Short answer
Commissioning is the acceptance test that stands between an installed device and a trusted number. It verifies four things: the calibration curve rises consistently, the interface pair genuinely matches, the ingress rating suits the mounting position, and a baseline reading exists to compare every later figure against.
Key takeaways
- A calibration curve that goes up and down instead of consistently in one direction is a fault to find before continuing, per Teltonika's own documented diagnostic.
- RS-232 and RS-485 are not interchangeable: our FMC003 record lists 1× RS-232, while the OMNICOMM sensor is RS-485 Modbus RTU.
- An IP rating grades solids and liquids only, the IEC code says nothing about temperature, so it cannot justify a hot mounting position.
- Our catalogue shows how wide the ingress spread is on one site: FMC003 is IP41 while the OMNICOMM sensor is IP68.
- A baseline reading recorded at handover is the only reference a later drift complaint can be measured against.
01What commissioning is for
Commissioning is the step between a device being physically installed and its readings being treated as facts. It is easy to skip because nothing appears to be wrong: the device powers up, a position arrives on the map, and the installation is declared finished. The problem is that a wrong specification produces plausible data. A sensor on the wrong interface stays silent, which is at least visible. A sensor with a bad calibration table reports numbers that look entirely reasonable and are simply not true, and those numbers propagate into consumption reports, theft alerts and maintenance schedules for months. The economics of commissioning are therefore lopsided in one direction: an error found at handover costs an hour on site, while the same error found in a quarterly report costs a site visit plus every decision taken on the bad data in between.
02Check one: the calibration curve must be monotonic
The first check needs no equipment beyond the calibration screen itself. Teltonika's documented analog fuel sensor procedure has the installer fill the tank in five to ten per cent increments with roughly thirty seconds of settling per step, and the output is a set of cubic coefficients. The diagnostic that matters is stated plainly in that documentation: the recorded voltages must change consistently in one direction, and values that go up and down mean there is a fault to find before continuing. That instruction is worth repeating because it converts a subtle problem into a visible one. A non-monotonic curve is evidence of a loose connection, a probe touching the tank wall, air in the sensor, or a wrongly ordered set of reference points, none of which the resulting fuel figure would announce on its own.
03Check two: the interface pair, not the interface name
Serial interfaces sound interchangeable in a quotation and are not. The FMC003 catalogue record lists one RS-232 serial port; the OMNICOMM fuel level sensor in the same catalogue is RS-485 Modbus RTU. Those are different electrical standards, and a sensor speaking one will not appear on a port expecting the other. The catalogue is explicit about which pairing works, the FMC003 record names its RS-232 port as connecting directly to OMNICOMM fuel sensors, but the general lesson is to confirm the pair as a pair at commissioning rather than confirming each side separately. The same discipline applies to CAN: the FMC003 record notes that CAN bus is optional and reached via a separate FMB0YX adapter, so a CAN requirement discovered on site is a missing part, not a configuration setting.
04Check three: the ingress rating against the mounting position
The mounting position is decided on site, and it has to be checked against the device's own rating rather than against the installer's judgement of how sheltered a spot looks. The spread within one catalogue makes the point: our FMC003 record is IP41 while the OMNICOMM sensor is IP68. Those two devices belong in very different places, and the second numeral is where the difference lives. It also matters what the code does not cover. The IEC ingress protection code grades the first numeral 0–6 for solids and the second 0–9 for liquids under IEC 60529, and temperature appears in neither. A high IP number therefore says nothing about whether a position near an exhaust or on a black surface in the sun is survivable, which is a separate specification with its own limits.
05Check four: record the baseline before you leave
The last check is administrative and the most frequently skipped. A baseline is the set of readings taken at handover with their conditions written down: tank level and the volume that produced it, engine hours as the counter reads them, the ambient temperature, the input states, the firmware version and the exact mounting position. Its whole value is comparative. When somebody reports six months later that the fuel figure looks wrong, the only way to answer is to compare against a known-good starting point, and without a baseline the investigation degenerates into opinion. The baseline also settles the ownership question that always follows a drift complaint: a reading that has moved away from a recorded baseline is a maintenance problem, while a reading that was never right is an installation problem, and those have different remedies and different owners.
06The metadata the readings depend on
Some commissioning entries are not readings at all but numbers the readings will be divided by later, and a wrong one corrupts everything downstream silently. Tank geometry is the clearest example. A horizontal cylindrical tank holds 19.55 per cent of its capacity at a quarter of its height, not 25 per cent, so treating level as linearly proportional to volume is wrong by several percentage points across most of the range. Recording the tank shape and the capacity at commissioning is what allows the platform to apply the right relationship. The same applies to the calibration table: the Teltonika procedure produces cubic coefficients specific to one tank in one vehicle, so those coefficients are a property of the installation and belong in the record with the installation, not in a shared device template.
07Turning the checklist into a handover document
A checklist that lives in an engineer's head is not a control. The useful form is a short signed record per asset, and the fields follow directly from the four checks: the calibration points with a note that they were monotonic, the interface pair as installed, the device rating alongside the chosen mounting position, and the baseline readings with their conditions. Add the firmware version and the input assignments, because both change the meaning of later data and neither can be reconstructed afterwards. The document does not need to be long. It needs to be per asset rather than per fleet, because every one of these fields differs between two vehicles that look identical, and the value of the record is precisely in the differences it captures.
| Check | What you verify | Failure signature if skipped | Cost of finding it later |
|---|---|---|---|
| Calibration curve | Readings change consistently in one direction | Plausible but wrong volumes | Every report since handover |
| Interface pair | RS-232 to RS-232, RS-485 to RS-485 | Sensor silent, or absent from the record set | One site visit, plus a wrong part |
| Ingress rating | Rating matches the chosen position | Water or dust ingress weeks later | Device replacement and lost history |
| Baseline reading | A known-good starting point exists | No way to prove drift either way | An unresolvable dispute |
Figures and sources
- Teltonika monotonicity diagnostic during calibration: voltages must change consistently; up-and-down values mean a fault to find before continuing (Source: Teltonika Telematics Wiki, Analog Fuel Sensor, as of 2026-08-31)
- Teltonika documented calibration procedure and the polynomial it produces: 5–10% tank increments, ~30 s settling; yields cubic coefficients a0, a1, a2, a3 (Source: Teltonika Telematics Wiki, Analog Fuel Sensor, as of 2026-08-31)
- Teltonika FMC003 serial interface, as listed in our catalogue, not interchangeable with RS-485: 1× RS-232 (Source: ENTEK.AI product catalogue, as of 2026-08-31)
- OMNICOMM fuel level sensor interface, as listed in our catalogue: RS-485 Modbus RTU (Source: ENTEK.AI product catalogue, as of 2026-08-31)
- Ingress rating spread between two devices on the same installation: Teltonika FMC003 IP41; OMNICOMM fuel level sensor IP68 (Source: ENTEK.AI product catalogue, as of 2026-08-31)
- What the two IP numerals grade, temperature is absent from both: first numeral 0–6 solids; second numeral 0–9 liquids; IEC 60529 (Source: IEC, Ingress Protection (IP) ratings, as of 2026-08-31)
- Volume held by a horizontal cylindrical tank at quarter height (exact circular-segment geometry): 19.55% of capacity at h/D = 0.25, and 80.45% at h/D = 0.75 (Source: Computed from V/Vtot = (θ − sin θ)/2π, θ = 2·arccos(1−2h/D); ENTEK.AI computation, independently reproducible, as of 2026-08-31)
- Teltonika FMC003 CAN bus availability, an added part rather than a setting: CAN bus optional, via a separate FMB0YX adapter (Source: ENTEK.AI product catalogue, as of 2026-08-31)
Questions
What exactly makes a calibration curve unacceptable?
Readings that do not change consistently in one direction. Teltonika's documented procedure fills the tank in five to ten per cent steps with about thirty seconds of settling, and states that voltages going up and down mean there is a fault to find before continuing, typically a loose connection, air in the sensor, or a probe touching the tank wall.
Can I connect an RS-485 sensor to an RS-232 port with an adapter?
That is a conversion, not a connection, and it becomes a component in the installation with its own power, failure mode and documentation. The cleaner answer at commissioning is to confirm the pair matches as specified, our FMC003 record lists 1× RS-232 and names OMNICOMM fuel sensors as connecting directly to it.
Is a high IP rating enough for an exposed position in the Gulf?
No, because the IP code does not grade temperature at all. Under IEC 60529 the first numeral covers solids from 0 to 6 and the second covers liquids from 0 to 9, and neither addresses heat. A position near an exhaust or on a sun-facing surface has to be checked against the device's stated temperature range separately.
What should a baseline record contain?
The readings and the conditions that produced them: tank level with the volume it represents, the engine-hours counter as it reads, ambient temperature, input states, firmware version and the mounting position. Its only purpose is comparison, so a value without its conditions is not a baseline, it is just a number in a file.