Encoder Wiring — Differential vs Single-Ended, Signals, and Shielding
Position and velocity feedback lives or dies on the feedback cable — why differential beats single-ended over distance, how to match the input card, and why the counts jump only when the motor runs.
Safety. This guide is educational reference material, not a work instruction. Electrical work is performed de-energized and verified by qualified personnel under your site’s LOTO procedures, following the device manufacturer’s manual and the authority having jurisdiction. An encoder often shares an enclosure with drive and control voltages — treat the panel as energized until proven otherwise.
Overview
An encoder turns shaft angle into electrical pulses, and the control system derives position and velocity from them. However many wires the device has, the landing falls into three terminal groups:
- Power — a supply pair (+V and 0 V), commonly 5 V or 24 V DC.
- Signal channels — the quadrature pair A and B, plus the index / marker Z (one pulse per revolution). On a differential device each is a complementary pair: A/A̅, B/B̅, Z/Z̅.
- Shield / ground — the cable screen and where it lands.
Incremental vs absolute is a real distinction, but mostly a protocol one (incremental = counted pulses needing a home reference; absolute = a coded position, often on a serial channel such as SSI, BiSS, or EnDat). This guide is about wiring practice for the electrical channels — not the protocol.
flowchart LR
PS[Supply<br/>5 V or 24 V DC] --> ENC[Encoder]
ENC -->|A / A̅| RX[Counter / PLC HSC<br/>or encoder input card]
ENC -->|B / B̅| RX
ENC -->|Z / Z̅| RX
Scope: this guide covers general-purpose incremental encoders landing on
counters, PLC high-speed-counter (HSC) inputs, and dedicated encoder input
cards. Servo / drive feedback wiring — resolver, absolute serial feedback,
commutation (U/V/W) channels, and drive-matched cable — lives on the
servo drive wiring guide.
The counts/units/RPM/linear-speed math is handled by the
cst encoder tool — this
guide is about the wiring, not the scaling.
Before You Start
The output type is THE decision — it dictates cable, distance, and which input card you can use. Have on hand:
- Encoder output / interface type. Confirm whether each device is differential (RS-422 / line-driver) — a complementary pair per channel, the choice for long or noisy runs — or single-ended: push-pull / HTL (typically 24 V logic), open-collector (needs a pull-up), or TTL / 5 V. The receiving input card must match. The detail is in Control / Signal Wiring below.
- Supply voltage — 5 V vs 24 V. A 5 V TTL encoder over a long run is trouble: cable voltage drop browns out the device and its line driver. 24 V devices carry far more headroom. Confirm the supply the encoder needs and what the run can deliver at the encoder.
- PPR / resolution and speed. Pulses-per-revolution times RPM sets the output frequency, which trades off against maximum cable length and the input card’s maximum count frequency. Note PPR, top speed, and whether index (Z) and any commutation channels are used.
- Drawings and environment. Cable route relative to VFD/servo motor cabling, the input-card type already selected, and mechanical mounting decided upstream.
Sizing & Protection
Encoder current is small, so this is a voltage-integrity problem, not an ampacity one — there is no conductor-ampacity table to look up.
- Supply voltage drop is the real failure. The +V/0 V pair has
resistance; on a long run at 5 V, a fraction of a volt lost in the cable
browns out the encoder and its line driver, and counts degrade or stop. The
arithmetic is conductor voltage drop over the run — the concept behind
cst voltage-drop; never invent a resistance-per-length value, take it from the wire tables. Some line-driver encoders offer a remote-sense pair so the supply compensates for cable drop — consult the manual and use it on long 5 V runs. - Max cable length falls as PPR/frequency rises and as supply drops. A run fine at low PPR fails when a higher-count encoder is fitted or the speed is raised. Verify length against the encoder’s frequency-vs-length curve and the input card’s maximum count frequency — both datasheet values.
- Protect the supply. Fuse or current-limit the encoder supply per device and panel practice — generally accepted practice, verify for your installation.
Power Wiring
- Clean, dedicated supply. Feed the encoder from a clean DC source and keep its +V/0 V pair out of the motor-power routing; a noisy or sagging supply shows up as unstable counts before it shows up as a dead encoder.
- Mind the drop on the +V/0 V pair. Size the supply conductors so the voltage at the encoder stays in spec — especially at 5 V — and measure it at the device on long runs, not just at the panel terminal. On long line-driver runs, wire the encoder’s remote-sense pair per the manual so the supply corrects for cable drop.
- Torque discipline. Terminal torque and wire ranges are vendor values; record what you used.
Control / Signal Wiring
This is the core of encoder wiring: pick the right output type and pair the conductors correctly.
Output types (verify electrical type on both device manuals — the roles are universal, the labels and thresholds are not):
| Output type | Wires per channel | Noise immunity / distance | Notes |
|---|---|---|---|
| Differential (RS-422 / line-driver) | Complementary pair (A/A̅) | Highest; the choice for long / noisy runs | Receiver reads the difference; terminate fast/long lines |
| Push-pull / HTL | One (24 V logic) | Moderate; robust logic swing | Common 24 V single-ended; better margin than TTL |
| Open-collector | One (needs pull-up) | Low–moderate | Input or external resistor supplies the pull-up |
| TTL / 5 V | One (5 V logic) | Lowest; short runs only | 5 V swing over distance is trouble |
- Differential rejects common-mode noise. Each channel is a twisted pair carrying complementary signals. Noise couples roughly equally into both conductors of a tight pair, shifting them together; the receiver reads only the difference, so equal noise cancels. This is the same balanced-signaling idea that lets RS-485 run through noisy plant — the full explanation lives on the RS-485 physical-layer page.
- Twist by channel complement, not at random. Pair A with A̅, B with B̅, Z with Z̅ — that pairing is what makes common-mode rejection work. Pairing A with B (a classic field error) destroys the rejection and can cross-couple the channels. Single-ended has no such rejection — its whole margin is the logic swing against a shared 0 V, which fades with distance and nearby power switching, so use it only on short, quiet runs.
- Match the receiver to the driver. A TTL (5 V) output into a 24 V HTL input under-drives it (no counts); a 24 V HTL output into a 5 V TTL input over-drives and can damage it. Open-collector needs a defined pull-up. Confirm driver and input electrical type on both manuals before energizing.
- Terminate fast / long differential lines. A long line-driver run at high edge rate reflects off an unterminated end, corrupting later edges; terminate across each receiver pair (commonly ~120 Ω, matching the cable) where length or frequency warrant. Short, slow runs may not need it — same reasoning as RS-485 termination.
- Quadrature and index. A and B are 90° out of phase; their lead/lag gives direction, and the ×4 edges give resolution. Z is one pulse per revolution for homing / reference. Wire every channel the counter uses; land unused complementary inputs per the card manual. Terminal designations, logic thresholds, pull-up values, and termination provisions are vendor-specific — consult the device manual.
Grounding, Shielding & EMC
Device-specifics here; the deep treatment is owned by the noise & EMC mitigation guide and the shield-landing theory by panel grounding & bonding.
- Shield landed at ONE end only for the low-level encoder signal, commonly the controller / panel end. The shield drains coupled noise without becoming a conductor between two grounds. This is deliberately the opposite of the both-ends 360-degree rule for high-frequency VFD/servo motor cable — a signal cable screened at both ends across any ground-potential difference just carries 50/60 Hz hum into the counts. The frequency reasoning behind the two rules is owned by grounding & bonding.
- Keep the encoder cable away from VFD / servo motor power. This is the classic encoder-counts-jumping-under-motor-load fault: the drive’s PWM output couples into the feedback cable, and the counts drift or jump only when the motor runs. Separate the routes, cross at right angles, and never share a duct with motor cable — separation classes are in the EMC guide and the motor-cable side is covered on the servo drive wiring guide. Maintain shield continuity through every junction box, and do not land the shield on a signal conductor.
Common Mistakes
- Single-ended encoder on a long, noisy run. A push-pull or TTL device where the distance and environment demanded differential leaves no common-mode rejection — counts come and go, worst when nearby equipment switches. Specify a line-driver (RS-422) encoder for the run.
- Wrong twisted-pair grouping. Pairing A with B instead of A with A̅ throws away the cancellation that makes differential work and can cross-couple the channels. Twist and land each channel with its complement.
- 5 V supply drop over a long cable. The +V/0 V drop browns out the 5 V encoder and its line driver — counts degrade, freeze, or read erratically, looking like a flaky encoder. Measure voltage at the encoder; size the supply pair, or use remote sense / a 24 V device.
- Encoder cable in the motor-cable tray. Running feedback alongside VFD or servo motor cable couples PWM noise into the counts — the textbook counts jump under motor load symptom. Separate routes, cross at right angles.
- Shield landed at both ends. On a low-level signal across any ground-potential difference, the screen becomes a ground-loop conductor and injects hum that corrupts counts. Land the encoder shield at one end only.
- TTL into an HTL input (or vice versa). A 5 V output under-drives a 24 V input (no counts); a 24 V output over-drives a 5 V input and can damage it. Match the electrical type on both manuals before energizing.
- Missing termination on a fast differential line. A long, high-edge-rate line-driver run left unterminated reflects and miscounts. Terminate each receiver pair where length/frequency warrant.
Verification Checks
Prove the feedback before handing the axis to the machine (evidence-retaining sheets in templates):
- Output type confirmed on both manuals; input card electrical type matches (differential/HTL/open-collector/TTL)
- Supply voltage measured at the encoder on long / 5 V runs; within spec with margin
- Twisted pairs grouped by complement (A/A̅, B/B̅, Z/Z̅), not A-with-B
- Differential pairs scoped for clean, fast edges at the receiver end — same practice as reading an RS-485 bus with a scope
- Count-per-rev check: one shaft revolution equals the expected counts
(PPR × decode factor) —
cst encodergives the expected scaling - Direction / quadrature check: A leads/lags B in the expected sense; count direction matches machine direction; Z pulse once per rev where used
- Termination fitted on fast/long differential lines; shield landed at one end only; encoder run separated from VFD/servo motor cabling
- Run the machine and watch for count jumps under motor load — the tell for coupling from drive/motor cable
- Terminal torques per the device manual, recorded
Standards References
- RS-422 / RS-423 line-driver convention — the differential encoder output type; cited as a category, not a specific clause.
- NFPA 79:2024 — machine wiring-practice chapters: conductor identification, routing, and separation of sensitive signal circuits from power (chapter-level citation).
- IEC 60204-1:2016+AMD1:2021 — wiring practices for machine electrical equipment, the international counterpart; clause-level citation only.
- Device manuals are the authority for terminal designations, output electrical type, supply and remote-sense wiring, logic thresholds, and termination values.
Related Pages
- Servo drive wiring — motor-feedback, resolver/absolute feedback, and drive-matched cable; the motor-cable side of the noise story
- Noise & EMC mitigation — separation classes and hardening the victim circuit
- Panel grounding & bonding — shield-landing policy and the frequency reasoning behind one-end vs both-ends
- RS-485 physical layer — the balanced-signaling and termination theory encoders share, plus the oscilloscope method
- Engineering toolkit —
cst encoderfor counts ↔ units ↔ RPM ↔ linear-speed scaling
This site is an independent educational reference for industrial automation standards. It paraphrases and organizes guidance in original language; it is not a substitute for the official published standards, professional engineering judgment, or legal review, and content has not been independently verified against current published editions.
Items marked Review pending await technical review against the identified edition. Items marked Planned have no reviewed source material behind them. Do not rely on this site for compliance determinations, safety-critical design decisions, or legal interpretation.