.png)
Modbus RTU Gap Errors Above 19200 Baud: Check the Timing Rule
A Modbus RTU link behaves at one serial speed but rejects frames after a baud-rate change. Before changing registers or adding retries, separate the gap inside a frame from the silence between frames. They serve different purposes and are not the same as an application response timeout.
Key takeaways |
Measure the serial timing at the relevant interface. An application log’s timestamp spacing may reflect buffering rather than the gaps seen by the RTU receiver. |
1. Identify which timer is failing
Record the exact endpoint models, firmware, baud rate, parity, stop bits and the documented driver timing configuration. Keep the serial-to-network path visible. A network packet capture alone may not show the spacing of individual characters at the serial receiver.
The Modbus serial guide describes t1.5 for an inter-character timeout and t3.5 for frame separation. Its guidance distinguishes rates up to 19,200 bit/s from higher rates, where it recommends fixed values of 750 microseconds and 1.750 milliseconds respectively. Confirm how the actual driver implements that guidance.
For application scoping, start with ProSense networking and gateway solutions. Confirm the actual supported interface and acceptance requirements; the generic protocol explanation does not establish product compatibility.
2. Keep frame timing separate from response waiting
A response timeout governs how long a client waits for the server’s reply. It is not a substitute for correct framing of transmitted bytes. Increasing that timeout does not necessarily repair an unwanted gap inserted inside the outgoing request.
Similarly, an inter-frame interval is not an instruction to insert that same delay after every byte. A configuration screen can use imprecise labels, so identify the setting’s documented meaning before adjusting it. Preserve the previous values and the reason for a change.
Hypothetical arithmetic check: with an explicitly assumed 11-bit serial character at 9,600 bit/s, one character takes approximately 1.146 milliseconds and 3.5 characters approximately 4.01 milliseconds. This calculation illustrates the time basis; it is not a recommendation to apply the same scaled formula at every higher baud rate.
3. Test at the interface that owns the gap
Use a suitable isolated test arrangement and capture equipment under the site’s access and electrical procedures. Ask a competent technician to establish the serial observation point. Do not expose live terminals or remove protective barriers to obtain a convenient trace.
Compare a known complete request at the selected speed with the driver’s accepted response. Retain the raw frame and timing evidence. If a converter buffers data, document whether it forwards a complete serial frame or releases portions as network data arrives.
Do not infer a serial timing fault solely because application timestamps are several milliseconds apart. Scheduling, buffering and timestamp resolution can alter a log without establishing the physical serial gap. Use the evidence appropriate to the boundary under investigation.
4. Repeat the acceptance checks after a change
Change one documented configuration variable at a time in a controlled test. Repeat normal reads, boundary conditions and an intentional malformed fixture where supported. Retain valid rejection behaviour as well as successful results; accepting damaged frames is not an improvement.
Separate CRC failures, incomplete-frame errors, no reply and protocol exceptions in the record. Each points to a different stage. If the device does not expose those diagnostics, document the limitation rather than inventing precision from a generic timeout count.
Before restoring operational polling, verify serial settings at both ends, the complete converter path and the application’s intended values and units. Timing acceptance establishes a communication condition; it does not validate the instrument calibration or a certified machine safety function.
5. Diagnostic checks
Observation | Check next |
Fails only after a speed change | Review driver timing guidance and endpoint settings. |
Longer timeout changes nothing | Check framing rather than waiting time. |
Log shows irregular timestamps | Confirm buffering and timestamp resolution. |
Frames pass; values are wrong | Investigate mapping and measurement separately. |
6. Recommended equipment / related solutions
ProSense automation solutions suit projects that need a documented device-to-application boundary, controlled testing and a complete acceptance record. ProSense Solutions is a solutions provider; exact hardware and protocol support must be confirmed.
Related guide: understand the limits of a Modbus echo test. A valid communication exchange and a proven application measurement are different acceptance questions.
Next step: request application and system-selection advice. Provide exact equipment, firmware, the documented interface and sanitised test evidence.
7. Practical questions
Is t3.5 the response timeout?
No. Frame separation and response waiting are separate controls.
Can I infer byte gaps from a cloud log?
Usually not without knowing buffering and timestamp behaviour.
Primary reference: the official protocol specification. Use the relevant sections and exact product documentation when implementing the checks.
Need help with your application?
Email: sales@prosense.com.au
Email: info@prosightsoftware.com
Phone: 1800 560 854
