝑅𝑜 ∼ 𝑉2, 𝑜 ≠ 0 (ie. base resistance is not zero)

𝑉2 1 𝑅2
○ For reading 𝑥 = 0, 𝑉
= 1+𝑘
(from 𝑉2 = 𝑅1+𝑅2
𝑉)

Wheatstone bridges:
● Constant offset produced by the voltage divider is removed by adding a voltage divider
○ Called wheatstone bridge
● Produces differential voltage output
● 2 voltage dividers in parallel
○ Voltage is read across midsection of bridge
𝑅2 𝑅4
● 𝑉𝑜𝑢𝑡𝑝𝑢𝑡 = 𝑉𝑏 − 𝑉𝑒 = 𝑅1+𝑅2
𝑉− 𝑅3+𝑅4
𝑉
𝑅1 𝑅3
● 𝑘= 𝑅2
= 𝑅4
⇒ bridge is balanced (𝑉𝑜 = 0)
○
Called bridge ratio
○
Ratio of individual voltage dividers are equal
■ This means one can be a scaler of the other
● One resistance is variable (doesn't matter which one)
○ Variation will affect our reading of V (voltage across the bridge) → measurand
○ When this is 0, (𝑉𝑜 = 0) this is no reading
○ Which is why we use wheatstone bridges (gets rid of offset present in voltage dividers)
δ𝑅
● 𝑥= 𝑅𝑜
 ○ We have R4 as our variable resistance ∴𝑅4 = 𝑅𝑜 + δ𝑅
𝑉𝑜
● We’re interested in output voltage ratio 𝑉
● The change in V related to the measurand is extremely small
○ ⇒ amplified by a operational amplifiers (op amps)
Deriving output voltage equations
● Givens
○ 𝑅1 = 𝑘𝑅2
○ 𝑅3 = 𝑘𝑅𝑣
○ 𝑅𝑣 = 𝑅𝑜 + δ𝑅

● Voltage dividers
𝑉𝑜 𝑅2
○ 𝑉
= 𝑅1+𝑅2
𝑉𝑜 𝑅𝑜+δ𝑅
○ 𝑉
= 𝑘𝑅𝑜+𝑅𝑜+δ𝑅
δ𝑅
𝑉𝑜 𝑅𝑜(1+ 𝑅
)
○ 𝑉
= δ𝑅
𝑘𝑅𝑜+𝑅𝑜(1+ 𝑅
)
δ𝑅
𝑉𝑜 (1+ 𝑅 )
○ 𝑉
= δ𝑅
𝑘+(1+ 𝑅 )
𝑉𝑜 1+ 𝑥
○ 𝑉
= 1+𝑘+ 𝑥
● Wheatstone bridge
𝑉𝑜 𝑅2 𝑅𝑣
○ 𝑉
= 𝑅1+𝑅2
− 𝑅3+𝑅𝑣
𝑉𝑜 𝑅2 𝑅𝑜+δ𝑅
○ 𝑉
= 𝑘𝑅2+𝑅2
− 𝑘𝑅𝑜+𝑅𝑜+δ𝑅
δ𝑅
𝑉𝑜 1 𝑅𝑜(1+ 𝑅
)
○ 𝑉
= 𝑘+1
− δ𝑅
𝑘𝑅𝑜+𝑅𝑜(1+ 𝑅
)
δ𝑅
𝑉𝑜 1 (1+ 𝑅 )
○ 𝑉
= 𝑘+1
− δ𝑅
𝑘+(1+ 𝑅 )
δ𝑅 δ𝑅
𝑉𝑜 (𝑘+1+ )−(𝑘+1)(1+ )
○ 𝑉
= 𝑅
δ𝑅
𝑅

(𝑘+1)(𝑘+1+ 𝑅
)
𝑉𝑜 (𝑘+1+ 𝑥)−(𝑘+1)(1+𝑥)
○ 𝑉
= (𝑘+1)(𝑘+1+ 𝑥)
𝑉𝑜 (𝑘+1+ 𝑥)−(𝑘𝑥+𝑘+𝑥+1)
○ 𝑉
= (𝑘+1)(𝑘+1+ 𝑥)
𝑉𝑜 −𝑘𝑥
○ 𝑉
= (1+𝑘)(1+𝑘+𝑥)
Amplifier circuits:
● Amplify voltage
Impedance matching:
● We don’t want the amplifier circuit to
affect the previous circuit
○ Will act as a load ∴affecting it
(draws voltage from it)
○ Due to circuits having different
effective resistance
● ∴want to match impedance/resistance
○ Match ≠ equal
○ Impedances designed to maximize properties (max power, stable voltage, etc)
Equivalent resistance:
● Determined using Thevenin’s theorem
● Circuits can be reduced to:
○ Equivalent resistance
○ Equivalent voltage
● These need to match between amplifier circuit and measurement circuit
● Impedance (equivalent resistance) of amplifying circuit ≥ measurement circuit
○ Otherwise get loading error
● Ideal circuit (with perfect impedance matching):
○ ∞ input impedance (Z) ∴no current (I) drawn from original circuit
○ 0 output impedance/resistance (Z) ∴output voltage (V) is unaffected by output current (I)
○ 𝑉− = 𝑉+
○ 𝑖− = 𝑖+ = 0 (no current)
Operational amplifiers (Op-Amps):
● Amplify voltage difference
○ Uses open loop gain
○ Amplification: 𝐴𝑂𝐿 > 100 000
■ 𝑉𝑜 = 𝐴𝑂𝐿(𝑉+ − 𝑉−)
● Made with transistors in integrated circuit
● Open vs. closed loop Op Amps
○ Open: output is NOT connected to input
○ Closed: output is connected to input
■ ∴tends to equalize
● Ideal Op-Amp:
○ Infinite amplification AOL (for open loop) (called open loop gain)
○ No current in or out of op amps inputs
○ With negative feedback, voltage of inputs are equal
● Output voltage (Vo) < input voltage (Vs)
○ Otherwise op-amp is “saturated”
Uses of OpAmps:
● Comparator
○ Closed loop configuration
○ 𝑉+ > 𝑉− ⇒ 𝑉𝑜 =+ 𝑉𝑆 otherwise, 𝑉𝑜 =− 𝑉𝑆
● Inverter
○ Open loop configuration
● Non-inverting
○ Open loop configuration
● Summing (non-inverting)
○ Open loop configuration
● Summing (inverting)
● Subtractor
● Voltage follower
○ Input voltage = output voltage
○ Used to eliminate loading effects (since no current)
● Differentiator
● Integrator
Practice solving for output voltage in OpAmps

System response
● Interacting circuits often creates time dependency
● Input signal → output signal
○ But not always immediate
○ How long is the delay between a unit step input and output response
● 0 order response
○ Immediate response
● 1st order response
○ Gradual shift in output signal
nd
● 2 order response
○ Oscillating or damped oscillating response

Where x(t) is input, y(t) is output

Zero-order system (steady state):
● Differential equation: 𝑎𝑜𝑦 = 𝑏𝑥
● 𝑦 = 𝐾𝑠𝑥
● 𝐾𝑠 is static sensitivity
 ○ Defines amplification of input
● No time dependency
○ Immediate output response
● For multiple devices connected in series
○ Static sensitivities (Ks) are multiplied
First-order system:
𝑑𝑦
● Differential equation: 𝑎1 𝑑𝑡
+ 𝑎𝑜𝑦 = 𝑏𝑥
𝑑𝑦
○ τ 𝑑𝑡
+ 𝑦 = 𝐾𝑠𝑥
● τ is time constant
○ Defines time for system to respond to change in input
■ ↗τ ⇒ 𝑠𝑙𝑜𝑤𝑒𝑟 𝑠𝑦𝑠𝑡𝑒𝑚 𝑟𝑒𝑠𝑝𝑜𝑛𝑠𝑒
−𝑡/τ −1
○ Represents time at which τ = 𝑡 ∴𝑒 =𝑒 = 0. 36788
−1
■ For loading this is 1 − 𝑒 = 0. 63212
■ Represents time taken [s] to reach 63.2% of change of 𝑦𝑓 − 𝑦𝑖
○
−𝑡/τ
● Solved ODE: 𝑦(𝑡) = (𝑦𝑖𝑛𝑖𝑡𝑖𝑎𝑙 − 𝑦𝑠𝑠)𝑒 + 𝑦𝑠𝑠
○ yss is output at steady state value
RC circuit:
● Charging:
𝑑𝑉𝑜
○ 𝐼=𝐶 𝑑𝑡
𝑉𝑖−𝑉𝑜 𝑑𝑉𝑜
○ 𝑅
=𝐶 𝑑𝑡
−𝑡/𝑅𝐶
○ Solved ODE: 𝑉𝑜 = 𝑉𝑖(1 − 𝑒 )
■ Initial condition: 𝑡 = 0, 𝑉𝑜 = 0
■ Steady state condition: 𝑡 = ∞, 𝑉𝑜 = 𝑉𝑖
■ τ = 𝑅𝐶 (units: [s])
● How fast a system reacts to changes
● Discharging
−𝑉𝑜 𝑑𝑉𝑜
○ 𝑅
=𝐶 𝑑𝑡
−𝑡/𝑅𝐶
○ Solved ODE: 𝑉𝑜 = 𝑉𝑚𝑎𝑥𝑒

Solving ODEs:
● 𝑦 = 𝑦𝑔𝑒𝑛𝑒𝑟𝑎𝑙 + 𝑦𝑝𝑎𝑟𝑡𝑖𝑐𝑢𝑙𝑎𝑟
● ygeneral is based on system: assumes input x(t)=0
−𝑡/𝑅𝐶
○ Always: 𝑦𝑔𝑒𝑛𝑒𝑟𝑎𝑙 = 𝐶𝑒 (note: C is constant depending on initial conditions)
● yparticular is based on type of input: function x(t)
Resistance / Impedance:
● Resistors, capacitors and inductors all effect current and voltage
○ Give effective resistance (called “impedance”)
● In DC circuits, only resistors affect current & voltage
● Given constant voltage (V)
○ Resistors → decrease current (I)
■ Called resistance
○ Inductors → lagging current (I)
■ Called inductive reactance
○ Capacitors → leading current (I)
■ Called capacitive reactance
● 𝑍 = 𝑅 + 𝑋𝑖𝑛𝑑𝑢𝑐𝑡𝑜𝑟 + 𝑋𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑜𝑟
○ Z is impedance [Ω]
○ R is resistance [Ω]
○ X is reactance [Ω]
Using s-domain (Laplace transform):
𝑑𝑦 𝑥
● Laplace transform makes derivatives ( 𝑑𝑥 ) and exponentials (𝑒 ) easy to use (polynomials)
○ Inductive and capacitive reactance are governed by 1st order ODEs ⇒ ex
● In t-domain:
𝑉(𝑡)
○ 𝑅(𝑡) = 𝐼(𝑡)
● In s-domain:
𝑉(𝑠)
○ 𝑍(𝑠) = 𝐼(𝑠)
○ Acts like resistance: 𝑉(𝑠) = 𝐼(𝑠) 𝑍(𝑠)
○ Therefore gets ride of time dependance
■ Everything is constant
Frequency domain:
● All time-varying signals can be approximated into sum of cosine waves
○ Using fourier series
● Convert input signals to cosine wave of form: 𝑦(𝑡) = 𝐴 𝑠𝑖𝑛(ω𝑡 + ϕ)
○ ω = 2π𝑓
● Calculations
○ Assume current 𝐼(𝑡) = 𝐴 𝑐𝑜𝑠(ω𝑡 + ϕ) (use this for derivative)
● s-domain
𝑖𝑥 𝑖𝑥
○ Euler formula: 𝑒 = 𝑐𝑜𝑠(𝑥) + 𝑖 𝑠𝑖𝑛(𝑥) BUT no sin(x) ∴𝑒 = 𝑐𝑜𝑠(𝑥)
○ 𝑠 = 𝑖 ω = 𝑖 2π𝑓 note: i is sometimes written j
Transfer function:
● Ratio of output to input signal in s-domain
● Describes how system modify input
○ Treats system as black box
𝐾𝑠
● 1st order systems: 𝐻(𝑠) = 1+τ𝑠
Dynamic system response
● Types of inputs causing 1st and 2nd order systems
○ Impulse
○ Step
○ Ramp
Errors due to system response:
● Dynamic response error
○ Difference in input vs output at given time
○ Property of specific time
● Steady state error
○ Dynamic response error as time → ∞
○ Property of system and input type
● Use transfer function to relate input to output
Identifying order of system:
● Based on number of capacitors (C) and inductors (L)
○ Represent +1 increase in order
■ Since both have ‘s’ in impedance
Damping (2nd order systems):
● No damping means oscillates forever (no steady state error since doesn’t converge)
● ζ is damping ratio
𝑐 𝑐
○ ζ= 𝑐𝑐
= 2 𝑚 ω𝑛

○ Large ζ → decrease oscillations

● Types of damping
○ ζ < 1 → underdamped (overshoots and oscillates)
○ ζ = 1 → critically damped (no oscillation)
○ ζ > 1 → overdamped (no oscillation, slow response)
● We want correct damping in signal processing
○ Converge to correct value
■ Want critically damped
Solving system response (output) for dynamic systems:
● Determine transfer function
𝐾𝑠
○ 1st order: 𝐻(𝑠) = 1+τ𝑠
2
𝐾𝑠ω𝑛
○ 2nd order: 𝐻(𝑠) = 2 2
𝑠 +2ω𝑛ζ𝑠+ω𝑛

● Find input in s-domain (laplace)

● Find output in s-domain: 𝑌(𝑠) = 𝐻(𝑠)𝑋(𝑠)
● Find output in t-domain (inverse laplace)

Frequency response - sinusoidal inputs

● System response to sinusoidal/wave inputs
Amplification/gain from transfer function (H):
● 𝑎𝑚𝑝𝑙𝑖𝑓𝑖𝑐𝑎𝑡𝑖𝑜𝑛 (𝐴𝑂𝑆) = |𝐻(𝑠)|
○ BUT H(s) in t-domain (use 𝑠 = 𝑖 ω)
○ Consider real and imaginary part of H(s)
● H is often affected by frequency
○ ∴amplification is affected by frequency
Frequency based filters:
● Low-pass filters
○ As frequency → \inf ⇒ A → 0
○ Amplifies lowest frequencies
● High-pass filter
○ As frequency → 0 ⇒ A → 0
○ Amplifies highest frequencies
Measuring amplification/gain (GdB):
● Often uses logarithmic scale
● Units: [dB]
● 𝐺𝑑𝐵 = 20 𝑙𝑜𝑔10(|𝐻(𝑠)|)
○ note: for no amplification (𝐻(𝑠) = 1) ⇒ 𝐺𝑑𝐵 = 0
● Static gain (pass band gain):
○ GdB when no signal attenuation
● Cut-off frequency (ωc)
○ Frequency at which signal starts to attenuate
○ Arbitrarily chosen as: 𝑠𝑡𝑎𝑡𝑖𝑐 𝑔𝑎𝑖𝑛 − 3 [𝑑𝐵]
1
■ 3 [𝑑𝐵] ≈ 20 𝑙𝑜𝑔10( )
2
1
■ ∴|𝐻(𝑠)| =
2
○ For 1st order systems

2 −τω𝑐 2
1 1
■ |𝐻(𝑠)| = || 1+τ𝑠 || = ( 2 ) + ( 2 )
1+(τω𝑐) 1+(τω𝑐)
1
■ Considering 3dB is cut off, |𝐻(𝑠)| =
2
1
● ∴ ω𝑐 = τ
(frequency response \related time constant)

Frequency response:
● What frequencies are affected (usually attenuated) by system
● Used in low and high pass filters
○ See “bode plot” on right
● Removing unwanted frequencies
○ eg. removing high frequency sensor noise
○ Remove DC voltage component
○ Removing 60Hz noise from electrical outlets
Frequency response - 2nd order system:
2
𝑑𝑦 𝑑𝑦 2
● 2 + 2ζω𝑛 𝑑𝑡
+ ω𝑛 𝑦 = 𝐾𝑠 𝑥 ODE relating input (x) to output (y)
𝑑𝑡
○ ωn is natural frequency, ζ is damping ratio
○ Ks is static sensitivity (ie. steady state gain) (ie. amplification)
2
𝐾𝑠ω𝑛
● 𝐻(𝑠) = 2 2 transfer function: ratio relating input to output (ie. solution to ODE)
𝑠 +2ω𝑛ζ𝑠+ω𝑛

Resonance:
● When input frequency equals natural frequency of underdamped system
○ ω = ω𝑛
○ Makes voltage → ∞
● Big problem since P=VI, ∴ P→∞, ∴ wires ge super hot and burn
Corner frequencies:
● Corner frequencies (ωc)
○ Non-linear equivalent of cut-off frequencies
2 2
○ Roots of denominator of transfer function (ie. roots of 𝑠 + 2ω𝑛ζ𝑠 + ω𝑛 )
■ Only true for underdamped system ζ < 1
● If roots of H(s) are complex → underdamped system
● RC circuits act like low-pass filter
● nth order system → n corner frequencies
● Corner frequencies have compound effect
○ For ω > ωc2 > ωc1 effect = ωc2 + ωc1
Roll rate:
● dB/dec meaning dB per ω
○ ie. amount of attenuation (deamplification) as input frequency increases)
● -20dB/dec for 2nd order systems
Bode plot:
● Amplification (GdB) vs. input frequency (ω)
● If Ks=0, amplification is 0
● Cut-off frequency is as -3dB from steady state
● Corner frequencies are roots were graph slopes changes
● Low-pass filter on right
○ High frequencies get most attenuated
● High pass filter is mirror image
○ Low frequencies get most attenuated
● note: In real life, changes are less sudden (ie. smoothed out curve)
Active filters:
● Filters that use OpAmps
● Advantages
○ Don’t add electrical load (no impedance)
○ Connected in series for multiple order (ie. more than one corner frequency)
 ○ Can filter and amplifying at same time
● Recall: as s increases ⇒ means ω increases
● Most measurement systems act as low-pass filters
● note: filter frequency means cut-off frequency

Bandwidth:
● Range of frequency where no attenuation
● Types of filters
○ Band-pass: allows signal between 2 frequencies
○ Band-stop: stops signal between 2 frequencies
○ Notch filter: removes 1 frequency

Spectral analysis
● Time-domain to frequency-domain
● Gives decomposition of signal into individual sinusoidal waves
● All time-varying signals can be approximated as infinite sum of sinusoidal waves
○ Square wave → Σodd frequency sines
○ Triangle wave → Σodd frequency sines with alternating negatives
○ Sawtooth wave → Σeven frequency sines
● Typically use logarithmic scale for frequencies

Sampling frequency
● Real life is continuous signals
● Computers is discrete signals (ie. data points)
● ∴need to convert from analogue to digital
Sampling frequency (fs):
● Rate at which data point recorded
Aliasing error:
● Incorrect identification of signal due to discretization
● Nyqyuist criterion
○ Used to prevent aliasing error
○ 𝑓𝑠 > 2𝑓𝑚 (typically 2.5 in industry)
 ○ Nyquist frequency (fN):
1
■ 𝑓𝑁 = 2
𝑓𝑠
■ Ideally: 𝑓𝑁 > 𝑓𝑚
○ Nyquist rate = 2𝑓𝑚 (minimum fs satisfying nyquist criterion)
○ Most measurement take some time
● Need to make sure input frequencies are less than cut-off frequency
Experimentally obtaining frequency system response:
● Often conducted on entire system
● For known input measure output
○ Convert both signals to frequency domain
○ Find transfer function (H)
○ Make bode plot: amplification (= |𝐻(𝑠)|) vs. frequency
■ note: s-domain is frequency domain (since 𝑠 = 𝑖ω = 𝑖2π𝑓)
Sampling error:
● Alising error
● Response of measurement system is too slow → bandwidth too small
○ Higher bandwidth means can support faster frequencies in input

Analogue to digital converted (ADC)

● Discretization of data is both in time and amplitude
○ Time discretization is aliasing error
○ Amplitude discretization is quantization error
● Amplitude discretization
○ Input range
○ Bit count (⇒resolution of data)
Bit count:
● Digital signal is binary
● Bit-count is #of binary digits
𝑁
● #𝑜𝑓 𝑜𝑢𝑡𝑝𝑢𝑡𝑠 = 2 where N is bit count
ADCs:
● Use simultaneous OpAmp comparators to compare input voltage to pre-set voltage
○ Used with voltage divider to compare voltage multiple times
Saturation:
● ADC has input range (based on voltage of OpAmps)
○ Otherwise voltage is too large ⇒ OpAmp saturated
● Saturation limit (Vd, max)
○ When input voltage (Vi) is larger than range of outputs
■ Largest output is outputted
𝑁 −
○ 𝑉𝑑, 𝑚𝑎𝑥 = 𝑉𝑟 (2 − 1) + 𝑉𝑎 where Va─ is lower bound of range of device (ie. minimum
output)
Resolution:
● Step voltage between each OpAmp
● Certain small range of voltages all get encoded as single voltage
○ More OpAmps with same range ⇒ greater resolution

●
Quantization error:
● ε𝑞 = 𝑉𝑑 − 𝑉𝑖 where Vd is digitized output, Vi is analogue input at sampled point
○ But we don’t usually know Vi since that is the point of digitalization
● Max error is bounded by ½ resolution
𝑉𝑟
○ 𝑚𝑎𝑥(ε𝑞) =± 2
where Vr is resolution voltage
Conversion rate:
● Rate at which data is converted (from analogue to digital)
○ ie. time it takes to run through system
○ Typically <100ns (=10MHz)
● 2 types of conversion
○ Simultaneously
■ All channels sample at same time (expensive)
○ Sequentially
■ Each channel samples one after another
● Rate can be quoted for
○ All channels
○ Per channel
● Usually taken sequentially, therefore ↗range ⇒ ↗channels(comparators/OpAmps) ⇒ ↗time to
cycle through each one ⇒ ↗conversion rate

Strain gauges
● Longer wire → more resistance in wire
● Gauge factor
(δ𝑅/𝑅)𝐴
○ 𝐺𝐴 = ε𝐴
, ε𝐴 is strain

○ Typically: 𝐺𝐴 = 2. 1
○ Changes in resistance are very small
● Quarter bridge
𝑉𝑜 𝐺𝐴ϵ𝐴
○ 𝑉𝑠
≈ 4
, given k=1 for wheatstone bridge
○ Quarter bridge - type 1
■ 1 strain gauge used
○ Quarter bridge - type 2
■ Two strain gauges used
■ Compensate for temperature fluctuations
● Half bridge
 ○ 2 strain gauges
○ Half bridge - type 1
■ 2 strain gauges perpendicular (same surface)
𝑉𝑜 𝐺𝐴ϵ𝐴(1+𝑣)
■ 𝑉𝑠
≈ 4
, given k=1 for wheatstone bridge
○ Half bridge - type 2
𝑉𝑜 𝐺𝐴ϵ𝐴
■ 𝑉𝑠
≈ 2
, given k=1 for wheatstone bridge
■ 2 strain gauges parallel (different surface)
○ Half bridge - type 3
■ 2 strain gauges parallel (same surface)
● Full bridge
○ 4 strain gauges
○ More sensitivity
○ Temperature compensation
○ Full bridge - type 1
𝑉𝑜 𝐺𝐴ϵ𝐴
■ 𝑉𝑠
≈ 1

■ 2 sets of parallel gauges

○ Full bridge - type 2
𝑉𝑜 𝐺𝐴ϵ𝐴(1+𝑣)
■ 𝑉𝑠
≈ 2

■ 2 sets of perpendicular gauges

● Other methods of measuring strain exist
○ Optical sensing
○ Capacitive sensing
○ Photoelastic properties

Measuring displacement
Potentiometers:
● Adjustable length → increase resistance
● Voltage divider with adjustable wire length (∴resistance)
● Can be
○ Linear
𝑅2 𝑥
■ (𝑉2 =)𝑉𝑜 = 𝑉𝑠 𝑅1+𝑅2
= 𝑉𝑠 𝐿

○ Rotary
θ
■ 𝑉𝑜 = 𝑉𝑠 2π
● Rely on physical contact ∴wear and tear → limited life-span
Capacitive displacement sensor:
𝐴
● Capacitance (C): 𝐶 = ϵ𝑜ϵ𝑟 𝑑
, where A is area of plates and d is distance between plates
● Increase geometry between plates → changes capacitance → changes voltage
 ● 2 options
○ Vary d (distance between plates)
1
■ Gives 𝑦 = 𝑥
relationship (non-linear)
○ Vary A (overlap of plates)
■ Gives 𝑦 = 𝑥 relationship (linear)
● Best for very small displacements
○ Since bigger distance → more electrical power required
● Type of (dielectric) material affect sensitivity (to physical changes)
Linear variable differential transformer (LVDT):
● Uses inductors (∴magnetism)
● All solenoids have same #of turns
● Parts
○ 3 solenoids (1 powered, 2 unpowered linked)
○ Movable ferromagnetic core
● Displacement → output voltage (between connected solenoids)
○ Linear relationship (for whole core inside solenoid region)
○ Non-contact device
● Can be in circular path as well (RotaryVDT)
● Output voltage depends on input voltage
Photonic distance sensors:
● Light intensity based
○ Light displacement transducer
● Time-of-flight based
○ Light waves
■ LIDAR
■ RADAR
○ Sound waves
■ SONAR
Sensor vs. transducer:
● Transducer:
○ Converts energy to new form (eg. mechanical → electrical)
○ Often describes entire device
● Sensor:
○ Detects and responds to physical quantity (eg. strain → resistance)
Photonic transducer:
● Displacement from light source → affects light intensity
○ Also depends on surface finish (therefore require calibration)
● Sensitivity
○ Change in intensity → change in voltage
○ note: There is also a time delay
● Types:
○ Photo-resistor
 ■ Changes resistance
■ Cheap and simple
○ Phototransistors
■ Creates current
■ High sensitivity, non-linear, sensitive to temperature
○ Photodiode
■ Creates low current
■ Robust, high bandwidth
■ note: Solar panels are made of these
● Limited range (requires light hitting sensor)
○ ∴works for very short range (often detects sudden change in conditions)
○ eg. proximity sensors, path tracking (for robots)
1
● Light intensity (J): 𝐽 = 2
𝑥
● Gaussian beam profile

○
Lasers:
● Laser have:
○ Gaussian beam profile
○ Focused light
● Relate light spread → distance
𝑑−𝑑𝑜
○ θ = 2 𝑎𝑟𝑐𝑡𝑎𝑛( 2𝑥
) (beam divergence angle), do is initial beam diameter, d is beam
diameter, x is distance
● High spatial resolution
○ More detail for smaller grid cell

LIDAR:
● Stands for: Light Imaging Detecting and Ranging
● Time for light to travel → distance
○ Uses speed of light ∴ fast response time
● Uses light waves (high frequency waves)
● Spatial resolution: couple cm
● 2D LIDAR
○ Fast rotating LIDAR emitter
○ Gives x and y positions
RADAR:
● Stands for: RAdio Detection and Ranging
 ● Time for light to travel → distance
○ Uses speed of light ∴ fast response time
● Uses radio waves (frequencies between 3kHz to 1 GHz)
● Spatial resolution: couple meters
○ ∴hard to know what object is
● RADAR tower rotates
SONAR\SODAR:
● Stands for: SOund Navigation and Ranging (water)
● Stands for: SOund Detection and Ranging (air)
● Time for light to travel → distance
○ Uses speed of sound
○ note: speed of sound is way faster in water
● SONAR is often used in water since nature of sound waves
○ Light waves are easily scattered by organic material

Measuring velocities (of solids)

● Two main velocities to measure:
○ Linear
■ Use LIDAR, RADAR, SONAR
○ Rotary (rotational)
■ Use Tahometers (magnetic pickup, laser, stroboscopic)
■
● Velocity measures distance and time
○ Time resolution is important (closer in time distance measurements)
Doppler effect:
● Due moving source of sound waves
● Change in frequency of observed waved
○ Compressed in-front (higher frequency)
○ Spaced out behind (lower frequencies)
RADAR\LIDAR\SONAR doppler:
● RADAR
○ Uses radio waves (low frequency)
○ ∴good for objects at distance
○ ∴low spatial resolution (hard to tell what you are measuring if multiple moving objects)
● LIDAR
○ Uses light waves (high frequency)
○ ∴bad for objects at distance
○ ∴high spatial resolution
● Pulse of waves are sent out, hit object → compresses waves (extent varies with speed)
○ Measure how much waves are compressed
𝑓𝑑 𝑐
● 𝑣= 2𝑓𝑠 𝑐𝑜𝑠(θ)
 ○ fd is doppler (received) frequency, fs is sample (emitted) frequency, c is speed of light, θ is
angle between line of action of car and radar gun (assumed to be θ = 1)
● Relative error:
δ𝑣 δ𝑓𝑑 δ𝑓𝑠
○ ϵ𝑟𝑒𝑙𝑎𝑡𝑖𝑣𝑒 = 𝑣
= 𝑓𝑑
− 𝑓𝑠

○ Assumes small angle (θ = 1)

General calculation for generalized uncertainty:
𝑛1 𝑛 𝑛𝑖
● For any expression of form: 𝑦 = 𝑎 𝑥1 𝑥2 2 ... 𝑥𝑖
● Relative uncertainty:
δ𝑦 δ𝑥𝑖
○ ϵ𝑟𝑒𝑙𝑎𝑡𝑖𝑣𝑒 = 𝑦
= Σ𝑛𝑖 𝑥𝑖

Tachometers:
● Measures angular velocities
● 3 types
○ Magnetic pickup
○ Stroboscopic
○ Laser
● Units:
○ SI: [rad/s]
■ Denoted: ω
○ [RPM] rotations per minute
■ Denoted: NRPM
𝑟𝑎𝑑 2π
○ Conversion: [ 𝑠
]= 60
[𝑅𝑃𝑀]
Magnetic pickup:
● Detects strength of magnetic field
○ Strongest if metal is closest to sensor
○ ∴need non-round object
● Measures time between peaks
○ Need to specify #of teeth per rotation (or other irregularities)
Stroboscopic:
● Determines rotational speed visually
○ Sampling frequency is adjusted (by hand)
● Stroboscopic effect
○ Caused by aliasing error (incorrect sampling frequency)
○ eg. oscillating/rotating objects appears stationary
■ Since sampling frequency = frequency of oscillation/rotation
● Determines when image is stationary
○ ∴rotational frequency = sampling frequency
○ Sampling frequency is fixed (and known)
● Drawbacks
○ Harmonics (strobing at half of frequency)
■ ie. Once every 2 full cycles
 ○ Human error
○ Must visually see object
Laser tachometer:
● Detects changes in constant input signal
● Has
○ Laser emitter
○ Laser receiver
● 2 types
○ Reflective
■ Measures changes in distance (short range)
○ Through-type
■ Detects obstructions

Measuring Force, Torque, Vibrations

● Accelerometers
● Load-cells
● Force/torque transducers
● Accelerometers
● Static and dynamic torque
● note: Often are 2nd order devices
Accelerometers:
● Acceleration
○ Measure from distance-time graph
■ Use displacement measurements (LIDAR, RADAR, SONAR)
𝐹
○ Measure form force: 𝐹 = 𝑚𝑎 ⇒ 𝑎 = 𝑚
■ If known force and mass
■ ∴need to measure forces (indirectly by effect of forces)
● Resistive and capacitive strain gauges
● Piezoelectric sensors
○ Directly convert electric charge → mechanical strain
● Devices use 𝐹 = 𝑚𝑎
● Proof mass
○ Known mass suspended in device housing (using elastic element)
○ (when device accelerates) Elastic element defect → induced force on proof mass
■ This force can be measured ⇒ ∴acceleration determined
● Strain gauge accelerometers
○ Cantilever setup (with strain gauges)
○ Acts as spring-mass system
𝑘𝑥
■ 𝑎= 𝑚
strain gauges measure x
○ Typically function at <50Hz
● Piezoelectric sensors
○ Convert mechanical strain to electric charge
Load-cells:
● Load cell is a thin material with a strain gauge attached
● To increases sensitivity: force → ↗strain
○ Sensitivity inversely proportional to Young's modulus & area
● Z or S type
○ Removes material and makes load cell less stiff
Torques and moments:
● Moment: Linear load acting perpendicular causing rotation
● Torque: Angular load
○ Static/stall torque: no rotation
○ Dynamic torque: rotating shaft (𝑃𝑎𝑛𝑔𝑢𝑙𝑎𝑟 = 𝑇ω)
■ Measure power + rotation speed (indirect measurement)
● Dynamic torque measurements:
○ Dynamometer
■ Measures force (∴power and torque)
■ Resists rotary motion (dissipates energy)
○ AC generator
■ A dynamometer resisting motion using magnetic resistance
■ Amount of electric power generator
● Related directly to power of motor
○ Torque transducer
■ Measures torque directly
■ No resistance to rotary motion
■ 2 strain gauges mounted at 45o
● Gives shear strain → shear stress
τ𝐽
● 𝑇= 𝑟
J is polar moment of inertia, r is radius, τ is shear stress

System response:
● 2nd order systems
𝑘 𝐹
○ Natural frequency: ω𝑛 = 𝑚
= δ𝐿·𝑚
where k is stiffness
𝑐
○ Damping ratio: ζ = 2𝑚ω𝑛
where c is damping ratio (depends on physical properties of
system such as damping fluid density)

Measuring Fluid Pressure

Pressure:
● Fluid mechanics equivalent of stress
𝐹
● 𝑃= 𝐴
● Types of pressure
○ Absolute pressure (Pabs): relative to vacuum
○ Gauge pressure: relative to atmospheric pressure
Manometers:
● Hydrostatic pressure
○ Pressure cause by fluids (eg. air, water)
■ ∆𝑃 = ρ𝑔∆ℎ where ρ is density, g is gravity, h is distance
■ Only vertical height matters
● Manometers
○ Pressure measuring device using liquid displacement
○ If liquid density is known, change in pressure can be determined
○ Typically with reference to atmospheric pressure (since this is 0 point)
● Barometer
○ Measure atmospheric pressure
○ Uses mercury (Hg)
○ Compares atmospheric pressure to vacuum (P=0)
● Sensitivity:
𝑚𝑒𝑎𝑠𝑢𝑟𝑎𝑛𝑑 δ∆𝐿 1
○ 𝑑𝑒𝑠𝑖𝑟𝑒𝑑 𝑞𝑢𝑎𝑛𝑡𝑖𝑡𝑦
= δ∆𝑃
= ρ𝑔𝑠𝑖𝑛(θ)
since ∆𝑃 = ρ𝑔∆𝐿𝑠𝑖𝑛(θ)
○ ∴ ↘angle(θ) and ↘density (ρ) → ↗sensitivity
● Limitations
○ Bad for dynamic measurements
■ Due to high damping of liquid
● note: Many other mechanical pressure sensors exist
Strain gauge pressure transducer:
● Chamber with two openings separated by diaphragm
○ Diaphragm defects due to difference in pressure
○ Deflection is symmetric
■ Since pressure is uniform
● 2 strain gauges measure deflection
○ In ½ bridge configuration
○ Strain → stress → force → pressure (known area)
● Bad for low pressures
Capacitive pressure transducer:
● Use capacitor plate
Frequency response:
● 2nd order systems
● ωn is related to thickness and material of diaphragm
○ ω𝑛 = 𝑘/𝑚
● Slow response time
○ Measurement device is often physically removed from pressure site
○ Sensing line must be used
■ Caries pressure to different location
● Fluid used in sensing line determines response (not transducer itself)
○ Response of sensing line only:
𝑎
■ 𝑓𝑛 = 4𝐿
where a is speed of sound in fluid, L is length of sensing line
 ○ Response of entire system:
𝑎
■ 𝑓𝑛 = 𝑉 1/2
where Vt is volume of transducer, Vs is volume of sensing line
2π(0.5+ 𝑉𝑡 )
𝑠

○ ↗natural frequency → ↗measurement range

● Damping ratio:
16𝑣𝐿 𝑉𝑡 1/2
○ ζ= 2 (0. 5 + 𝑉𝑠
) where v is kinematic viscosity (15*10-6 of air), d is diameter of line
𝑎𝑑
● note: ↗natural frequency → ↘damping ratio (for pressure transducers)

Measuring temperature
Types:
● Thermocouples
● Resistance temperature devices
● Thermistors
● Infrared thermometers

Thermocouples:
● Two different metals of different temperatures
○ Induces voltage
○ Electrons move hot → cold
○ Rate of electron movement depends on material
● Measures relative temperature
○ Hot junction (source) vs cold junction
● There are different types (use different metal pairs) based on temperature range
○ Have different sensitivities
● Advantages
○ Inexpensive
○ Simple and robust
○ Wide operating range
○ Sensitive
● Disadvantages
○ Linear
○ Accuracy (±1oC)
○ Measured voltage with no current flow
○ Needs voltages measuring device
Resistance temperature devices (RTD):
● ↗temperature → ↗resistance
○ ↗vibrations → ↗collisions → ↗resistance
○ Temperature induced strain (longer wire → ↗resistance)
■ Uses strain gauge setup
● Used in ¼ bridge
● Advantages
○ Accurate
○ Linear output response
○ Measures temperature directly
○ Stable over time
● Disadvantages
○ Bulky and expensive
○ Poor spatial resolution
○ Slower response
Thermistors:
● Uses semiconductor
● 2 types
○ Positive temperature coefficient (PTC)
■ ↗T → ↗R
■ For circuit protection
○ Negative temperature coefficient (NTC)
■ ↗T → ↘R
■ For temperature sensor
● NTC thermistors
○ More energy (T) → more conductive material
● Advantages
○ High sensitivity
■ Sensitivity drops at high temperatures
○ Doesn’t age
● Disadvantages
○ Strong non-linearity
○ Needs calibration
○ Low temperature range
○ Self-heating if high current
○ High time constant (τ = 1𝑠)
Infrared thermometers:
● Non-contact
● Detect black-body radiation
○ IR for room temperature
○ Visible light above 500oC (this is why hot things are red)
● Advantages: Very fast measurements
● Disadvantages: Only measures surface temperature
System response of temperature sensors:
● Based solely on rate of heat transfer between medium and device
𝑑𝑇𝐶𝑉
○ Governed by: 𝑞 = 𝑚𝐶 𝑑𝑡
where TCV is temperature of control volume (device)
○ ∴ 1st order system
● Time constant:
𝑚𝐶
○ τ= ℎ𝐴
where m is mass, C is specific heat capacity, h is heat transfer coefficient, A is
surface area

Measurement of fluid flow/velocity

Types of devices:
● Positive displacement flowmeters
○ Venturi tube
○ Flow nozzle
○ Orifice plate
● Pitot-static tube
● Hot-wire anemometry
Positive displacement flowmeters:
● Measures pressure difference by (temporarily) constraining pipe
○ An average velocity of flow in tube
● ↘area → ↗velocity → ↗dynamic pressure
● 𝑄 = 𝑉𝐴 where Q is volumetric flow rate, V is velocity, A is area
2[(𝑃1+ρ𝑔ℎ1)−(𝑃2+ρ𝑔ℎ2)] 𝐷2
● 𝑄 = 𝐶𝐴2 2 C is discharge coefficient (depends on diameter ratio β = 𝐷1
ρ[1−(𝐴2\𝐴1) ]
1 2
○ 𝑑𝑦𝑛𝑎𝑚𝑖𝑐 𝑝𝑟𝑒𝑠𝑠𝑢𝑟𝑒 = 2
ρ𝑉
○ ℎ𝑦𝑑𝑟𝑜𝑠𝑡𝑎𝑡𝑖𝑐 𝑝𝑟𝑒𝑠𝑠𝑢𝑟𝑒 = ρ𝑔ℎ (usually =0)
● Venturi tubes
○ Build according to specific standard
○ Cause lest fluid energy loss
○ Expensive
● Flow nozzle
○ Simpler venturi tube ∴less expensive
○ More energy loss
■ Due to turbulent conditions at outlet
● Orifice plate
○ Least expensive and modular
○ Single point of reduced diameter
■ Smallest diameter occurs past physical point
○ Largest loss of energy
■ Turbulent flow at front and back
Pitot-static tube:
● Measures fluid velocity at single point
○ Increase spatial resolution
● Measures velocity by isolating dynamic pressure
1 2
○ 𝑃𝑑𝑦𝑛 = 2
ρ𝑉
○ 𝑃𝑡𝑜𝑡𝑎𝑙 = 𝑃𝑠𝑡𝑎𝑡𝑖𝑐 + 𝑃𝑑𝑦𝑛𝑎𝑚𝑖𝑐
○ Note: static and hydrostatic are equivalent
● Measures static pressure and total pressure
○ Static pressure is taken perpendicular to fluid flow
■ ie. No fluid motion
● Need to determine density (ρ)
○ For ideal gas, depends on temperature and pressure
Hot-wire anemometer:
● Also called Constant temperature anemometer (CTA)
● Principal: flowing fluid → cools wire → lowers resistance → requires power to heat up
● Working details
○ Thin wire is heated (using electric resistance)
○ Fluid flow cools wire ∴lowers resistance
○ Change in resistance is measured using wheatstone bridge
■ ¼ bridge configuration
● Static calibration required
○ Determine amount to voltage required to keep resistance steady with velocity
■ Measure using pitot-static tube
● Often wire is plated
○ Reduces heat reduction (due to less area)
○ Accurate sensing lengths
○ More rigid design