S7-1200 Temperature Scaling: NORM_X and SCALE_X Error Fix

Problem Summary

On a Siemens SIMATIC S7-1200 CPU 1214C AC/DC/Rly (catalog 6ES7214-1BG40-0XB0), a 0–10 V temperature transmitter is field-scaled to 0–200 °C inside a cyclic OB using NORM_X and SCALE_X. The PLC computes a temperature that disagrees with a hand-held reference thermometer by approximately 10 °C, sometimes drifting further when the process load changes. The block faceplate looks correct - NORM_X between 0 and 27648, SCALE_X between 0 and 200 - so the offset must come from outside the math: the transducer/signal conditioner, the field wiring, the reference instrument, or the CPU analog front-end.

Both NORM_X and SCALE_X are documented in the TIA Portal help and in the S7-1200 system manual at the Siemens Industry Online Support portal (support.industry.siemens.com). Use this article to localize which side of the loop the error is on before re-touching software.

Why ~10 °C Maps to a 0.5 V Offset

Span math is the first sanity check. With a 0–200 °C range and a 0–10 V output, every volt is exactly 20 °C. A 10 °C reading error therefore corresponds to ~0.5 V of systematic error. That magnitude is typical of one of the following:

• A 5 % gain drift on the signal conditioner output stage (e.g., 9.55 V at 200 °C instead of 10.00 V).
• A live-zero transmitter rated 0.5 V to 10 V (or 1 V to 10 V) being scaled as if it were 0 V to 10 V - a cold reading is shifted up by 0.5 V ≈ 10 °C.
• A ground-potential difference between the transducer and the CPU's 0 V analog common that injects a constant bias when load equipment switches.
• A thermocouple input stage with the wrong TC type selected (J vs K, K vs N), which changes the mV/°C slope.

CPU 1214C Onboard Analog Input - Electrical Limits

The CPU 1214C AC/DC/Rly integrates six onboard analog inputs (AI0 through AI5) that are software-selectable for 0–10 V or 0–20 mA mode. For voltage mode the relevant limits from the S7-1200 system manual are:

The transducer signal common must be tied to the same 0 V reference as the CPU analog input common. A floating common picks up the ground-potential difference as a constant offset - exactly the symptom reported.

NORM_X and SCALE_X - Reference Behavior

TIA Portal V13 introduced the IEC-standard scaled-value pair that replaced the legacy SCALE/NORM FBs. Use them as a pair on the S7-1200:

For a 0–10 V, 0–200 °C transmitter the canonical arrangement is:

1. Read %IW64 (CPU 1214C onboard AI channel 0) into an INT/DINT tag.
2. Convert INT → REAL with INT_TO_REAL (or DINT_TO_REAL for the 32-bit field %ID64). Forgetting this conversion is the single most common cause of an "always zero" reading.
3. NORM_X with MIN = 0.0, MAX = 27648.0, VALUE = the REAL raw count.
4. SCALE_X with MIN = 0.0, MAX = 200.0, VALUE = the normalized 0.0…1.0 result.

Likely Root Causes Ranked by Field Frequency

Diagnostic Workflow

Walk through this decision flow with the process offline (cyclic OB stopped or AI channel forced) before touching the program:

Step-by-Step Verification Procedure

1. Inject a known voltage at the AI terminal block. Stop the cyclic OB or use a watch table to monitor %IW64. Disconnect the field wiring and apply a calibrated DC source (e.g., 0.000 V, 5.000 V, 10.000 V) directly to the AI+ and M terminals.
2. Read the raw count. Record %IW64. Apply the formula count ≈ V × 2764.8 (because 27648 counts span 10 V). 0 V should read 0 ± 3 counts, 5.000 V should read 13 824 ± 3 counts, 10.000 V should read 27 648 ± 3 counts.
3. If raw counts match: the CPU scaling block chain is correct. The fault is in the signal conditioner output, the field wiring, or the reference thermometer.
4. If raw counts do not match: the analog front-end is suspect. Move the wiring to a different onboard channel (AI1…AI5) and retest. If a second channel also drifts, evaluate the SM 1231 RTD/AI module as a sanity reference or replace the CPU base unit.
5. Measure at the transducer end. With the field cable reconnected, place a multimeter across the transducer output terminals while the PLC is reading. Walk the value from 0 °C to 200 °C using the calibrator. The multimeter voltage × 2764.8 should track %IW64 within tolerance.
6. Cross-check with a true 0 °C reference. Place the thermocouple in a melting-ice bath (or have the calibrator output 0.000 V / 0.000 mV). Confirm the PLC now reads exactly 0.0 °C. If it does, the hand-held thermometer was the uncalibrated instrument.
7. Trim the signal conditioner. Adjust ZERO with the sensor at a real 0 °C reference (not at the trim-pot's labelled detent), then adjust SPAN at 200 °C. Repeat 2–3 times to converge. Most industrial conditioners have both trim pots accessible behind a sealed cover.
8. Re-verify across the full span. Sweep 0, 50, 100, 150, 200 °C and confirm ±0.5 °C before handing the panel back to operations.

Hardware Wiring and Signal Conditioner Checks

A 0–10 V single-ended input requires three wiring decisions that all manifest as a constant offset when wrong:

• Common reference. The transducer GND / 0 V terminal must connect to the CPU's analog M terminal. A floating input reads the ground-potential difference between the cabinet and the field device, which shows up as a constant bias.
• Single-point shield ground. Earth the cable shield at one end only - typically at the cabinet entry gland, not at the field device - to break ground loops that often shift the reading 0.3–0.7 V when heavy loads switch.
• Channel mode. On S7-1200 onboard AIs the voltage vs. current mode is software-selected in the device configuration view under "AI configuration → Inputs". A mismatch presents a 2:1 or 5:1 ratio (not a constant offset); rule this out first before returning to the offset diagnosis.

If the transducer is bipolar (±10 V) and the signal conditioner is set for that range, the 0–10 V intermediate will clip at 0 V when the process is below the mid-span. The PLC then sees 0 °C at any sub-zero temperature, masking the offset diagnosis. Verify the conditioner output range with a manual volt-meter sweep before condemning the scaling.

Software Adjustments When the Hardware Is Acceptable

If a software trim must ship before the transducer can be re-calibrated, compensate the SCALE_X limits with a linear correction. Compute it from two calibration points (T_low_meas, V_low_meas) and (T_high_meas, V_high_meas):

T_corrected = a × T_measured + b

where:

a = (T_{high_real} − T_{low_real}) / (T_{high_meas} − T_{low_meas})

b = T_{low_real} − a × T_{low_meas}

Implement the linear block in SCL inside the same FB that hosts NORM_X / SCALE_X. Keep a and b in a separate global DB so the HMI can re-tune them after each maintenance visit without re-compiling the program.

Firmware and Library Version Considerations

NORM_X and SCALE_X require TIA Portal V13 (released 2014) and CPU firmware V4.0 or later. If the project was migrated from an older STEP 7 Basic release, or the CPU was swapped for a V3.x spare while the project targets V4.x, the legacy SCALE FB is sometimes resurrected by the migration. Open the cyclic OB and confirm the block types:

Check the installed library version in TIA Portal under "Options → Support packages…" or "Project information → History". Siemens maintains backward compatibility notes for the S7-1200 in the public release notes at support.industry.siemens.com; cross-check the firmware-version release notes for the exact NORM_X / SCALE_X availability before deploying.

Field-Proven Code Snippet

This SCL function block reads %IW64 (CPU 1214C onboard AI channel 0), normalizes it to 0.0…1.0, scales it to °C, and applies a software trim stored in a global DB so the HMI can re-tune it at runtime:

FUNCTION_BLOCK "Temp_RTD_Scale"
VAR_INPUT
  iRaw : WORD;        // bound to %IW64
END_VAR
VAR_OUTPUT
  oCelsius : REAL;
  oValid : BOOL;
END_VAR
VAR
  sNorm : REAL;
  sScale : REAL;
END_VAR
BEGIN
  IF iRaw = 0 THEN
    sScale := 0.0;
    oValid := FALSE;
  ELSE
    sNorm := NORM_X(MIN := 0.0,
                    MAX := 27648.0,
                    VALUE := WORD_TO_INT(iRaw));
    sScale := SCALE_X(MIN := 0.0,
                      MAX := 200.0,
                      VALUE := sNorm);
    oValid := TRUE;
  END_IF;

  oCelsius := "cal_gain" * sScale + "cal_offset";
END_FUNCTION_BLOCK

Bind cal_gain to 1.0 and cal_offset to 0.0 initially, then refine from a two-point ice-bath / 200 °C calibration sweep.

Verification and Acceptance Test

After correcting either the wiring, the signal conditioner, or the scaling constants, run the following acceptance test and sign off only when each step passes:

1. Injector = 0 °C, PLC reads 0.0 ± 0.5 °C, stable for ≥ 60 s.
2. Injector = 100 °C, PLC reads 100.0 ± 0.5 °C.
3. Injector = 200 °C, PLC reads 200.0 ± 0.5 °C.
4. Apply a step from 0 °C to 200 °C and confirm the PLC follows within the configured AI integration time.
5. Repeat steps 1–3 on the spare/backup AI channel if the installation is redundant.
6. Record the calibration coefficients (a, b, or cal_gain / cal_offset) in the maintenance log alongside the as-left %IW64 value for traceability.

FAQ