Symptom
A level, pressure, or flow reading shifts or becomes noisy when a drive runs and cleans up when it stops; the noise changes when the carrier frequency is changed; a speed reference to a drive is noisy and the drive hunts; the affected signals share a panel or a route with the drive or its motor cable.
The short answer
VFD Noise on Analog Signals
A drive switches its output between the DC bus rails in a fraction of a microsecond thousands of times a second, and those edges couple into any nearby signal wiring through capacitance and through common-mode current returning in the ground. An analog input samples that noise and aliases it into a slow wander or a step that looks like a process change; a noisy speed reference makes the drive hunt. Prove it by trending the raw input while the drive starts and stops and while its carrier frequency is changed. Fix the signal side with separated routing, shielded twisted pair grounded at one end, isolation, and filtering, and fix the drive side with shielded motor cable bonded at both ends, a short drive ground, and if needed an output reactor or filter and a lower carrier frequency.
Key points
- Starts with the drive and stops with the drive: trend the raw input and prove it before touching anything.
- Changing the carrier frequency changes the noise: a cheap test that confirms the drive as the source.
- The input aliases kilohertz noise into a slow wander; the reading looks like process, not noise.
- Signal side: separate route, shielded twisted pair, one shield ground, isolation, then filtering.
- Drive side: shielded motor cable bonded at both ends, short drive ground, output reactor, lower carrier.
- The speed reference cable inside the drive panel is the most common victim; route it away from the motor leads.
Possible causes and what to check
| Possible cause | What to check |
|---|---|
| Signal wiring routed with the drive output cable, in the field or in the panel | Trace both; inside the drive panel, look at whether the reference and feedback wiring share a wireway with the motor leads. |
| Unshielded motor cable, so the drive current returns through the plant ground | Check the motor cable type and its shield termination at both ends. An unshielded or pigtail-terminated cable is the first drive-side fix. |
| Drive ground long, thin, or missing | The drive manual calls for a short, wide ground connection to the panel ground; a long green wire is a high impedance at the frequencies involved. |
| High carrier frequency | Note the carrier frequency setting; lower it temporarily and watch the noise. Higher carriers are quieter to the ear and noisier to the instruments. |
| Shield of the signal cable grounded at both ends or not at all | Lift and measure; the drive common-mode current on a shield grounded twice couples straight into the pair. |
| Speed reference from the controller analog output on an unshielded or shared cable | Look at the reference wiring; a noisy reference shows up as the drive speed wandering rather than as a bad reading. |
| Input with no filtering and a fast sample rate | Check the module input filter setting; a wide-open input samples the noise and aliases it. |
| Long motor cable increasing capacitive current | Motor cable length from the drive manual limits; long runs need an output reactor or filter. |
What the drive does to a signal
Every output pulse of a drive is a voltage step of several hundred volts in a fraction of a microsecond. That rate of change drives current through any capacitance it can find: between the motor cable and adjacent cables, between the motor windings and the frame, between the drive and its enclosure. The current that flows through the motor cable and windings to ground has to return to the drive, and it returns on the shield if there is one bonded at both ends, and through the plant ground if there is not. A signal cable that runs near the motor cable receives the capacitive coupling directly; a signal circuit that touches the plant ground receives the returning current as a common-mode voltage. Either way the noise is at the carrier frequency, typically a few kilohertz, with harmonics far higher.
Why it looks like the process moved
An analog input converts the signal at a sample rate, often tens or hundreds of samples per second. Noise at several kilohertz, sampled at that rate, does not show up as several kilohertz; it shows up as a beat between the noise frequency and the sample rate, which can be a slow wander of seconds or minutes, or as a constant offset if the noise happens to be rectified in the input stage. The trend shows the level creeping up and down or stepping, in a way that looks exactly like a process change, and the operator adjusts something. A trend of the raw counts at the fastest rate the module supports, with the drive started and stopped, is what reveals it.
Proving it
- 1
Trend the raw input
The unfiltered raw counts, at the fastest rate available. Note the noise band with the drive stopped.
- 2
Start the drive
Watch the trend as the drive starts, ramps, and runs at several speeds. Noise that appears with the drive and changes with speed is the drive.
- 3
Change the carrier frequency
Within the range the drive manual allows, set the carrier to a different value and watch the noise change in character or amplitude. This is the confirming test; nothing else in the plant does this.
- 4
Check the signal side
Route of the signal cable relative to the motor cable, field and panel. Cable type. Shield termination at each end. Input filter setting.
- 5
Check the drive side
Motor cable type and shield termination. Drive ground. Motor ground back to the drive. Carrier frequency. Motor cable length against the manual limit. Output reactor or filter present.
- 6
Fix in order
Cheapest and most general first: input filter and carrier frequency to confirm; then shielded motor cable and drive grounding, which helps every signal; then signal routing and cable; then isolation and filtering on the loops that still need it.
- 7
Verify
Raw trend again under the same conditions, at the speeds that were worst. Then the filtered value the loop actually uses.
Fixes
| Side | Fix | What it does |
|---|---|---|
| Drive | Shielded motor cable, shield bonded 360 degrees at the drive and the motor | Returns the common-mode current on the shield instead of through the ground and the signal circuits |
| Drive | Short, wide ground from the drive to the panel ground bus; motor ground conductor back to the drive | Gives high-frequency current a low-impedance path that is not the instrument ground |
| Drive | Lower carrier frequency within the manual range | Less high-frequency energy; more audible motor noise |
| Drive | Output reactor, dv/dt filter, or sine filter | Slows the edges at the source; needed for long motor cables anyway |
| Drive | Common-mode core on the motor leads at the drive | Impedance to common-mode current at the source |
| Signal | Separate route from the motor cable, in the field and in the panel | Removes the capacitive coupling path |
| Signal | Shielded twisted pair, shield grounded at the panel end only | Blocks capacitive coupling; avoids shield current |
| Signal | Loop isolator or isolated input | Removes the common-mode component |
| Signal | Input filter set for the process, and a ferrite at the panel entry | Removes what remains, at the cost of response time |
Where the shield goes on drive control wiring
Drive manufacturers differ on where to ground the shield of the control wiring to the drive. Many specify the drive end, at the shield clamp provided on the control terminal strip, because the drive control common is the reference for its analog input. The instrument convention of grounding at the panel also works when the drive and the controller share a ground bus. What does not work is both ends, or a long pigtail. Follow the drive manual for the drive control wiring and the instrument convention for the field loops, and write on the drawing which was done.
Frequently asked questions
- The drive has an EMC filter. Why does it still cause noise?
- The input filter reduces what the drive puts back on the supply line and does nothing for the output side, which is where the motor cable noise originates. The output side needs the shielded motor cable, the grounding, and if needed an output reactor or filter. Also check that the filter is connected; some drives ship with the filter ground disconnected for use on ungrounded supplies.
- Will a software filter on the input fix it?
- It will hide it, at the cost of response time, and the drive will go on injecting current into the ground system that reaches other signals. Use the filter to make the loop usable while the drive installation is corrected, and then reduce it.
- Is a lower carrier frequency bad for the motor?
- Within the range the drive manual allows, a lower carrier means more audible noise and slightly more motor heating from current ripple, and less drive heating and less electrical noise. Pump and fan applications run fine at the lower end of the range, and many drives default to a value chosen for quiet rather than for instruments.
- The noise is on a 4-20 mA loop. Current loops are supposed to be immune.
- They are immune to induced voltage in series with the loop, but not to current injected through capacitance into the loop conductors, and not to a common-mode voltage beyond the input range. A drive does both. The shielded twisted pair with one ground removes most of the injection, and isolation handles the rest.
Related topics
- Drive Causes Instrument NoiseWhen many instruments became noisy after a drive was installed, the drive installation is the place to look: the checklist for motor cable, shielding, grounding, filters, and carrier frequency that decides how much noise a drive puts into the plant.
- Cable Routing ProblemsNoise that comes from where a cable runs: signal cables sharing a conduit, tray, or wireway with power or drive cables, long parallel runs, and shallow crossings. How coupling works, the separations that prevent it, and how to find the shared path.
- Common-Mode NoiseNoise that rides on both conductors of a signal pair together, relative to ground: where it comes from, why a differential input rejects it only up to a limit, how to tell it from normal-mode noise with a scope, and the fixes that work.
- Filtering Analog InputsHow to damp a noisy analog input in the controller without hiding a real process change: the first-order filter, the moving average, the median, where each belongs, and what never to filter.
- Variable Frequency Drives in Pump PanelsWhat a drive adds to a pump panel and what it costs: speed control, soft start, reduced inrush, and diagnostics against heat, harmonics, cable and motor stress, bypass, and the minimum speed below which a pump does no work.
- 4-20 mA Signal Is UnstableAn analog reading that will not settle. How to separate a genuine process swing from induced noise, a ground loop, or a loop running out of voltage.
Direct contact
Have a controls question?
Reach Eric Sullivan directly about anything on this site, a controls or automation topic, or one of his personal projects.