Troubleshooting
Why DF1 reads work on an SLC 500 but not a PLC-5
Short answer
The DF1 link layer is identical on both processors, so your cable, baud and channel settings are not the fault. The difference is the PCCC command dialect: SLC 500 and MicroLogix answer protected typed logical reads that address data by byte offset, while the PLC-5 answers typed reads that address words. An RTU speaking only the SLC dialect reads nothing from a PLC-5 — set it to PLC-5 mode or get PLC-5 support from the vendor.
On this page
- What is the flickering COMM light telling me?
- Is the 2-3-5 to 2-3-7 cable the right null modem?
- How do I prove the PLC-5 before blaming the RTU?
- Which CH0 settings can stop the PLC-5 replying?
- Is a PLC-5 N register different from an SLC 500 N register?
- What exactly differs in DF1 between the two processors?
- How do I capture what the RTU actually sends?
- What is the fix when the RTU speaks only the SLC dialect?
- Questions asked
Your cable and channel settings are almost certainly not the fault. The DF1 link layer is the same protocol on both processors, and a COMM light that flickers once per poll proves characters are arriving and framing. The difference sits one layer up, in PCCC: an SLC 500 or MicroLogix answers a protected typed logical read that addresses data by byte offset, while a PLC-5 answers its own typed read family that addresses data by word. An RTU that speaks only the SLC dialect sends a command your PLC-5 will not execute, so every polled value sits at zero while the RTU reports comms healthy. If the RTU has a PLC-5 mode, set it; if it does not, no CH0 setting will fix this — you need vendor driver support or a translating device in between.
What is the flickering COMM light telling me? #
Characters are reaching the port and the UART is clocking them in, so baud, parity, stop bits and the transmit path are alive; the light cannot show whether the processor ever executed the command inside the frame.
A blink at the poll interval followed by an RTU timeout is precise information. A flag reading DF1 comms status good reports the link layer only — bytes framed, checksums passing, an acknowledgment returned — and says nothing about the read itself, which completes in a separate application reply. A poll, its acknowledgment and the reply each cost a round trip on a single-master line, the same arithmetic that makes Modbus RTU drive polling on one serial port slow, so the rhythm you are watching is normal, not a fault.
Do not re-terminate the cable and do not drop the baud: taking CH0 from 19.2k to 1200 only slows the blink and never fixes a fault at this layer. Triage what you actually see:
- COMM flickers each poll, RTU status good, all values zero — the link answers, the application does not; capture the wire.
- COMM flickers, the RTU retries, then faults the port — BCC or CRC mismatch; the processor discards the frame without replying. Compare Error Detection on both ends.
- COMM dark, no activity at all — baud or parity mismatch, or an open transmit line or ground; run RSLinx Auto-Configure first.
- First poll answered, everything after ignored — Duplicate Packet Detect dropping an RTU that reuses its transaction number; disable it as a test.
- Receive activity only, the PLC-5 never transmits — Control Line left in a modem mode with no handshake path on three wires.
- Values arrive but they are the wrong words — address fields resolved as words against bytes; see the dialect section.
Is the 2-3-5 to 2-3-7 cable the right null modem? #
Yes — that pinning is the correct crossover, because a DE9 DTE port receives on pin 2 and transmits on pin 3 while a DB25 DTE port transmits on pin 2 and receives on pin 3, so the matching numbers already cross over.
Ground rides pin 5 on the 9-pin end to pin 7 on the 25-pin end; do not leave it open, because a missing signal ground produces intermittent framing errors that imitate a protocol fault. Once RSLinx has been online on that cable, the data path is settled. Stop rewiring it.
Handshaking is the one cable-side trap. If Control Line on CH0 is set to anything other than No Handshaking, the PLC-5 waits on CTS before keying its transmitter, and three wires give it nothing to wait on: it receives the poll, the light flickers, and it never answers. Either set Control Line to No Handshaking, or jumper the handshake pairs at the 25-pin end — 4 to 5 (RTS to CTS) and 6 to 8 to 20 (DSR to DCD to DTR).
How do I prove the PLC-5 before blaming the RTU? #
Take the RTU out of the path and drive the processor from a PC: configure an RS-232 DF1 driver in RSLinx, run Auto-Configure, browse to the PLC-5, and read the exact words the RTU scans — N30:0 and N31:0, one element each.
Set the driver up as a DF1 polling master addressed to the same station number the RTU uses, and let it poll. If RSLinx reads those words over that cable, the processor, its CH0 configuration and the wiring are cleared end to end, and everything after this point lives on the RTU side. Confirm the data table while you are there: a request for element 1 of a one-element file draws an error reply, and a polled file number that does not exist draws an error status, not zeros. Proving the path with the vendor's own serial tool before touching third-party gear is the same first move as commissioning CP1L Host Link serial reads to SCADA.
Which CH0 settings can stop the PLC-5 replying? #
Five fields decide it — Mode, Protocol, Station Address, Error Detection and Duplicate Packet Detect — and Control Line left in a modem mode will silence the transmitter entirely on a three-wire cable.
Replicating the SLC configuration field for field is where this fault hides, because the PLC-5 channel configuration carries fields the SLC 500 configuration does not:
| CH0 field | What a mismatch does |
|---|---|
| Mode | User mode is ASCII; DF1 never runs whatever else you set. Leave it in System mode. |
| Protocol | A half-duplex DF1 slave only transmits when a half-duplex master polls it, so a full-duplex RTU never sees its replies. Match the RTU's mode. |
| Station Address | A poll addressed to a node the processor does not own is never answered. Confirm the radix each tool displays before comparing numbers. |
| Error Detection | BCC on one end and CRC on the other discards frames silently: no reply, no visible NAK, master times out. |
| Duplicate Packet Detect | An RTU stack that reuses its transaction number has every poll after the first dropped. Disable it as a test. |
| Embedded Responses | Replies packed in a form a third-party parser ignores on a point-to-point link. |
| ACK Timeout, ENQ retries, NAK retries | Premature retries collide with the reply and trip duplicate suppression. |
| Diagnostic File | Not a fault. Assign one — it counts retries and NAKs and is your instrument for every row above. |
Make these changes over the Ethernet path you already have, never through the serial port you are online on: reconfiguring CH0 under your own session drops the connection mid-write.
Is a PLC-5 N register different from an SLC 500 N register? #
No — an N-file element is a 16-bit signed integer on both platforms and N30:0 names the same storage either way; the register is not the problem, the command that reaches it is.
Your one-element files exist and read correctly, which RSLinx just proved on the same cable. The split shows up in software, not in the data table: a Logix MSG instruction makes you choose between PLC5 Typed Read and SLC Typed Read for the same destination, and telemetry DF1 interface cards have long carried a physical PLC-5 versus SLC 500 register selector — the PR27 card in a Serck 2000M is one example. An RTU with no such switch was built to speak one dialect. The message-type choice and the address fields it carries are laid out in this DF1 protected typed logical read MSG configuration guide.
What exactly differs in DF1 between the two processors? #
The link layer is identical; the PCCC command subsets differ — the PLC-5 serves its typed read family with word logical addressing, while SLC 500 and MicroLogix controllers serve the protected typed logical read family, whose three address fields count bytes.
Which command families your processor answers is documented per command in the DF1 Protocol and Command Set reference, publication 1770-RM516, including its byte-versus-word addressing discussion. Read it for your processor instead of assuming the SLC 500 command set transfers. The same choice follows these platforms onto Ethernet, where PCCC is carried over the network stack — the decision made when a SCADA client reads a small PLC over EtherNet/IP.
Worked example, labeled illustrative because it depends on the RTU's internals. An integer element is 2 bytes, so in the SLC dialect the byte offset for element n of an N file is n × 2. To read element 3, the RTU sends byte offset 3 × 2 = 6. The PLC-5 typed read reaches that same element at word address 3. A 6 landing where the processor resolves words points at element 6, three elements deeper than intended — and on a one-element file any nonzero offset is past the end, which returns an error status, not zeros. That arithmetic is the whole fault, and it is why a stack with no word mode cannot be aimed at a PLC-5 by editing file numbers.
How do I capture what the RTU actually sends? #
Put a PC in line on a breakout or monitor cable, log both directions with a serial analyzer, and decode the fields in order: DLE STX, DST, SRC, CMD, STS, TNS, FNC, then the address bytes.
DF1 is only the envelope — DLE STX, payload, DLE ETX, checksum, then DLE ACK, DLE NAK or ENQ at the link layer — and PCCC rides inside it carrying the read. Two questions in the capture decide the case. Does the PLC-5 transmit an application reply at all? If it acknowledges the frame at the link layer and never sends one, the processor is refusing that command family and the dialect fault is confirmed. If a reply does come, is STS non-zero? STS is the status byte the processor sets when it accepts the frame but cannot execute the command; its meaning is in the command set reference. Either way you leave with the name of the command the RTU sends, which is exactly what your vendor will ask for.
What is the fix when the RTU speaks only the SLC dialect? #
Stop editing CH0 — the fault is not there. The decision is whether your RTU has or can acquire PLC-5 typed read support; if it cannot, you put a translator in between or you replace the RTU.
- The RTU has a PLC-5 mode or a register-family selector: set it to PLC-5, keep the proven 2-3-5 to 2-3-7 cable, and expect values on the next poll cycle. Nothing else changes.
- The RTU has a driver option or firmware update: get PLC-5 typed read support with word logical addressing from the vendor, then retest with the analyzer — the command family change shows up in the CMD and FNC fields.
- Neither exists: interpose a small SLC-family processor that reads the PLC-5 over its own link and serves the SLC dialect to the RTU, or replace the RTU with one that documents PCCC support for both families.
The link layer you spent the morning proving will carry either dialect without complaint, which is why every channel edit looked promising and none of them worked.
Questions asked
Does a DF1 acknowledgment mean the read succeeded?
No. The acknowledgment your RTU counts as good comms is a link-layer DLE ACK, which only confirms the frame arrived intact. The read completes in a separate application reply frame that carries the data plus an STS status byte. An RTU can hold a healthy status flag while silently discarding replies it cannot parse, which is exactly the good-status, zero-values pattern.
Can I work around the difference by remapping my data files?
No. The file numbers are not the fault: N30:0 and N31:0 exist and read correctly on the PLC-5. Remapping changes which file a request names, not which command family carries it. If the RTU only issues protected typed logical reads and your processor answers a different family, no data table layout makes the values appear. The fix is in the RTU driver, not the data table.
When would changing baud or parity actually help?
Only when the port is dark or RSLinx Auto-Configure will not lock on, because those are framing and electrical symptoms. With the COMM light flickering at the poll interval and the RTU reporting healthy status, baud, parity and stop bits are already proven end to end. Dropping from 19.2k to 1200 baud only slows the blink.
What exactly should I ask the RTU vendor for?
Ask which PCCC commands their DF1 driver implements, then request support for the PLC-5 typed read family with word logical addressing, the form documented for your processor in publication 1770-RM516. Vendors deliver this as a driver option or a firmware update; older telemetry cards did it with a physical PLC-5 versus SLC 500 switch.
Should the PLC-5 port be half duplex or full duplex?
Match the RTU, not the old SLC setup. A full-duplex master runs point to point and expects a reply on the same pair; a half-duplex master polls and the slave transmits only when addressed, which is how multi-drop telemetry radios work. Both pairings move bytes and link acknowledgments, so the wrong pairing still looks healthy while data never lands.