A DF1 full-duplex HMI and PLC plan 15 to 20 ft apart on direct RS-232. The 15 m (50 ft) figure at 19200 baud is the standard's cited length and 100 ft is its theoretical limit. Past that, an RS-232-to-RS-485 converter at each end reaches several thousand feet.
A panel and a PLC that talk DF1 full duplex over RS-232 belong within 15 to 20 ft of each other on a plant that is actually running. The 100 ft figure quoted for the RS-232 specification is a theoretical ceiling. The length the standard is commonly cited at is 15 m (50 ft) at 19200 baud, on cable whose capacitance fits the driver's budget and in air quieter than most plants. When the route is longer, change the physical layer: an RS-232-to-RS-485 converter at each end turns the long span into RS-485, which is realistic at several thousand feet. This page sets out the three numbers, what limits the cable, and the decision at each distance.
The distance figures to plan to
There are three numbers and one fix, and each number answers a different question.
| Figure | What it is | How to use it |
|---|---|---|
| 15 to 20 ft | The practical range on a working plant | Plan the direct RS-232 run here |
| 15 m (50 ft) at 19200 baud | The length the RS-232 standard is commonly cited at | The most a direct run should be asked to do |
| 100 ft | The theoretical limit quoted for the RS-232 specification | A ceiling for a tested overshoot, never a plan |
| Several thousand feet | What an RS-485 span between converters covers | The plan past the standard-cited length |
The planning range is field practice, and the standard is the reason it sits so far below the paper figure. RS-232 is single-ended: each bit is one conductor swinging against a signal common, so the driver charges the cable's capacitance on every edge and any induced voltage arrives as if it were data. The 15 m figure at this baud rate is what that capacitance budget allows on paper. In a plant, drive output cables, contactor coils and power ducts eat the margin first, which is why the range you can plan on sits well below it.
What actually limits a direct RS-232 run
Cable capacitance sets the paper limit. A driver moves a fixed amount of charge per edge, a longer cable adds capacitance, and slow edges start reading as errors at the receiver. Low-capacitance cable buys part of that margin back, which is the one way to stretch a direct run honestly.
Noise sets the real limit. A single-ended link carries induced voltage the same way it carries signal, so a drive cable running parallel to the serial cable couples its switching edges straight into the receive pair. That is why a link that holds on the bench drops the moment the line starts. Route the serial cable in its own duct, cross power runs at right angles where crossing is unavoidable, and terminate the shield the way the two device manuals specify.
What DF1 full duplex needs from the wire
DF1 full duplex is the protocol: the panel sends a request, the PLC answers, and the two ends talk point to point over one cable, with a separate conductor for each direction. Selecting the DF1 driver in the panel defines those messages; the wire still sets the distance, and the driver buys no feet back.
Configure both ends to the same baud rate, parity and stop bits, and matching station settings wherever the driver exposes them, and read the values off the configuration screens instead of assuming them. A framing mismatch mimics distance trouble: no answer at all, or answers that will not decode. Message size is never the reason a short run fails, because a typed read or write in the common DF1 implementations carries at most 1000 bytes, which is 500 integers or 250 floats.
Worked example: a 100 ft run at 19200 baud
Inputs:
- Proposed panel-to-PLC distance: 100 ft, the theoretical figure quoted for the RS-232 specification.
- Planned baud rate: 19200 baud, the rate the standard-cited 15 m figure is tied to.
- Conversion factor: 1 ft = 0.3048 m.
The rule in words: plan a direct RS-232 run against the field range from the table, treat the standard-cited length at this baud as the most the cable should be asked to do, and convert to RS-485 past it.
Arithmetic: 100 ft × 0.3048 m/ft = 30.48 m. Against the standard-cited figure, 30.48 m ÷ 15 m = 2.03, so the proposed run is about twice the length the standard is cited at. Against the field range, it is five times the planning number.
Result: the run fails both tests, and no cable choice turns 30.48 m at this baud rate into a planned direct RS-232 run.
Decision: build the link as panel, short RS-232 stub, converter, RS-485 span, converter, short RS-232 stub, PLC. Give each stub a length inside the planning range, and give the long span to RS-485, which covers several thousand feet. The threshold to carry into the next job: past the standard-cited length of planned route, fit converters.
Spanning the distance with RS-485 converters
Two interface converters, one at the panel port and one at the PLC port. RS-232 on the stubs, RS-485 between them. RS-485 sends each bit as the voltage difference between two wires, so noise arriving on both wires cancels at the receiver, and the span is realistic at several thousand feet, trading length against baud rate.
Wire each converter to its own manual: end-of-line termination, shield and signal common, and the specified supply. One hard rule: never land RS-485 terminals on the RS-232 port's pins. The two electrical standards are incompatible and the port can be damaged. With power isolated as the manuals require, build the link, then prove it with the line running: issue normal reads and writes, start the drives, and watch the panel's communications status through a full cycle. A link proved on an idle line only has not been proved.
If the RS-485 segment will carry more devices later, the sharing rules tighten; the checks in two Modbus RTU masters on one RS-485 bus walk through a shared segment.
Move the panel or change the network
Distance is also a mounting decision. A panel close to the PLC with a short cable is the cheapest link you will ever build, and if the operators only need to watch a few live values, displaying a PLC counter value where operators can see it starts at the display end.
Panels and PLCs that also speak DH-485 have a second road: DH-485 runs over RS-485 wiring through interface converters, so distance behaves like the link above and node addressing joins the plan. The walkthrough for connecting an operator interface to a PLC over DH-485 covers the wiring and the addresses.
When the distance problem is really about I/O signals and not the panel's conversation, remote I/O puts the cards near the field devices; the page on remote I/O solutions for distances up to 1000 feet outlines that approach.
And if an existing RS-232 run only needs a few more feet, the checks in extending an RS-232 HMI cable safely cover what the extra cable adds and how to prove it.
The decision at each distance
- Up to the planning range: direct RS-232, DF1 full duplex, shielded cable in its own route. Build it.
- Between the planning range and the standard-cited length: the tested exception. Low-capacitance cable, a clean route and a witnessed test with the line running; if any of the three is missing, fit converters.
- Above the standard-cited length: a converter at each end, the long span on RS-485, stubs inside the planning range.
- Never: RS-485 landed on RS-232 port pins.
Recommendation: plan the panel within 20 ft of the PLC and run DF1 full duplex over direct RS-232 with shielded cable routed away from power wiring. When the planned route exceeds the standard-cited 50 ft length, fit an RS-232-to-RS-485 converter at each end, keep each RS-232 stub inside the planning range, and prove the link with the plant running before you call it finished.
Questions
Does a lower baud rate let a direct RS-232 DF1 link run farther?
Yes, within limits. The 15 m (50 ft) figure is tied to 19200 baud, and the real constraint is cable capacitance plus route noise, so dropping the port to 9600 baud buys margin on a marginal run. It will not rescue a route at the theoretical ceiling. Above the planning range a direct run is a tested exception, and beyond the standard-cited length the converter pair is the plan.
Can special cable make a long direct RS-232 run work?
Low-capacitance cable is the one lever that lengthens a direct run, and cable makers publish reaches up to 1000 ft (300 m) with it. The published reach assumes that specific cable and a quiet route, so a long direct run is something you test with the plant running. For new work past the standard-cited length, the RS-485 converter pair costs less than the risk.
How large can a DF1 message be, and does size change the distance limit?
In the common DF1 implementations, one typed read or write carries up to 1000 bytes, which is 500 integers or 250 floats. At 19200 baud and 10 bits per byte, a full 1000-byte transfer occupies the wire for about 0.52 s, a timing question rather than a distance question. The distance limit is set by capacitance and noise on the cable.
Planning information, not legal or engineering advice. Rules differ by place: check codes and permits with your local authority and your licensed trades.