LDDS75 Shows 20 mm? Reject Invalid Distance Before Calculating Tank Level

A tank display suddenly jumps to almost 100%, yet the independent level check shows no corresponding rise. Before changing the empty distance or smoothing the trend, inspect the original distance field and its quality. A received packet can contain an invalid measurement.

For the Dragino LDDS75, a 20 mm distance is a particularly important clue. Validate the model, firmware and raw value before converting distance into level. Otherwise, ordinary level arithmetic can turn a fault indication into a convincing full-tank number.

What 20 mm means on this model

Dragino’s LDDS75 documentation specifies a flat-object measurement range of 280–7500 mm. Section 2.3.2 explains that values below 280 mm are invalid; from firmware v1.1.4, those values are represented by 0x0014, or 20 mm. It separately identifies 0x0000 as meaning the system does not detect the ultrasonic sensor.

These are model-specific rules. Record the actual firmware and use its matching payload definition; do not apply an LDDS75 rule automatically to DDS75-LB/LS or another distance sensor. The manufacturer’s LDDS75 product page identifies the device and available regional variants. Confirm the supplied variant and network plan separately for an Australian installation.

Manufacturer product photograph of the Dragino LDDS75 enclosure, antenna and cabled ultrasonic probe. This is a product photograph, not a customer installation.

Worked example: an invalid value becomes 99%

Consider a hypothetical vertical tank whose zero-level reference is 2500 mm below the sensor face. Its configured full level is 2000 mm above that reference, leaving a 500 mm air gap at full. For a valid reading, level equals the empty-reference distance minus the measured air gap.

At a valid 900 mm gap, level is 2500 − 900 = 1600 mm, or 80% of the configured 2000 mm span. If software blindly substitutes the invalid 20 mm code, it calculates 2480 mm and 124%. A display that caps every result at 100% then conceals the problem behind a plausible full indication.

With a different, equally common configuration that divides by the entire 2500 mm reference distance, the same invalid value produces 99.2%. Neither result establishes that the liquid reached the top. The arithmetic is working on an input that should have been rejected.

Put the validity check before level arithmetic

Use this commissioning sequence with a controlled test dataset or an authorised non-operating test environment:

1. Capture the original uplink, receive time, decoded distance and firmware identification. Retain this evidence before changing the decoder or tank configuration.

2. Confirm which bytes represent distance for that firmware. Keep the native millimetre value available alongside any converted metres or percentage.

3. Classify sensor-not-detected and invalid-range values before subtracting distance from the empty reference. Produce a distinct invalid quality state rather than a replacement level of zero or full.

4. Check the remaining valid distance against the application’s expected geometry. A value inside the instrument range can still be implausible for this tank.

5. Apply the level and volume conversion only to accepted measurements. Test the operator display, trends, exports and alarm logic with the same invalid dataset.

The preferred response to missing or invalid data depends on the process. Agree it with the responsible engineer and operators. If a display retains the last good level, show its timestamp, age and current invalid status prominently; an old value must not appear to be a fresh successful measurement.

Distinguish a measurement fault from a radio outage

A new packet containing an invalid distance calls for a different investigation from no packet arriving at all. Preserve both communication freshness and measurement quality. Treating every received message as proof of a healthy measurement loses that distinction.

After recording the evidence, competent personnel can assess the sensor’s mounting position, closest expected surface and clear measuring path under the site’s safe access arrangements. Do not enter a tank, reach into operating equipment or force a real high-level condition merely to reproduce a code. Software test inputs can verify the conversion chain without changing the process.

Common fixes that hide the evidence

Do not recalibrate the empty distance to make an invalid sample look reasonable. Do not average a sentinel value together with good measurements, clamp it to 100%, or discard it silently from a report. Each approach can obscure how long the measurement was unavailable.

An independent high-level safeguard must remain a separately engineered and validated function where the application requires one. This guide describes monitoring data handling; it does not establish the suitability of a wireless sensor or dashboard as an overfill protection system.

Before handover, save a small acceptance pack containing one valid sample, the relevant invalid and sensor-missing cases, the expected quality states, and screenshots of the resulting displays. Include the firmware, decoder revision, tank reference dimensions and agreed response to stale data. That gives the next technician a repeatable test instead of a mysterious percentage.

Need help with your application?

Discuss the sensor model, firmware, tank dimensions and fault behaviour with ProSense Instruments. Email sales@prosenseinstruments.com.au or call 1800 560 854.

For complete monitoring and automation projects, contact ProSense Solutions at sales@prosense.com.au. Explore ProSight Software or email info@prosightsoftware.com for software enquiries. Device integration and application suitability must be confirmed separately.