A vibration transmitter reads unexpectedly high for two distinct reasons: the machine really vibrates more (looseness, imbalance, misalignment, bearing or process change), or the measurement chain amplifies or invents the signal (loose or resonant mounting, wrong measurement axis or location, frequency-range mismatch, wrong scaling in the PLC). The first safe check is comparative: inspect the mounting for looseness and compare the reading with a portable reference measurement at the same point, machine state unchanged. Do not loosen or remount a sensor while the machine runs; mechanical work at the machine follows the site's lockout procedures.
Key Takeaways
- Mounting is the first suspect, not the last. A loose stud, a magnet on a curved surface, or a long bracket creates its own vibration — the transmitter faithfully reports its own motion, not the machine's.
- Resonance amplifies silently. Every mounting (bracket, adapter, adhesive pad) has a natural frequency; if machine vibration excites it, readings multiply at specific speeds while the machine itself is normal.
- Axis and location are part of the measurement. A sensor moved from the bearing housing to a thin cover, or mounted off-axis, reads a different machine. Remounting changes the baseline — compare like with like.
- Scaling errors are perfect impostors. A PLC scaled for the wrong range or unit (velocity vs. acceleration, mm/s vs. in/s) produces a permanently "high" reading with a healthy machine.
- Trend, not one value. Judge against the machine's own baseline at the same point, axis and operating state — not against a generic limit alone.
Symptom Definition
This article covers readings persistently or suddenly higher than the established baseline:
- Sudden step up — high since a specific event (remounting, machine repair, new bracket);
- Speed-dependent high — readings multiply near certain speeds (resonance signature);
- Permanently high since commissioning — never matched expectations (scaling/selection signature);
- Genuinely rising trend — gradual increase over weeks (possible real machine degradation — do not dismiss).
Pattern 4 is the one this guide must not talk you out of: if mechanical checks pass and the rise is real, escalate to machine diagnosis — the transmitter may be doing its job.
Cause Table
| Cause category | Why it creates the symptom | Safe check | Corrective direction | Technician required? |
|---|---|---|---|---|
| Loose mounting | The sensor/adapter moves relative to the machine; relative motion adds to the measured signal | Visual and manual inspection with the machine stopped and locked out; compare with portable reference | Re-mount per manual: correct surface preparation, stud torque or adhesive procedure | Yes — machine-adjacent mechanical work |
| Mounting resonance | Bracket/adapter/magnet natural frequency sits inside the machine's excitation range; amplification peaks at specific speeds | Note whether the high reading peaks at certain speeds; compare a rigid reference mounting at the same point | Stiffen/shorten the mounting path; move the mounting natural frequency above the excitation range | Yes |
| Wrong location or surface | Thin covers, cooling fins or painted/uneven surfaces vibrate differently from the bearing housing | Compare mounting point with the intended measurement location (bearing housing, solid structure) | Relocate to the specified measurement point; prepare the surface per manual | Yes |
| Wrong measurement axis | A single-axis sensor mounted off its sensitive axis under-reads one direction and over-reads another; triaxial setups wired to wrong channels | Confirm axis orientation against the manual's marking; confirm channel mapping | Re-mount/re-map per manual | Yes |
| Frequency-range mismatch | Machine excitation sits outside the transmitter's stated frequency range; integration/filter behavior distorts the value | Compare the machine's characteristic frequencies (running speed, blade pass, bearing tones if known) with the model's frequency range | Reselect a model whose frequency range covers the excitation; the selection guide (https://www.kjt-sensors.com/show-27151.html) covers the parameters | Yes for reselection validation |
| Real machine change | Looseness, imbalance, misalignment, bearing wear or process change genuinely raised vibration | Correlate with machine events; listen/inspect per site practice; compare across machine points | Escalate to machine diagnosis — this is the transmitter working correctly | Yes — machinery diagnosis |
| PLC scaling / unit error | Scaling constants assume a different range or unit; a "high" reading exists only in the PLC | Recompute the expected output (e.g., mA at the measured value) from the data-sheet range; compare with the PLC's raw input | Correct scaling constants and units; verify against the raw loop value | Yes for energized verification |
| Signal-chain noise | Ground loops or cable noise add a false component to the 4–20 mA output | Check whether the raw loop value is stable when the machine is stopped | Correct grounding/routing per manual | Yes — electrical work |
Diagnostic Sequence (Least Invasive First)
- Freeze the context. Record machine type, speed, load, measurement point, axis and the baseline history. No comparison is valid without the same operating state.
- Classify the pattern. Sudden step, speed-dependent peak, permanent-since-commissioning, or gradual trend — each maps to a different cause family above.
- Inspect the mounting (machine stopped, locked out). Stud torque, adhesive integrity, magnet seating, bracket stiffness, cable strain relief.
- Compare with a portable reference. A handheld vibration meter or reference sensor at the same point and state separates "machine is high" from "installed chain is high".
- Verify point, axis and surface. Confirm the sensor sits at the intended measurement location, on a prepared surface, in the marked axis.
- Check speed dependence. If readings peak at specific speeds, treat mounting resonance as primary; verify with the reference mounting.
- Verify the electrical/scaling chain. Recompute the expected loop output from the data-sheet range; compare with the raw PLC input before trusting any scaled value (qualified personnel for energized checks).
- Escalate appropriately. Machine-side causes → machinery diagnosis; measurement-chain causes → re-mount, reconfigure or reselect with the information below.
Conditional Corrective Actions
- If the reference measurement is normal but the installed reading is high → the chain is at fault: re-mount per manual, stiffen the bracket, then re-verify.
- If high readings peak at specific speeds → mounting resonance; shorten/stiffen the mounting path or relocate before touching any electronics.
- If the reading has been high since commissioning with a healthy machine → verify scaling and units first, then the model's frequency range against the machine's excitation.
- If both installed and reference readings are high → the machine genuinely vibrates more; escalate to machinery diagnosis (imbalance, misalignment, looseness, bearings) — do not "fix" the sensor.
- If readings jumped after a model substitution → re-run the compatibility checks: range, frequency response, output type, axis convention.
Replacement and Reselection Criteria
Reselect when:
- the machine's characteristic frequencies fall outside the installed model's stated frequency range;
- the mounting point cannot be prepared to the manual's requirements (space, surface, temperature);
- the application needs velocity, acceleration or displacement in a range the current model does not cover — the selection guide (vibration-sensor selection guide) maps measurement type, range, frequency response and output;
- the transmitter will feed a formal condition-monitoring program — the application information checklist (application-information checklist) lists what the selection record needs.
Information to Send KJT Sensors Support
- Exact model number, measurement type (acceleration/velocity), range and output (nameplate/data sheet);
- Machine: type, speed range, and the measurement point's location and axis (photo of the mounting);
- Mounting method (stud/adhesive/magnet/bracket) and surface condition;
- Reading history: baseline, the step/trend, speed dependence;
- Portable reference comparison result, if taken;
- PLC raw input value and scaling constants.
Frequently Asked Questions
Can loose mounting amplify readings?
Yes — and it is the most common non-machine cause. If the sensor can move relative to the machine, that relative motion is measured as vibration; compliant mounts (long brackets, magnets on curved or painted surfaces, degraded adhesive) make it worse. The fix is mechanical: prepared surface, correct stud torque or adhesive procedure, short stiff mounting path — verified with the machine stopped and locked out, then confirmed against a portable reference.
Does measurement axis matter?
Decisively. A single-axis sensor responds along its sensitive axis; mounted off-axis it measures a projection of the true motion, and different axes of a machine genuinely differ (horizontal vs. vertical readings on the same bearing are not interchangeable). Confirm the axis marking against the manual, keep the axis consistent with the baseline's axis, and never compare values across axes without a deliberate reason.
How can PLC scaling create an apparently high value?
If the PLC's scaled range does not match the transmitter's actual output range — or the units differ (velocity vs. acceleration, metric vs. imperial) — every value is multiplied by the error while the loop itself is healthy. The test is to bypass the scaling: recompute what the raw loop value (e.g., mA) means against the data-sheet range and compare it with the reference measurement. Scaling errors are permanent and uniform; real vibration is neither.
The reading rose gradually over two months — is that drift?
Treat a gradual rise as real machine behavior until proven otherwise. Sensor drift exists, but looseness, imbalance growth and bearing wear are far more common — and they are the reason the transmitter is installed. Run the mounting and reference checks; if they pass, escalate to machinery diagnosis rather than adjusting the instrument.
Conclusion
Unexpectedly high vibration readings are resolved by separating the machine from the measurement chain: freeze the operating context, inspect the mounting, compare with a portable reference at the same point and state, then verify axis, frequency range and scaling. Loose or resonant mounting and scaling errors are the frequent impostors; a genuine rising trend is the transmitter doing its job — escalate it. All mechanical work at the machine is stopped-and-locked-out work.
Sources
- KJT Sensors, Vibration Sensors category (piezoelectric speed, acceleration, integrated and three-axis transmitter models listed, incl. A45J89 integrated vibration transmitter and N31F20 three-axis transmitter) — https://www.kjt-sensors.com/list-zdcgq.html (verified 2026-10-10)
- KJT Sensors, Service & Support — https://www.kjt-sensors.com/list-services.html (verified 2026-10-10)
- KJT Sensors internal category knowledge base: vibration sensors — version current at drafting
- General vibration-measurement practice (mounting resonance, axis sensitivity, frequency range; consistent with ISO 20816-series machine-vibration evaluation concepts; no third-party claims)
Contact Technical Support
Send your installation photos and machine-operating data via the KJT Sensors service page — include the reference comparison and speed-dependence pattern, and a vibration specialist will help separate mounting, resonance and selection causes from real machine behavior.
https://www.kjt-sensors.com/
KJT Sensors
