Troubleshooting
Fixing HMI PLC Communication Timeout Over Serial
Short answer
A serial HMI-to-PLC timeout means the panel sent a request and no valid reply arrived before its timer expired — not that the link is slow. Match the baud rate and parity in the PLC port setup and the HMI driver settings first, then verify the cable pinout, the selected ports and the ground bond. At 9600 baud a full poll cycle takes about 28 ms of wire time, so a failing one-second timeout points to settings or wiring, never the timer.
On this page
An HMI that reports a PLC communication timeout has sent a request on the serial link and received no valid reply before its timer expired. The message does not mean the link is slow, and it does not prove the cable is dead; it means one request went out and nothing the panel could decode came back in time. In the installation behind this article, the fix was to set both ends to the same 9600 baud, 8 data bits, even parity, 1 stop bit — the panel project had been left at no parity, so every reply was discarded. That corrected the fault with no cable work and no timer change. Work the checks in this order: match the serial settings at both ends, verify the cable and its pinout, confirm the selected ports and the ground bond, and only then touch the timeout value. At 9600 baud a full poll and reply take about 28 ms on the wire, so a healthy link never comes near a one-second timer.
What a serial timeout actually means #
A timeout is a one-way verdict: the request left the panel, the timer ran out, and no reply the panel could decode arrived. It bundles every cause of silence or garbage on the wire: the PLC port powered down or set to the wrong mode, serial settings that do not match, a broken or miswired cable, the cable plugged into the wrong port, or a ground reference that corrupts the signalling. It does not prove the wire is electrically dead; it establishes only that no valid response was recognized in time. Panels word the same condition differently — some print a message like "PLC Communication Timeout", others raise a numbered code such as PLC-001 — but the mechanism underneath is identical.
The useful split is between silence and a rejected answer. If the PLC replies but the reply is an error, the panel normally shows a fault or exception code rather than a timeout; a Modbus RTU illegal-function reply is that failure mode, and it points at the request, not the link. A timeout says the panel saw nothing valid at all — which is why the settings comparison comes before any cable work.
Match the serial settings at both ends #
Both ends of an RS-232 or RS-485 link must run the same baud rate, parity, data bits and stop bits; otherwise the receiver slices the stream into bytes the wrong way and discards every frame. If the panel expects 10-bit characters (8 data bits, no parity, 1 stop bit) and the PLC sends 11-bit ones (even parity added), the framing slips from the first byte and each reply lands in the discard pile. The panel then reports a timeout even though bits are genuinely arriving.
- Record the baud rate, parity, data bits and stop bits configured on the PLC serial port the panel is wired to.
- Record the same four values in the HMI project's driver settings for that controller connection.
- If baud or parity differs, you have found the fault. Make both ends match, and change the HMI project rather than the PLC port — other devices may depend on the port's current settings.
- Download the corrected project to the panel and retest before touching anything physical.
On a Modbus RTU link, respect the stop-bit convention as well: parity none pairs with two stop bits, and even or odd parity pairs with one — the guide to parity-none stop-bit selection on Modbus RTU explains why the pairing exists. The same class of fault appears on other controllers too; the walkthrough of a Modbus master parity error on an S7-1200 shows a parity-none setting answered with a communication error code instead of data.
If the recorded values differ, stop there. In the installation behind this article, matching the baud rate and parity at both ends was the entire fix — no cable was replaced and no timer was changed.
Verify the cable and its pinout #
With the settings matched, the next suspect is the path. Check connectors fully seated, conductors not loose or crossed, the cable in the ports you think it is in, and continuity from end to end. A request can leave the panel and never reach the PLC, or the PLC's reply can die on the way back — the timeout looks identical either way.
- If the cable was built in the field, verify every conductor against the pinout for those two exact ports; never infer pins from wire colors or connector shape. The guide to building the RS-232 Modbus cable between an HMI and a controller shows how the pins pair on a standard HMI-to-controller link.
- For a DH-485 link, the DH-485 RS-232 pin map for HMI replacements gives the matching diagram.
- If a known-good cable is on the shelf, substitute it. Communication returning convicts the original cable; the timeout persisting clears the cable and moves the suspicion back to ports or ground.
- Do not shuffle signal wires between pins by trial and error; random swaps have killed more ports than they have fixed.
Confirm the ports and the ground bond #
Verify in both configurations that the connection uses the ports you intend. On PLCs with more than one serial port, each port carries its own settings and its own role; a panel wired to the wrong one faces correct-looking settings and still receives silence. Confirm the port's role in the PLC's configuration instead of moving wires between ports to see what happens.
Then check the reference. The field guidance for the reported installation called for the PLC and the panel to be grounded at the same point: a shared, solid reference keeps both receivers measuring the same voltages, while separate or corroded ground paths can corrupt serial signalling even when settings, cable and pinout are all correct.
If the two devices are grounded at different points, leave re-bonding to a qualified electrician — a ground path can carry fault current, and bonding is installation work, not troubleshooting.
Worked example: wire time at 9600 baud #
Take a typical Modbus RTU poll — read two holding registers at 9600 baud, 8 data bits, even parity, 1 stop bit, so every character is 1 start + 8 data + 1 parity + 1 stop = 11 bits. The request is 8 bytes (address, function 03, start register, register count, CRC) and the reply is 9 bytes (address, function, byte count, 4 data bytes, CRC), with the standard 3.5-character silent gap before and after each frame.
- Request: 8 × 11 = 88 bits; 88 ÷ 9600 bit/s ≈ 9.2 ms
- Reply: 9 × 11 = 99 bits; 99 ÷ 9600 bit/s ≈ 10.3 ms
- Two silent gaps: 2 × 3.5 × 11 = 77 bits; 77 ÷ 9600 bit/s ≈ 8.0 ms
- Total: 9.2 + 10.3 + 8.0 ≈ 27.5 ms on the wire, before a few milliseconds of PLC scan time to service the port
Even a slow scan leaves the whole exchange under 100 ms, so a 1000 ms timeout carries more than 30 times the margin it needs. The arithmetic gives you the decision: if a one-second timer expires, the reply is not being decoded — settings, wiring or ground — and raising the timer only waits longer for the same silence. The timeout value never fixes a mismatch.
The check order and the decision at each step #
Run the checks in this sequence and stop at the first one that finds a fault:
- Read the PLC port's baud rate, parity, data bits and stop bits.
- Read the HMI driver's settings for the connection and make them match.
- Apply the project and retest.
- Inspect the cable end to end; verify the pinout or substitute a known-good cable.
- Confirm the cable sits in the intended ports and both devices share a ground reference.
- Power-cycle both devices once and note whether the link returns.
| What you find | What it means | The decision |
|---|---|---|
| Baud or parity differs between panel and PLC | Frames cannot decode; every reply is discarded | Match the values at both ends, download the project, retest — this was the fix in the reported installation |
| Settings match, field-built cable unverified | Request or reply may not travel the intended conductors | Verify the pinout and continuity, or substitute a known-good cable |
| Settings and cable check out, ports or grounding uncertain | The signal path or its reference is still suspect | Confirm port roles in configuration; verify the common ground bond |
| All checks pass, timeout remains | The cause sits outside these checks | Record the settings and escalate to the panel vendor's support with that record |
| Timeout clears after a power cycle, returns weeks later | Loose connector, marginal bond or coupled noise | Reseat connectors, inspect the bonding, reroute the serial cable away from drive and motor wiring |
When the timeout comes and goes #
Settings that match either work or fail immediately; they never fail on Tuesday and behave on Wednesday. A fault that a power cycle clears and that returns weeks later is physical or electrical: a connector loosening with temperature, a marginal ground bond, or noise coupled into the cable run. Reseat the connectors at both ends, check that the shield and ground path are intact, and route the serial cable away from drive output and motor conductors.
Two neighbouring symptoms mean different things. If the panel updates but lags or freezes rather than timing out cleanly, the link is congested rather than silent — see why an HMI freezes or updates slowly over serial. On an RS-485 multidrop the same silence is often reported as device no response, and the fixes for HMI Modbus no response over RS-485 run the same settings-first order.
The decision in every case follows the same order: settings, then path, then reference, then — and only then — the timer. Never edit PLC logic to chase a serial mismatch; the logic is innocent.
Questions asked
Does a communication timeout always mean the cable is broken?
No. The message only says that no valid reply arrived within the timer. A seated, undamaged cable with a baud-rate or parity mismatch produces exactly the same error, because the panel cannot decode what comes back. Compare the serial settings on both ends before unplugging or replacing anything, and move on to the cable only if the settings already match.
Where do I find the serial settings I need to compare?
On the PLC they sit in the serial port or system setup for the port the panel is wired to. On the HMI they sit in the project's controller or driver settings for that connection. Compare baud rate, parity, data bits and stop bits, and also the station or node number: an HMI polling a station number the PLC never answers will time out with every other setting correct.
Should I increase the timeout value until the error stops?
Only after a healthy link is proven. At 9600 baud a full request-and-reply cycle takes under 30 ms on the wire, so a 1000 ms timeout carries more than 30 times the margin it needs. If that timer expires, the reply is not being decoded, and a longer timer simply waits longer for the same silence. Correct the settings, wiring or grounding instead.
Why does the link run for weeks and then time out once or twice a month?
A settings mismatch fails constantly or not at all, so an occasional timeout has another cause. The usual suspects are a connector working loose as the cabinet heats and cools, a corroded ground bond, or noise coupled into an unprotected cable run. A power cycle that restores the link confirms the pattern: reseat connectors, verify the bonding, and reroute the cable away from drives and motor wiring.