LHT65N Shows 327.67°C? Preserve the Disconnected-Probe Fault
A dashboard showing 327.67°C from an LHT65N external probe needs a data-quality investigation before anyone treats the number as a process temperature. For the Ext=1 format, Dragino documents 0x7FFF as the result when the external temperature sensor is not connected. The practical response is to preserve the evidence, identify the exact message layout and represent the channel as unavailable while the connection fault is investigated.
Confirm the mode before interpreting the number
Dragino’s LHT65N/S manual describes the Ext=1 E3 temperature format and its disconnected-sensor value of 0x7FFF, numerically 327.67°C after division by 100. Other external-sensor modes and timestamp formats have different layouts. Check model, firmware, application port and external-sensor configuration against the matching documentation; do not apply one byte offset to every packet.
Save a sanitised raw uplink and the decoder revision before making changes. Trace that same message through decoder output, stored telemetry, dashboard and export. The aim is to establish whether the fault value came from the device or was introduced by an incorrect mapping. Device security keys are unnecessary for this investigation and should stay out of ordinary support records.
The photograph identifies the LHT65N housing offered in ProSense Instruments’ external-probe listing. It does not establish which probe, firmware or regional radio variant is installed. Confirm those details from the ordered configuration and device record.

Keep a failed measurement visibly different from a valid one
Use a separate quality field or equivalent application status for the external channel. Retain the raw value for diagnostics, but exclude a confirmed disconnected-probe indication from calculations that assume a valid temperature. Agree how that condition appears in graphs, reports and current-value tiles before deploying the mapping.
Replacing the value with zero invents a measurement. Holding the previous value without an age or quality warning makes an old observation look current. Silently dropping the message can conceal that the device is communicating while the intended sensing point is unavailable. A clear unavailable-channel indication avoids those ambiguities.
The manufacturer’s v1.5.6 decoder contains a separate No_connect indication and external-temperature decoding. Inspect the matching decoder’s complete output rather than forwarding only the numeric temperature into the application. This file is a reference to verify against your firmware, not an instruction to replace every installed decoder.
Separate an instrument fault from a temperature excursion
Agree two response paths with the application owner: one for a valid temperature outside its approved band and another for an unavailable sensing channel. A disconnected probe does not prove that the monitored space is hot, cold or safe. If monitoring is lost, use the site’s established contingency process rather than waiting for a plausible number to return.
Do not widen an operational temperature limit to accommodate 327.67°C. That changes the process rule while leaving the measurement fault unresolved. Similarly, suppressing all high values with a broad filter can hide a genuine excursion. The rejection rule should match verified fault semantics and retain an explicit reason for rejecting the sample.
Prove the behaviour with an isolated replay
Use a spare device or a controlled packet-replay facility that cannot affect operating stock or equipment. Keep the current decoder and mapping configuration so the team can restore them. These are proposed application acceptance checks, not manufacturer-certified test results.
First replay a verified valid external-temperature sample and confirm its channel and unit. Next replay a confirmed Ext=1 disconnected-probe sample. Check that the diagnostic record retains the original evidence, the live external-temperature tile becomes unavailable and the agreed instrument-fault response is visible. Inspect the exported data as well as the screen.
Then replay a fresh valid sample. Confirm the quality state recovers according to the agreed policy, without rewriting the fault interval as valid data. Include a different supported message mode in the test set to ensure the Ext=1 rule is not being applied to unrelated bytes.
A worked example and a useful handover
Hypothetical example: a cabinet’s last valid external reading is 6°C, followed by a verified disconnected-probe indication. The receiving application should not display 0°C as a replacement or continue showing 6°C as fresh. It should show the external channel as unavailable, retain the last valid observation with its age, and raise the agreed fault indication. These numbers illustrate data handling, not recommended storage limits.
Check the physical connection only through the exact product’s supported maintenance procedure and an authorised arrangement. Do not disturb a critical monitoring point without a contingency. A normal housing-temperature reading does not replace the remote probe’s observation of a different location.
For technical assistance, provide ProSense with the model, firmware, probe type, mode, sanitised payload, decoder revision and expected fault/recovery behaviour. Ask for the complete receiving-platform path to be checked. ProSight integration and its handling of channel quality require separate verification for the proposed system.
Need help with your application?
ProSense Instruments - explore instruments, sensors, data loggers and test equipment.
ProSense Solutions - discuss complete monitoring and automation systems.
ProSight Software - explore cloud dashboards, alarms and reporting. Device integration must be confirmed for your application.
Instrument enquiries: sales@prosenseinstruments.com.au | 1800 560 854
Project enquiries: sales@prosense.com.au. Software enquiries: info@prosightsoftware.com.




Comments