LOGO! 8 Analog Inputs I7/I8 Fluctuation: Root Cause and Field Fix

1. Problem Statement

On a LOGO! 8 base module (for example, 6ED1052-1MD08-0BA1 LOGO! 12/24 RCE or 6ED1052-1CC08-0BA1 LOGO! 24 RCE), the on-board analog inputs I7 and I8 show a noisy, fluctuating process value when wired to a PT100-to-0-10 V transmitter. The fluctuation survives a stable supply voltage and disappears when the same input is driven from a digital potentiometer (a clean, low-impedance bench source). The observed jitter is typically 3-12 counts on a 0-1000 normalized scale (3 mV-12 mV on a 10 V span), which is well outside the LOGO! 8 data-sheet repeatability when the analog front end is properly terminated.

This pattern is the classic signature of common-mode and capacitive noise injection on a high-impedance analog node, not a failed module. The first thing to confirm is that the symptom is not a hardware defect by swapping the wiring between I7 and I8: if the noise stays with the cable, the cable is the source; if the noise stays with the input, the input is suspect. In the field, the cable wins almost every time.

2. Root Cause Analysis

The LOGO! 8 base module exposes I7 and I8 as 0-10 V single-ended analog inputs. The Siemens data sheet specifies an input impedance of approximately 80 kΩ on these channels. That high impedance is necessary to avoid loading the field device, but it also makes the node a good antenna: every volt of coupled common-mode noise appears directly as a process-value excursion.

The PT100 transmitter in the user's rig, by contrast, typically presents a low-impedance output (10-500 Ω) with a clean, driven 0-10 V rail. The digital potentiometer is even cleaner. So the transmitter itself is not the culprit; the path between the transmitter and the LOGO! terminal is.

The four real root causes seen in this scenario, in order of frequency, are:

1. Unshielded analog pair routed in parallel with switching conductors. Capacitive coupling from adjacent AC or DC switching wires (relay coil returns, contactor legs, VFD output cables, even 24 V DC power) injects an AC component directly into the high-Z analog node. A 50 Hz/60 Hz mains field typically shows up as 3-6 count drift; a nearby PWM drive shows up as a 4 kHz-16 kHz ripple superimposed on the DC value.
2. No or poor shield bonding. Even when shielded cable is used, a shield that is not bonded at one end (and only one end) acts as a pickup loop rather than a barrier. Shielding the analog pair is meaningless if the drain wire is floating.
3. Shared or "dirty" ground reference. The 0-10 V return (M of the analog input, terminal M on the base module) tied to a ground bus that also carries load currents will have a few hundred millivolts of IR drop riding on it. The transmitter "sees" that ground bounce and the LOGO! "sees" it again, so it double-counts.
4. Source impedance too high relative to input impedance. A passive or weakly-buffered PT100 transmitter with output impedance above 1 kΩ, paired with a long unshielded run, forms an RC low-pass that integrates line noise. Verify the source impedance spec; if the transmitter cannot drive 80 kΩ without measurable attenuation, the front end is technically out of spec.

3. Confirming the Diagnosis

Use this matrix to localize the noise source before changing anything. Each test is non-destructive and takes less than five minutes on a running machine.

If the channel swap and source swap both indicate the cable/transmitter side, proceed to Section 4. If both indicate the LOGO! side, check the terminal seating, then the firmware version, then consider RMA.

4. Step-by-Step Resolution

Apply these in order. Each step is independent, and you can stop as soon as the noise clears.

4.1 Re-seat the analog terminal block

Power down the LOGO!, remove the detachable terminal block for I7/I8, inspect for discoloration or loose screws, and re-install with the recommended torque. The Siemens manual specifies 0.6 Nm for the LOGO! 8 plug-in terminals. A loose M (analog common) is the single most common cause of intermittent analog noise on LOGO! installations.

4.2 Enable the on-board analog filter FB

LOGO! Soft Comfort (LSC) provides a dedicated Analog Filter function block (FB) under Special Functions > Analog > Analog Filter. This block performs a moving average over a configurable number of samples, which suppresses both high-frequency noise and 50/60 Hz pickup.

1. Open the LSC program and place the Analog Filter FB on the AI input tag (for example, AI1).
2. Set Sn (number of samples) to a value appropriate for the response time you need. Each sample on a LOGO! 8 base module is one scan cycle, typically 50 ms when only digital I/O is configured, and longer when the AM2 PT100 expansion module is on the bus.
3. Select Average as the filter type. The block will then output the arithmetic mean of the last Sn samples.
4. Compensate for the added lag in the downstream control loop. As a rule of thumb, Sn = 20 (1 second of averaging) is a sensible starting point for temperature loops driven by a PT100.

4.3 Replace the cable with a shielded twisted pair

For 0-10 V analog runs longer than 0.5 m, use a shielded twisted pair (e.g., LiYCY 2 x 0.34 mm²) with the shield bonded to the LOGO! M terminal at the LOGO! end only. The transmitter end of the shield must remain open, or be capacitively bypassed (10 nF + 1 MΩ) to chassis ground, to avoid a ground loop. Per the LOGO! 8 system manual, the analog inputs are referenced to M, and the cable shield should be treated as an extension of that reference at the LOGO! end.

4.4 Reroute away from noise sources

Maintain at least 200 mm of physical separation between the analog pair and any of the following:

• AC power conductors (especially 230 V coil return lines from contactors and solenoid valves).
• LOGO! relay output conductors carrying inductive loads (the dry-contact outputs on the relay variants of the base module).
• Variable frequency drive output cables (PWM common-mode is broadband and aggressive).
• 24 V DC power trunks shared with solenoid valves.

If crossing is unavoidable, cross at 90 degrees rather than running parallel.

4.5 Provide a clean dedicated ground for the transmitter

Run a dedicated 0 V / M conductor from the LOGO! M terminal to the transmitter's GND terminal, and bond the cable shield at the LOGO! end to that same M terminal. The transmitter's chassis ground (if present) should be bonded to the panel chassis at the transmitter end, not at the LOGO! end. This is the standard "single-point shield ground" pattern that prevents shield currents from modulating the signal return.

4.6 Verify the transmitter output impedance is in spec

The transmitter's 0-10 V output must be able to drive the LOGO! 80 kΩ input without measurable attenuation. For a 4-wire (loop-powered or separately-powered) industrial PT100 transmitter, output impedance is typically ≤ 50 Ω and the loading error is below 0.07 %. If the transmitter in use is a low-cost 2-wire device with an output impedance above 1 kΩ, replace it with a 4-wire type or add an op-amp buffer (e.g., a single LM358 powered from the LOGO! 24 V rail) between the transmitter and the LOGO! terminal.

4.7 Add a hardware RC filter at the terminal as a last resort

If the noise is broadband and the filter FB is not enough, install a 10 kΩ resistor in series with the AI+ line and a 100 nF X7R ceramic capacitor from AI+ to M, both at the LOGO! terminal. This forms a ~1.6 kHz first-order low-pass that does not load a 50 Ω source but blocks RF and PWM noise. Combined with the filter FB, this is usually overkill in the good sense of the word.

5. Verification

After each fix, watch the LOGO! display or LSC online monitor for at least 10 minutes with the system in its normal operating state (drives running, contactors cycling). The AI value should be steady to within ±2 counts (2 mV) on a stable temperature.

Quantitative acceptance criteria:

• Steady-state noise: Peak-to-peak variation ≤ 0.5 % of span (5 counts on a 0-1000 scale).
• Step response: Time to reach 90 % of a 10 % input step within the expected loop time constant, plus filter delay. If the step is excessively slow, the filter FB or RC filter is over-tuned.
• Cross-channel isolation: Toggle a relay output on the same base module and confirm the AI value does not move more than 1-2 counts.

6. Firmware-Specific Notes

LOGO! 8 firmware has been updated several times since launch. The current long-term branch as of 2024 is FS: 03 (8BA8) for the LOGO! 8.2 line, with the newer LOGO! 8.3 (8.FW4) line shipping firmware FS: 04 on the 6ED1052-xMD08-0BA2 part numbers. The on-board AI behavior and the Analog Filter FB have not changed in a way that affects this troubleshooting path, but if the LOGO! is on a very early 8BA6 firmware, Siemens released an I/O timing fix that may slightly reduce noise coupling on the AI channels.

7. Comparison with Other Small-PLC Analog Behavior

The same root cause pattern — fluctuating analog readings on a high-impedance input when driven by a real-world sensor — is well documented across the small-PLC landscape. The underlying physics is identical, and the remediation principles are identical.

The shared lesson is that all single-ended high-impedance analog inputs on compact controllers require shielded, separately-routed, single-point-grounded wiring for reliable sub-percent measurement. Treating the analog pair as a digital wire is the root cause of nearly every fluctuation ticket in the field.

8. Specification Reference

9. Preventive Checklist for Future Installations

1. Always run a shielded twisted pair for any AI on I7/I8, even for short runs inside the panel.
2. Bond the shield at the LOGO! M terminal only, never at both ends.
3. Provide a dedicated analog M conductor back to the LOGO!; do not share the analog return with a load's 0 V.
4. Default to enabling the Analog Filter FB at Sn = 10 for any temperature loop and increase only if needed.
5. Maintain ≥ 200 mm of physical separation from AC, relay-output, and VFD cables; cross at 90° when necessary.
6. Torque the LOGO! detachable terminals to 0.6 Nm and inspect annually for thermal cycling loosening.
7. Verify that the PT100 transmitter is a 4-wire type or has an output impedance low enough to drive 80 kΩ without attenuation.

10. Quick Troubleshooting Matrix