IMDTouchKnowledge BaseShop main store

Troubleshooting

Fixing RS-232 Cable Disconnected Errors on a Replacement HMI

HMI and touchscreens Published 8 min read

Short answer

Jumper RTS to CTS and DTR to DSR at the HMI end of the RS-232 cable before touching communication settings — that pair cleared 'cable is disconnected' and 'response timed out' on a replacement panel polling a Host Link PLC. The different first command, @00MS5E versus @00MM40, is not the fault: both are valid reads and the PLC answers either one passively.

On this page
  1. Why does the replacement HMI say the cable is disconnected?
  2. Is the different first command why the PLC stays silent?
  3. Which pins get jumpers at the HMI end of the cable?
  4. Does the PLC end need the same jumpers?
  5. How do you verify the jumpers and the wiring?
  6. Which settings must match once the wiring is right?
  7. How do you tell a cable fault from a settings fault?
  8. Questions asked

When a replacement operator panel reports cable is disconnected or response timed out over RS-232 to a Host Link PLC, the usual cause is a custom cable that carries transmit, receive and ground but leaves the handshake pairs open. The case this article follows — an Omron CQM1-CPU41-V1 that answered an NT30C for years and ignored a Pro-face GP4301-TM wired with a new cable — was fixed with two jumpers at the panel end: RTS tied to CTS, and DTR tied to DSR. Nothing else changed. The tempting suspect, the panel's different first command (@00MS5E versus the old panel's @00MM40), is a red herring, and the checksum arithmetic below shows why. Work the physical link first and the settings second.

Why does the replacement HMI say the cable is disconnected? #

Because the panel's driver is watching handshake inputs that the custom cable never drives: tie RTS to CTS and DTR to DSR at the panel end, and the error clears — that exact pair of jumpers restored the link in this installation.

TIA-232, the standard behind the RS-232 name, assigns each circuit a function and a direction: TD and RD carry the data, SG is the common return, RTS and CTS manage flow between the ends, and DTR and DSR announce that each side is present. The standard leaves whether a port uses those control circuits to the equipment, which is why two ports can disagree across a cable that is wired correctly for data. A driver that expects CTS and DSR to be asserted reads their absence as no link and reports cable is disconnected; when it transmits anyway and hears nothing inside its timeout window, it reports response timed out. Both messages on one installation is the fingerprint of a data path that is intact and handshake pairs that are open. The retired panel's cable satisfied that panel's port; the replacement brings its own expectations, and a connector change alone can break the match.

Is the different first command why the PLC stays silent? #

No: @00MM40 and @00MS5E are both valid Host Link read requests — model type versus operating status — and neither is a prerequisite, because the controller sits passive on the port and answers whatever well-formed frame arrives.

A Host Link frame is @, a two-digit node number, a two-letter command header, a two-character checksum, then the terminator. MM reads the controller model; MS reads the operating status, meaning program, monitor or run mode. The checksum is an exclusive-OR of the ASCII codes of every character from @ through the header, printed as two hexadecimal characters.

Run it for the old panel's @00MM40:

  • @ = 0x40
  • 0 = 0x30
  • 0 = 0x30
  • M = 0x4D
  • M = 0x4D

0x40 XOR 0x30 = 0x70, then 0x70 XOR 0x30 = 0x40, then 0x40 XOR 0x4D = 0x0D, then 0x0D XOR 0x4D = 0x40 — which prints as 40, the exact ending captured from the old panel.

Run it for the new panel's @00MS5E: the chain is identical until the last character, where S = 0x53, and 0x0D XOR 0x53 = 0x5E — which prints as 5E, again matching the capture.

Both frames pass their own arithmetic, so both are well-formed requests, and the controller answers whichever one arrives: it waits on the port for requests, with no startup exchange required. Host Link master drivers do offer a startup option for whether the panel commands the controller into MONITOR on connect, and that choice shapes polling, not responsiveness. The decision here is firm: stop trying to change the first command and meter the cable.

Which pins get jumpers at the HMI end of the cable? #

Two pairs: RTS to CTS, and DTR to DSR, both closed at the panel's own connector — on a 9-pin D-sub those are pins 7 with 8 and 4 with 6, but the panel's connector drawing is the authority for its shell.

Signal Direction at a DTE 9-pin D-sub pin
DCD in 1
RD in 2
TD out 3
DTR out 4
SG — 5
DSR in 6
RTS out 7
CTS in 8
RI in 9

The pin column applies only where the cable actually uses that shell. The panel's serial connector may be a different one, and the pinout that governs is on the panel's own cable drawing — jumper by signal name, never by a number borrowed from another device. Common no-handshake practice also ties DCD in with DSR at the same end, so a port that checks carrier sees it asserted; one extra jumper prevents a second visit. The data path is untouched: TD and RD still cross between the devices, and signal ground runs through.

If you are building this cable rather than repairing one, the signal-by-signal method used for building the RS-232 Modbus cable between an HMI and a controller is the discipline to copy — the protocol differs, the method does not.

Does the PLC end need the same jumpers? #

No: in this installation both jumpers went in at the panel end only, and the controller answered immediately — jumper only at a port whose driver actually reads those inputs, which is the complaining end and any PC you insert as a bridge.

Jumper wherever a port's configuration enables hardware handshaking, because that is the port reading those inputs. The controller's Host Link port answered requests with nothing driven on its side, so its end stayed TD, RD and ground. A PC inserted as a sniffer or bridge is the classic second candidate: loop the same two pairs at its adapter, or the bridge itself can stall the panel's transmitter mid-frame, and a failure you introduced will masquerade as a panel fault.

How do you verify the jumpers and the wiring? #

Power the link down, open the panel-side connector, and confirm with a continuity tester that RTS-CTS reads short, DTR-DSR reads short, and TD and RD each reach the controller — five minutes with a meter beats an hour of settings changes.

Power both ends down before opening the connector: RS-232 levels are harmless to you, but soldering on a live port can push a transient into the UART and trade a jumper fix for a board repair.

  1. Power the panel and the controller down, or at least unplug the cable at both ends.
  2. Open the panel-side backshell and lay the connector against the panel's cable drawing.
  3. Confirm RTS-to-CTS is a short and DTR-to-DSR is a short at that connector; add DCD to the DSR group if the drawing or the port documentation calls for carrier.
  4. Confirm TD at the panel reaches RD at the controller, and RD at the panel reaches TD, with signal ground common to both.
  5. Close the connector, restore power, and watch the error screen.

In the case this article follows, the error cleared at this step and live controller data appeared. If yours does not, the settings section is next, with the cable now proven.

Which settings must match once the wiring is right? #

Four things: the Omron Host Link driver selected in GP-Pro EX, identical baud, parity and stop bits at both ends, and the polled node matching the controller's station — 00 in the frames captured here.

A frame that arrives at the wrong rate is noise on the wire, and a frame addressed to the wrong node is ignored — both produce the same timeout. The polled node is visible in the capture itself, in the two characters after @: 00 in both frames here, which must equal the controller's station. The capture also pins the panel's rate indirectly: if your terminal bridge shows clean ASCII, the panel and the bridge port already match, and that value is what you compare against the controller's configured rate. Set baud, parity and stop bits identical at both ends from the controller's documented port settings, then leave them alone.

How do you tell a cable fault from a settings fault? #

Sniff the line: a frame that arrives byte-perfect at the controller with no reply points at the return path and handshake; a frame that arrives corrupted points at rate or wiring; no frame at all points at the port configuration or handshake inputs holding the transmitter.

What you observe What it points to Next move
Panel reports cable is disconnected Handshake inputs never asserted Close RTS-CTS and DTR-DSR at the panel end
Panel reports response timed out A request left; no reply inside the window Verify station and serial settings, then the RD path
Old panel still talks, new one silent Controller and its port are proven good Compare the new cable signal-by-signal with the working one
Sniffer shows a clean, checksum-valid frame Panel transmitter and framing are good Stop editing the driver and meter the cable

Work the table top to bottom, one change at a time, reconnecting and observing between changes. The order is deliberate: the jumpers come first because they cost nothing and they were the fix here; settings come second because a meter cannot see them. A symptom that survives every single change is usually two faults stacked, so re-measure the closed connector before suspecting the panel.

Keep reading:

Questions asked

Why did tying RTS to CTS alone leave the error on screen?

RTS/CTS and DTR/DSR are independent pairs: one governs flow control, the other readiness. In this installation RTS-to-CTS was already jumpered and the panel still reported a disconnected cable; closing DTR-to-DSR at the same connector brought the link up. Treat them as two separate requirements, and tie DCD in with DSR if the port also checks carrier.

Can you configure the panel to send the old HMI's first command instead?

No, and there is nothing to gain. The first request belongs to the driver's startup sequence — a model read on one panel, a status read on the other — and the controller answers either one. Some Host Link drivers expose an option deciding whether the panel commands the controller into MONITOR at connect; that choice shapes polling behavior, not the controller's willingness to respond.

Does a captured frame that passes its checksum prove the cable is good?

No. The checksum only proves the bytes between the panel and your sniffing port were self-consistent, so the transmitter and its serial settings are coherent. It says nothing about whether the controller received the frame, whether the RD return path is intact, or whether handshake inputs at either end are satisfied — those are cable properties, and the continuity meter is the tool for them.

What should you check if both jumpers are in and the panel still times out?

Work in this order: the node number inside the captured frame against the controller's station, then baud, parity and stop bits at both ends, then continuity of TD to RD and the signal ground. Change one item, reconnect, and observe before the next. If nothing moves the symptom, re-measure with the connector closed — a jumper that reads short on the bench can open under a strained backshell.

Keep reading

  • Troubleshooting

    Why an HMI Freezes or Updates Slowly Over Serial

    A panel-wide freeze or slow updates over serial is a link fault — driver, PLC port, wiring, or startup timing — not a screen-object fault. Test the two field-proven fixes first: a 10-second startup delay before the first poll, and a port set to four-wire RS-422 at 19,200 baud. Fix an isolated stale tag separately, by address and data type.

    HMI and touchscreens

  • How-to

    DH-485 Wiring for HMI Replacement: RS-232 Pin Map

    DH-485 is a protocol, not one electrical standard: it runs over RS-232 or RS-485. The serial version on a 9-pin PanelView port is RS-232, and a replacement HMI connects pin-for-pin on pins 2, 3 and 5 to the controller's channel 0.

    HMI and touchscreens

  • Troubleshooting

    Building the RS-232 Modbus cable between an HMI and a controller

    Map the pins by signal name, not by number: the HMI's transmit on 25-pin pin 2 goes to the controller's receive, the HMI's receive on pin 3 comes from the controller's transmit, and signal common runs end to end. The jumper between pins 8 and 12 stays at the HMI connector. Generic straight-through and null-modem cables cause silence because the 25-pin and 9-pin standards number the data lines differently.

    Industrial networks