The receiver drops frames, reports corrupt data, or stops responding as cable length or signaling rate increases. Follow the data path from the sending UART through the RS-485 transmitter, connectors, trunk cable, receiver, and receiving UART. RS-485 defines the electrical interface; it does not provide one universal distance-versus-data-rate table for every cable.
Where does the RS-485 data path stop?
First identify whether the failure occurs before transmission, on the physical link, or after reception. Confirm that the sender produces the intended serial frame and that the driver-enable signal keeps the transmitter active for the complete frame. Then compare the differential waveform at the transmitter pins with the waveform at the farthest receiver.
| Check point | Reading to take | Outcome | Next check |
|---|---|---|---|
| Sending UART | Frame format and data timing | Incorrect locally | Correct the UART or application configuration |
| Driver output | Differential waveform while transmitting | Missing or truncated | Check driver enable, supply, and local wiring |
| Far receiver | Differential amplitude, edge shape, ringing, and common-mode movement | Degraded relative to the driver | Inspect cable, topology, termination, and loading |
| Receiving UART | Framing, parity, and error counters | Electrical waveform is usable but decoded data is wrong | Compare serial settings and protocol timing |
Probe the differential signal directly or derive it from properly acquired conductor measurements. A single conductor measured against local ground can look noisy even when the differential receiver sees a valid signal. Conversely, a clean-looking conductor-to-ground trace can hide insufficient differential voltage.
Does the cable pass a layer-one inspection?
Layer one first. Isolate power, then check continuity conductor by conductor, conductor-to-conductor shorts, conductor-to-shield faults, polarity through every connector, and unintended branches. Record the actual routed length rather than estimating from drawings. A star, long stub, or unterminated branch can create reflections even when the total installed cable length appears acceptable.
The proposed 1000 ft, 1 MHz, and 26 AWG values describe a design target, not a guaranteed RS-485 operating point. Wire gauge primarily affects conductor resistance. It does not specify characteristic impedance, capacitance per unit length, propagation velocity, attenuation versus frequency, shield construction, twist consistency, or pair balance. Read those properties from the exact cable datasheet.
| Cable property | Mechanism | Diagnostic reading |
|---|---|---|
| Loop resistance | Increases voltage loss and interacts with termination and node loading | Measure end-to-end resistance and compare it with cable length and datasheet resistance |
| Characteristic impedance | A mismatch produces reflections at cable ends and branches | Compare the cable value with installed termination |
| Capacitance per unit length | Increases driver loading and slows transitions as length rises | Calculate total cable capacitance from the datasheet and routed length |
| Propagation velocity | Sets reflection delay and end-to-end latency | Calculate one-way delay from routed length and datasheet velocity |
| Balance and shielding | Affect conversion of external interference into differential noise | Compare errors with machine states and inspect shield bonding |
Does the far-end waveform fit inside one bit period?
Determine what 1 MHz means. It might describe a clock, symbol rate, edge repetition rate, or bit rate. Record the configured bit rate as R_b and calculate the nominal bit period as t_bit = 1 / R_b. Coding, framing, and protocol turnaround affect throughput but do not change that bit-period calculation.
Capture several alternating-bit patterns because repeated identical bits can hide bandwidth limitations. At the farthest receiver, examine differential amplitude near the sampling region, transition time, overshoot, ringing, baseline movement, and noise. Trigger on the transmitter waveform and compare near-end and far-end traces. If transitions occupy too much of the bit cell or ringing crosses the receiver decision region, reduce the signaling rate, shorten the cable, correct the topology, or select cable with more suitable transmission characteristics.
Do not convert a scope trace into a pass/fail distance by wire gauge alone. The decisive reading is the waveform presented to the receiver under the worst operating conditions. Repeat it while expected drives, contactors, and other interference sources change state.
Do termination, bias, and node loading set the correct operating point?
With power removed, measure resistance across the differential pair at representative points. Interpret the result from the actual termination network, cable resistance, bias network, connected nodes, and any protective components. A resistance value without the circuit topology is not diagnostic by itself.
| Setting or condition | Symptom | Action |
|---|---|---|
| Missing end termination | Large reflection after an edge, often worse at higher rates or longer lengths | Terminate the ends of the main trunk to match the cable design |
| Extra termination | Reduced differential amplitude and excessive driver load | Remove termination from intermediate nodes or stubs |
| Incorrect bias | Unstable receiver state while no driver is active | Review the complete idle-state network and receiver loading |
| Long stub or star branch | Delayed echoes and location-dependent errors | Use a trunk arrangement and minimize branch electrical length |
| Excessive node loading | Amplitude falls as nodes are connected | Calculate the combined load from the actual transceiver datasheets |
RS-485 does not define device addresses, serial framing, or application ports. Check those only after the differential path passes.
| Logical item | Compare at both ends | Typical failure signature |
|---|---|---|
| Device address | Configured destination and expected responder | Valid electrical traffic with no intended reply |
| Serial port | Bit rate, data bits, parity, and stop bits | Framing errors or consistently corrupt characters |
| Turnaround timing | Driver-enable release and responder start | Truncated request, collision, or missing first response byte |
| Protocol settings | Frame format and integrity check | Complete frames rejected by the application |
How do you qualify the resolving branch?
- Document the required routed length, configured bit rate, node count, topology, termination locations, and environmental limits.
- Obtain resistance, characteristic impedance, capacitance, attenuation, and propagation data for the exact cable. Treat 26 AWG only as its conductor-size descriptor.
- Inspect continuity, polarity, shields, connectors, trunk ends, and every branch before changing protocol settings.
- Confirm the sending UART frame and driver-enable timing at the transmitter.
- Capture differential waveforms at the transmitter and farthest receiver using patterns that exercise frequent transitions.
- Correct topology, termination, bias, or loading when the waveform shows reflections, insufficient amplitude, slow edges, or an unstable idle state.
- If the corrected physical network still lacks timing or noise margin, reduce the bit rate, shorten the route, divide the network, or select a cable whose transmission data fits the required length.
- Run the application protocol with all intended nodes connected. Check receiver diagnostics, missing responses, retries, integrity-check failures, and framing errors.
- Repeat the waveform capture and error test at the maximum installed length and under the worst permitted operating conditions. Record cable identity, settings, traces, and counters as the acceptance baseline.
FAQ
Can I calculate maximum RS-485 distance from wire gauge alone?
No. Gauge gives only part of the resistance information; the calculation also needs cable impedance, capacitance, attenuation, propagation velocity, topology, termination, transceiver loading, and required noise margin.
Does 1 MHz mean the RS-485 link runs at 1 Mbit/s?
Not necessarily. Identify whether 1 MHz denotes a clock, symbol rate, edge rate, or bit rate; for bit rate R_b, the bit period is 1 / R_b.
Can I use 26 AWG cable for a 1000 ft RS-485 run?
26 AWG and 1000 ft do not predict operation by themselves. Qualify the exact cable and transceivers by calculation, then measure the differential waveform at the farthest receiver.
Does termination fix every long-distance RS-485 error?
No. Correct termination controls end reflections, but it cannot correct excessive attenuation, slow edges, extra loads, poor bias, long stubs, polarity errors, or mismatched serial settings.
Can I accept the link because it communicates at nominal conditions?
No. The final verification step is a far-end waveform capture and error-count run at the worst permitted cable length, node count, data rate, supply condition, and temperature.