PLC Analog Input Scaling and Calibration: Formula, Code and Field Proof

PLC analog input scaling in one answer

PLC analog input scaling maps the number delivered by an input channel onto the instrument's engineering range. For two known endpoints, use:

Scaled = ScaledLow
       + (Raw - RawLow) × (ScaledHigh - ScaledLow)
       / (RawHigh - RawLow)

For an S7-1200 channel configured so that 4–20 mA is represented by raw 0–27,648, and a pressure transmitter ranged −10 to 90 bar, raw 13,824 is halfway through both spans:

Pressure = -10 + (13,824 - 0) × (90 - (-10)) / (27,648 - 0)
         = -10 + 0.5 × 100
         = 40 bar

That mathematical result is necessary, but it is not proof that the installed loop measures correctly. A complete job also confirms the exact module, channel range and data format; preserves under-range and over-range evidence; checks numeric types; compares several injected or applied reference points; records as-found and as-left results; and verifies what the HMI, alarm and control logic actually consume.

Critical distinction: scaling is a software/data transformation. Calibration compares an indication with a reference under stated conditions and may lead to an authorized adjustment. Changing PLC endpoints until one field reading “looks right” can hide a transmitter, wiring, module, range or reference problem.

Use the free analog scaling calculator to check the arithmetic in either direction. Then use this guide's five-point loop-check worksheet to prove the implemented channel rather than trusting a single midpoint.

What this page owns

This is the scaling-and-calibration implementation owner. Adjacent pages retain their narrower jobs:

Need	Use this specialist	Ownership boundary
current-loop wiring and electrical behavior	4–20 mA current-loop guide	loop power, burden, polarity and live-zero fundamentals
analog signal/module selection	PLC analog I/O signals	current versus voltage, resolution, isolation and module fit
failed installed channel	PLC I/O troubleshooting	first-disagreement diagnosis across field and controller
transmitter adjustment procedure	transmitter loop calibration	instrument-specific reference, adjustment and certificate workflow
reusable arithmetic	PLC math instructions	numeric operations beyond analog channels
quick independent calculation	analog scaling calculator	endpoint conversion, not physical calibration or target validation

Define the six endpoints before writing code

“A 4–20 mA input is 0–32,767” is not an engineering specification. The current range, module representation and transmitter range are three different contracts. Document all six endpoints and their units.

Layer	Low endpoint	High endpoint	Example	Source of truth
physical measurand	PVLow	PVHigh	−10 to 90 bar	approved instrument data sheet / P&ID / range record
loop signal	SignalLow	SignalHigh	4 to 20 mA	transmitter configuration and receiver range
controller representation	RawLow	RawHigh	0 to 27,648	exact module, channel setting and selected data format

Also record the tag type, input update rate/filter, module diagnostics, PLC task period, HMI unit/precision and downstream consumers. The same part number may support raw/proportional counts, engineering units, percent or PID-oriented formats. A copied raw endpoint can therefore be wrong even when the hardware family is correct.

Siemens' S7-1200 system manual, for example, documents raw 0 at 4 mA and 27,648 at 20 mA for the 4–20 mA current range. The same table documents different electrical meanings for a 0–20 mA range. Rockwell's 1769 documentation exposes selectable Raw/Proportional, Engineering Units, Scaled-for-PID and Percent Range formats. Treat format as part of the interface—not a cosmetic display choice.

Endpoint manifest

Field	Example value	Why it matters
loop ID / channel	PT-204 / Local AI 3	connects code to drawing and field label
transmitter range	−10…90 bar	defines the physical span and units
signal range	4…20 mA	defines electrical endpoints and live-zero behavior
module and revision	SM 1231, exact order number	resolves input type, resolution and diagnostics
channel configuration	current, 4…20 mA	distinguishes it from 0…20 mA or voltage
data representation	raw integer	decides what the PLC tag means
raw nominal endpoints	0…27,648	formula inputs for this exact configuration
diagnostic endpoints	module-specific	keeps faults separate from valid process values
PLC source tag/type	%IW... : INT	prevents address and conversion mistakes
scaled tag/type	PT204_PV : REAL	preserves fractional engineering values
update/task period	recorded values	supports filter and stale-data reasoning
range authority	approved instrument record	prevents casual HMI re-ranging from changing control

Download the analog-channel contract (CSV) and fill it before copying any instruction block.

Derive the two-point formula instead of memorizing it

Two points define a straight line. Let x be the input/raw value and y the engineering value:

y = m × x + b
m = (ScaledHigh - ScaledLow) / (RawHigh - RawLow)
b = ScaledLow - m × RawLow

The endpoint form used at the top of this guide is algebraically identical, but it is easier to review because all four configuration constants remain visible.

Worked example: 4–20 mA, −10 to 90 bar

Assumptions are explicit:

• the installed transmitter is approved and configured for −10 to 90 bar;
• the exact S7-1200 input channel is configured for 4–20 mA;
• its representation table maps nominal 4 mA to raw 0 and 20 mA to raw 27,648;
• the source tag is read as an integer, then converted to floating point before division; and
• out-of-range quality is evaluated separately from the calculated value.

Test point	Applied current	Expected raw	Span fraction	Expected pressure
0%	4.000 mA	0	0.00	−10.00 bar
25%	8.000 mA	6,912	0.25	15.00 bar
50%	12.000 mA	13,824	0.50	40.00 bar
75%	16.000 mA	20,736	0.75	65.00 bar
100%	20.000 mA	27,648	1.00	90.00 bar

At the 75% point:

Pressure = -10 + (20,736 - 0) × 100 / 27,648
         = -10 + 75
         = 65 bar

The negative lower range makes a common bug visible: scaling only by span (raw × 100 / 27,648) gives 75, not 65, because it omits the −10 bar offset.

Reverse and non-zero ranges

The same formula supports 100 to 0%, 20 to 4 mA, −50 to 50 °C or any other linear endpoint pair. Do not force ScaledHigh > ScaledLow; reverse-acting ranges can be intentional. Do reject a zero raw span, because division by zero means the configuration contract is invalid.

Range shape	Valid?	Review point
raw 0…27,648 → 0…100%	yes	ordinary rising scale
raw 0…27,648 → −10…90 bar	yes	non-zero engineering offset
raw 0…27,648 → 100…0%	potentially	confirm reverse action is intentional
raw 1,000…1,000 → 0…100%	no	zero denominator; flag configuration fault
raw −27,648…27,648 → −10…10 V	yes	bipolar range and signed type required

A reusable Structured Text scaling block

The following IEC-style function separates calculated value, limited value and quality. Adapt types, conversion functions and diagnostic inputs to the exact target; it is a behavioral reference, not a drop-in certified library.

FUNCTION_BLOCK FB_AnalogScale
VAR_INPUT
    RawValue       : DINT;
    RawLow         : DINT;
    RawHigh        : DINT;
    EULow          : LREAL;
    EUHigh         : LREAL;
    UnderLimitRaw  : DINT;
    OverLimitRaw   : DINT;
    ModuleFault    : BOOL;
    ClampOutput    : BOOL;
END_VAR
VAR_OUTPUT
    ValueEU        : LREAL;  // extrapolated diagnostic value
    LimitedEU      : LREAL;  // optional bounded consumer value
    Good           : BOOL;
    UnderRange     : BOOL;
    OverRange      : BOOL;
    ConfigFault    : BOOL;
END_VAR
VAR
    Fraction       : LREAL;
    LowBound       : LREAL;
    HighBound      : LREAL;
END_VAR

ConfigFault := RawHigh = RawLow;
UnderRange  := RawValue < UnderLimitRaw;
OverRange   := RawValue > OverLimitRaw;
Good        := NOT (ConfigFault OR ModuleFault OR UnderRange OR OverRange);

IF NOT ConfigFault THEN
    Fraction := DINT_TO_LREAL(RawValue - RawLow)
                / DINT_TO_LREAL(RawHigh - RawLow);
    ValueEU := EULow + Fraction * (EUHigh - EULow);

    LowBound := MIN(EULow, EUHigh);
    HighBound := MAX(EULow, EUHigh);
    IF ClampOutput THEN
        LimitedEU := LIMIT(LowBound, ValueEU, HighBound);
    ELSE
        LimitedEU := ValueEU;
    END_IF;
ELSE
    ValueEU := EULow;
    LimitedEU := EULow;
END_IF;
END_FUNCTION_BLOCK

Download the IEC-style scaling oracle (ST). Compile or translate it on the selected target, because IEC 61131-3 establishes language concepts while conversions, overloads, MIN/MAX/LIMIT availability and diagnostic interfaces remain implementation-specific.

Numeric implementation failures

Failure	Symptom	Mechanism	Prevention
integer division before conversion	output remains near low end, then jumps	fractional result is truncated	convert numerator/denominator to REAL/LREAL before division
intermediate integer overflow	wrong sign or implausible value	multiplication exceeds source type	convert before subtract/multiply; inspect target overflow behavior
wrong signedness	high input appears negative	16-bit word interpreted incorrectly	match module-defined type and representation
stale raw tag	smooth but delayed/frozen PV	mapping or update path is wrong	capture module tag, copied tag and timestamp/change counter
rounding each stage	systematic steps	repeated conversion loses resolution	keep full precision; round only at presentation boundary
hidden clamping	flat value at limits	evidence is discarded before diagnosis	retain extrapolated value and quality separately
duplicate scaling	correct at one point, wrong elsewhere	module already provides engineering units	record data format and scale exactly once
copied constants	fixed span/offset bias	endpoints belong to another module/range	bind constants to channel contract and review version

Vendor implementation boundaries

Siemens S7-1200 / S7-1500

The S7-1200 manual defines NORM_X as (VALUE - MIN) / (MAX - MIN) and SCALE_X as VALUE × (MAX - MIN) + MIN. It also documents extrapolation behavior outside the nominal range. A typical flow is:

Normalized := NORM_X(MIN := RawLow, VALUE := RawInput, MAX := RawHigh)
Pressure   := SCALE_X(MIN := -10.0, VALUE := Normalized, MAX := 90.0)

Select the block types deliberately. Preserve the input module's diagnostics and 7FFF/overflow behavior instead of treating every integer as a valid process sample. The exact representation depends on module/order number and selected range.

Rockwell SLC / MicroLogix and Logix families

Rockwell's SLC 500 reference documents SCP (Scale with Parameters) for SLC 5/03, 5/04 and 5/05, while legacy manuals also document SCL. That does not make SCP a universal built-in Studio 5000 Logix instruction. For a Logix controller, use the exact controller/instruction set, module-supplied engineering format, reviewed math or a governed Add-On Instruction as appropriate.

Rockwell 1769 analog modules can expose several data formats, and their under-range/over-range channel bits are valuable evidence. Resolve the actual tag definition and channel configuration before deciding whether software scaling is required.

CODESYS, Schneider, Beckhoff and other IEC targets

Use the same endpoint contract, but confirm each target's conversion, overflow, LIMIT, function-block call and retained-value semantics. Do not label generic ST “vendor neutral” and assume it compiled unchanged. Record the compiler, runtime, library and task used to verify the translated block.

Platform decision	Evidence to capture
native normalize/scale instructions	exact help/manual version and parameter types
module engineering-units format	channel configuration and actual tag meaning
reusable custom function block	source version, call order, instance ownership and tests
Add-On Instruction / library block	revision, signature, change approval and deployed target
manual ladder arithmetic	execution order, intermediate types, overflow and divide-by-zero behavior

Scaling is not calibration

NIST distinguishes calibration, adjustment and verification and emphasizes that traceability belongs to a measurement result supported by a documented chain and uncertainty—not merely to an instrument with a sticker. In plant work, terminology and required records depend on the quality system, procedure, regulator and instrument.

Activity	Question answered	May change something?	Minimum evidence
scaling configuration	what EU corresponds to this controller value?	yes, program/configuration	approved endpoints, version, code review and tests
verification / loop check	does the observed chain meet stated acceptance criteria?	normally no	applied reference, conditions, readings, tolerance and result
calibration	what is the relation between reference values and indication, with associated uncertainty?	not necessarily	method, reference, results, uncertainty/traceability as required
adjustment	should the instrument be altered to reduce indication error?	yes	authorization, as-found data, adjustment and as-left data
trim/compensation in PLC	should software correct a characterized bias?	yes, high governance	approved rationale, bounds, ownership, audit trail and revalidation

Never make an undocumented PLC offset the default response to a bad loop check. First locate the disagreement: process reference, sensor, transmitter, current path, input module, raw representation, scaling, HMI formatting or downstream use.

Five-point field verification

Only trained and authorized people should connect test equipment, open panels, disturb process instruments or work near hazardous energy. Follow site isolation, electrical, process and calibration procedures. A browser exercise cannot authorize or validate field work.

Before applying a reference

1. identify the loop, process consequence and responsible owner;
2. confirm permits, isolation/override and alarm/interlock handling;
3. record module, channel, signal range, data format and current code version;
4. identify whether the test injects current at the PLC, at marshaling, or through the transmitter;
5. record reference instrument ID, range, calibration status and relevant uncertainty;
6. capture as-found raw, scaled, HMI and quality values; and
7. define point-by-point tolerance and restoration criteria before testing.

Example evidence table

The table uses the worked −10…90 bar S7-1200 contract. Acceptance values below are examples only; use the project's approved tolerance and reference uncertainty.

Direction	Applied mA	Expected raw	Observed raw	Expected bar	Observed bar	Error bar	Quality	Result
up	4.000	0	3	−10.00	−9.99	+0.01	good	evaluate to project tolerance
up	8.000	6,912	6,905	15.00	14.97	−0.03	good	evaluate
up	12.000	13,824	13,816	40.00	39.97	−0.03	good	evaluate
up	16.000	20,736	20,742	65.00	65.02	+0.02	good	evaluate
up	20.000	27,648	27,640	90.00	89.97	−0.03	good	evaluate
down	12.000	13,824	record	40.00	record	calculate	record	evaluate hysteresis

At minimum, repeat the low and high endpoints after any adjustment, then restore the process path and prove alarm/interlock/quality behavior under the approved procedure. Do not leave a forced value, bypass, calibrator connection or disabled alarm behind.

Diagnose by injecting at successive boundaries

Injection or observation point	If correct here	If wrong here	Next boundary
physical reference at sensor	process/reference is plausible	reference, impulse line or sensing point suspect	transmitter indication/output
transmitter output terminals	transmitter conversion is plausible	transmitter range, trim, power or sensor path suspect	field/marshalling terminals
PLC-side loop current	cable/current path is plausible	wiring, barrier, isolator, burden or grounding suspect	module channel raw tag
module raw tag	input conversion/configuration is plausible	wrong channel/range/format, module or termination suspect	scaled program tag
scaled tag	formula/constants/type are plausible	software mapping/scaling defect	HMI/historian/controller consumer
HMI/controller value	end-to-end data path is plausible	alias, unit, rounding, stale data or duplicate scaling suspect	output/control response under procedure

This first-disagreement method prevents random endpoint edits. For example, if 12.000 mA is correct at the PLC terminals but the raw tag is at 75%, inspect channel range/data format before touching the transmitter. If the raw tag is correct and the engineering value is biased by exactly 10 bar across all points, inspect EULow rather than wiring.

Preserve diagnostic zones before clamping

A 4–20 mA live zero creates room below 4 mA for under-range and failure signaling, but exact thresholds and meanings are device- and module-specific. Endress+Hauser documentation for one device using NAMUR NE 43, for example, distinguishes under-range at or below 3.8 mA and failure at or below 3.6 mA, with corresponding high-side ranges. That is evidence for that documented behavior—not permission to hard-code the same thresholds for every loop.

Value to expose	Purpose	Example consumers
RawValue	exact module evidence	commissioning and diagnosis
ValueEU	extrapolated engineering value	trend, diagnostics and fault context
LimitedEU	bounded value if a consumer requires it	explicitly designed calculation/control path
Good / quality enum	whether value is usable for its purpose	interlocks, alarming, display decoration
UnderRange / OverRange	directional electrical/process evidence	maintenance alarm and fault history
ModuleFault	hardware/channel diagnostic	first-out and device-health logic
LastGoodEU plus age	continuity aid, never disguised as live	display/history under an explicit stale policy

Do not overwrite ValueEU with zero on every fault and call that safe. Zero may command heat, speed, flow or makeup behavior through downstream logic. Define a per-consumer response: inhibit control, hold last good for a bounded time, transfer to manual, use a validated fallback, trip ordinary control, or request an independently engineered protection action.

Filtering without hiding the fault

A first-order filter can reduce display jitter, but it must use an intentional time constant and task interval:

alpha = dt / (tau + dt)
Filtered = Filtered + alpha × (ValueEU - Filtered)

For a 100 ms task and 2.0 s time constant, alpha = 0.1 / 2.1 ≈ 0.04762. Reusing that coefficient in a 20 ms task changes the real filter. Initialize the state deliberately at startup, call it at a known cadence, and decide what happens on bad quality. Filtering before fault detection can delay or conceal a broken-loop transition.

Design question	Required answer
where is filtering applied?	module, transmitter, PLC, HMI or several layers
what is the effective update interval?	measured/configured input plus task behavior
what dynamics must remain visible?	process event, control response and fault transition
how is startup initialized?	first good sample, configured default or retained state
what happens on bad quality?	freeze, invalidate, reinitialize or controlled fallback
who may change the time constant?	governed role, bounds and audit record

Build a measurement error budget

Scaling arithmetic can be essentially exact while the measurement result remains uncertain. Potential contributors include reference uncertainty, sensor/transmitter performance, ambient effects, installation, loop isolators, input-module accuracy, quantization, noise, sampling, wiring, linearization and repeatability.

For screening only, suppose approved specifications and conditions support independent standard-uncertainty contributions equivalent to 0.10%, 0.15% and 0.05% of span. A root-sum-square screening estimate is:

sqrt(0.10² + 0.15² + 0.05²) = 0.187% of span

A worst-case arithmetic sum would be 0.30% of span. Neither result is automatically the reportable uncertainty: specifications may use different confidence, distributions, reference conditions, denominators and correlations. Convert each contribution consistently and follow the approved uncertainty method. The example only shows why “the transmitter is ±0.1%, therefore the loop is ±0.1%” is incomplete.

Resolution is not accuracy

For raw 0–27,648 mapped across a 100 bar span, one count represents approximately:

100 bar / 27,648 counts = 0.003617 bar/count

That is the scaling increment, not the full installed accuracy. A display with three decimals can show quantization, noise and bias more precisely without measuring the process more accurately.

Acceptance matrix

Case	Stimulus	Expected result	Failure exposed
1	valid low endpoint	exact/within-tolerance low EU, good quality	offset or wrong raw low
2	valid high endpoint	exact/within-tolerance high EU	span or wrong raw high
3	midpoint	correct linear midpoint	slope/offset arithmetic
4	25% and 75%	correct intermediate values	nonlinearity or duplicate scaling
5	raw just below nominal	extrapolated evidence plus configured quality	premature clamp
6	module wire-break diagnostic	bad quality and approved consumer response	diagnostic ignored
7	raw above nominal	over-range retained and alarmed as designed	hidden over-range
8	RawHigh = RawLow	configuration fault; no divide	divide-by-zero
9	reversed EU range	intentional reverse mapping passes	invalid monotonic assumption
10	PLC restart at good input	defined initialization and correct value	stale/retentive filter state
11	quality bad then good	defined recovery and filter reinitialization	stale-value bump
12	input updates slower than task	age/stale behavior is correct	repeated sample mistaken for fresh data
13	range changed under authorization	versioned constants and consumers revalidated	undocumented re-range
14	HMI and historian comparison	same unit/source/quality policy	alias or duplicate conversion
15	alarm setpoint around boundary	correct engineering unit and hysteresis	setpoint left in old range
16	force/simulation removed	live source restored and documented	test state left active

Twelve recurring symptoms and first checks

Symptom	Likely boundary	First evidence to compare
correct at 4 mA, wrong at 20 mA	span/raw high	configured format, raw high and transmitter upper range
same bias at every point	offset/raw low	lower endpoint and EULow
reads 20% at live zero	0–20 versus 4–20 mismatch	channel range and raw representation
reads five times too large/small	engineering span or duplicate conversion	constants and upstream formatted value
negative above midpoint	signed type/word interpretation	module-defined tag type and raw hex/decimal
flat at zero	fault substitution, wrong address or no loop power	raw tag, module status and loop current
flat at upper limit	clamp or overflow sentinel	pre-clamp value and module diagnostics
correct PLC, wrong HMI	display/alias/unit	actual HMI source tag and conversion
value drifts only after warm-up	sensor/module/environment	field reference and component temperatures
noisy raw but smooth HMI	hidden filter	module, PLC and HMI filter settings
slow control response	stacked filters/update rates	timestamped raw/scaled/control trend
calibration passes at PLC injection but fails through transmitter	upstream physical chain	transmitter, reference, impulse/sensor path and wiring

Downloadable implementation pack

• Analog-channel endpoint and ownership contract (CSV)
• Five-point as-found/as-left loop-check worksheet (CSV)
• IEC-style scaling and quality oracle (ST)

The files are templates, not a substitute for the approved device manual, calibration method, change process or safety procedure.

Frequently asked questions

What is the PLC analog input scaling formula?

Use ScaledLow + (Raw - RawLow) × (ScaledHigh - ScaledLow) / (RawHigh - RawLow). Convert to an appropriate floating-point type before division and reject a zero raw span. Keep the four endpoints named and traceable to the exact module configuration and instrument range.

How do I scale 4–20 mA to engineering units?

First determine what raw values the exact configured channel returns at 4 and 20 mA. Then map those raw endpoints to the transmitter's lower and upper engineering values with the two-point formula. Do not assume every 4–20 mA module uses the same counts.

Why does an S7-1200 4–20 mA input start at raw zero?

The S7-1200 representation table documents raw 0 as 4 mA and raw 27,648 as 20 mA when the supported channel is configured for the 4–20 mA range. A 0–20 mA configuration has different electrical meaning. Verify the exact module/order number, firmware and selected range.

Is 5,530 always the Siemens raw value for 4 mA?

No. That value is associated with representing 4 mA as 20% of a 0–20 mA 0–27,648 span in some contexts. Siemens modules that support a configured 4–20 mA range can represent 4 mA as raw zero. Use the representation table for the exact module and range, not a remembered Siemens-wide number.

Does Studio 5000 have a universal SCP instruction?

No. Rockwell documents SCP for SLC 500 families, while Logix implementations depend on the target, module data format, available instructions, reviewed arithmetic or a governed Add-On Instruction. Do not paste an SLC instruction example into a Logix design without checking the target instruction set.

Should I clamp an analog input to its engineering range?

Only when a specific consumer requires a bounded value. Preserve the raw value, extrapolated engineering value and quality flags for diagnosis. If you clamp the only value at the nominal limits, you can hide under-range, over-range and failed-signal evidence.

Is PLC scaling the same as transmitter calibration?

No. Scaling maps defined endpoints in software. Calibration establishes a relation between a reference and indication under stated conditions, with the records/uncertainty required by the applicable system. Adjustment changes the instrument. A scaling edit should not conceal a failed calibration.

How many calibration points should I check?

Use the approved procedure and device requirements. A five-point up-scale check at 0, 25, 50, 75 and 100%, plus selected down-scale points, is a useful engineering verification pattern because it tests endpoints, midpoint, slope and possible hysteresis. It is not a universal regulatory requirement.

Why is my scaled value exactly right at one point but wrong elsewhere?

One matching point cannot prove both slope and offset. Compare low and high endpoints first, then intermediate points. Wrong span, duplicate scaling, incorrect data format and nonlinearity can all cross the expected line at one point.

Should a wire break force the process value to zero?

Not automatically. Zero may be a meaningful process value that drives unsafe or damaging downstream behavior. Mark quality bad and apply an approved response separately for each consumer: inhibit, bounded hold, fallback, transfer or ordinary trip as designed.

How do I choose an analog filter coefficient?

Choose a time constant from process and diagnostic needs, then derive the coefficient from the actual update interval, for example alpha = dt / (tau + dt). Document startup and bad-quality behavior. A coefficient copied between task periods produces a different filter.

Does more ADC resolution mean better accuracy?

No. Resolution describes available increments; accuracy and uncertainty also include reference, sensor, transmitter, module, environmental, installation, noise and repeatability effects. More digits can display a wrong result more precisely.

Can I calibrate a loop with a PLC simulator?

No. Simulation can verify formula, types, states, fault policy and acceptance cases. It cannot source traceable current, characterize an instrument, reproduce module electronics or validate the installed measurement chain.

What must I record after changing an analog range?

Record the approved range authority, instrument and module configuration, endpoint constants, code/configuration version, as-found/as-left evidence, alarm/interlock/setpoint review, HMI/historian units, restoration checks and approver. A range change affects every downstream consumer.

Primary and official sources

1. Siemens S7-1200 System Manual V4.5 — analog representation and NORM_X/SCALE_X
2. Siemens S7-1200 SM 1231 AI 4x13-bit product data
3. Rockwell Compact I/O Analog Modules User Manual, 1769-UM002
4. Rockwell Compact I/O Isolated Analog Modules User Manual, 1769-UM014
5. Rockwell 1769-IF4 module-defined data and channel diagnostics
6. Rockwell SLC 500 Instruction Set Reference — SCP
7. Rockwell SLC 500 Analog I/O Modules User Manual
8. Rockwell MicroLogix 1200/1500 Instruction Set Reference
9. Schneider Electric Analog Input Module scaling reference
10. CODESYS conversion operators
11. CODESYS LIMIT operator
12. Endress+Hauser device manual documenting NAMUR NE 43 ranges
13. NIST metrological traceability policy and FAQ
14. NIST: Traceability Considerations for Characterization and Use of Measuring Systems
15. NIST Engineering Statistics Handbook — calibration
16. NIST Engineering Statistics Handbook — uncertainty approach
17. IEC 61131-3:2025 overview
18. NI raw-data scaling and calibration note

Limitations: formulas, examples, code and worksheets are educational engineering references. They do not establish target compilation, device accuracy, calibration traceability, permissible error, process isolation, electrical authorization, hazardous-area suitability, functional safety or regulatory compliance. Use the exact project, hardware and instrument documents; approved procedures; competent personnel; and controlled field evidence.

Next: verify the endpoints in the analog scaling calculator, study the wider analog I/O signal chain, or practise the scaling and quality logic in the interactive PLC simulator. The simulator is operated by the same team; it does not emulate your target compiler, module electronics, transmitter, reference standard, installed process or safety system.