Resources / Application

TPMS After a Power Cycle: Data Returns in About 4:44

· 6 min read

The question every AGV operator asks at shift start

AGVs and construction machines are switched off at the end of every shift, and the receiver goes dark with the machine. That makes one question decisive for whether a fleet actually uses tyre pressure data or quietly ignores it: when the power comes back, does somebody have to walk out and re-pair the sensors?

This is the question our port, mining and multi-axle customers ask most often, and it is the one we could not answer honestly in a single sentence until we measured it. The sister article - packet loss by wheel position - covers what happens while the machine is running. This one covers what happens after it stops.

The short answer: no re-pairing, and about 4 min 44 s

No. Nobody has to re-pair anything.

On the machine we tested, all eight sensor readings dropped to zero at power-off and came back on their own after power-on. No reset, no pairing mode, no operator action. The last wheel position had data between 4 min 18 s and 5 min 17 s after power-on, and the average across five power cycles was about 4 min 44 s.

That is the number to put into a manual or an acceptance test. The rest of this page shows where it comes from.

Five power cycles, five recovery times

One machine, five power cycles, logged at the CAN output of the receiver - TPMS frames on protocol 0x18FDB1D8, in 29-bit extended frames:

TestTPMS framesLast wheel position has data
11,7045 min 17 s
21,3464 min 18 s
31,5684 min 42 s
41,6854 min 59 s
51,6074 min 26 s

Shortest 4:18, longest 5:17, average about 4:44. Every run came back without a single manual step; the only thing that changed between runs was the time it took.

Why it is not a fixed five minutes

This is the part worth reading twice, because "about five minutes" invites the wrong conclusion: that the system needs a fixed five minutes and that anything faster is luck.

Across the five runs, the slowest wheel position was not the same position.

  • Run 1 - position 31, 316.6 s
  • Run 2 - position 11, 258.2 s
  • Run 3 - position 11, 281.7 s
  • Run 4 - position 31, 298.3 s
  • Run 5 - position 21, 266.3 s

If one wheel were weak - a bad sensor, a shadow behind a steel plate, a poor antenna position - it would be last every run. It is not. Position 31 was last twice, position 11 twice and position 21 once, and the spread between the first and the last position inside a single run is tens of seconds, not minutes.

What is going on is scheduling, not signal strength. Each sensor runs its own wake-up cycle, and every transmission carries a random delay before it goes on air. That delay is deliberate: it is the anti-collision design that keeps eight sensors on one machine from talking over each other. It is not drift, it is not a defect, and it is not something a firmware revision should remove.

The consequence is that recovery time is a distribution, not a constant. Which makes this the honest sentence to put in a specification:

"About five minutes - 4 min 18 s to 5 min 17 s, all normal."

Anyone quoting a tighter number is quoting one lucky run, or has not run the test five times.

What was on the machine: 16 slots, 8 sensors

One detail changes how the numbers should be read: the machine has 16 wheel-position slots and 8 sensors fitted - two outer positions per axle, inner slots empty. The figures above describe those eight sensors, not a fully populated machine.

The test machine was an XCMG unit in service. A 16-slot machine with 8 sensors and a 16-slot machine with 16 sensors do not put the same load on the air, and the air is what the random delay is managing. On a fully populated machine each sensor still reports on its own cycle, but we would rather state what we measured than generalise from it.

The receiver here is our standard CAN unit: CAN 2.0B, 11-bit and 29-bit frames, RS485 Modbus RTU as a slave only, RS232 for a local display, IP68 housing. Sensors cover 0–1400 kPa gauge (100–1500 kPa absolute) at ±7 kPa, and the system addresses up to 256 wheel positions. CE marking is in progress.

What this means for your manual and your acceptance test

Two documents benefit from this number.

The manual. "After power-on, tyre data returns automatically in about five minutes; no pairing or reset is required." One line like that removes the most common support call on a fleet that is powered down every day, and it tells the operator to stop checking the screen at shift start and to trust it a few minutes later.

The acceptance test. Write "power cycle: automatic recovery in about five minutes, no operator action" instead of "requires manual intervention". The second wording ends up in a maintenance procedure and costs labour every shift. The measured spread does something better: it tells the customer what normal looks like, so a recovery at 4 min 30 s is not reported as a fault.

For the layout side - receiver placement, repeater routing, multi-axle coverage - the method is on the AGV TPMS solution page and on AGV installations. The same behaviour applies to any receiver that is powered down with the machine, which is most of the port machinery we see.

Ask for the raw CSV logs

We are not asking you to take five runs on trust. Ask me for the raw CSV logs - five files, CAN frames, and the script that extracts the last-wheel timestamp.

```

Tell me the machine type and the tyre count, and I will say what this data

implies for your layout. If we have not measured your class of machine, I will

say so instead of borrowing a number from this one.

```

Contact us for the logs. The products overview lists the receivers they were recorded from.

FAQ

Do the sensors need re-pairing after a power cycle?

No. Sensor identity is retained and the receiver rebuilds the wheel map on its own; nothing is entered by hand and no pairing mode is used.

Is 5 min 17 s a fault?

No. The five runs spread from 4 min 18 s to 5 min 17 s and the slowest position moved each time, which is the signature of the random transmit delay rather than a weak sensor.

What if one position never comes back?

Then look at the installation, not the recovery logic - mounting, battery, or steel in the signal path. A position that misses every run is the case to send us the log for: contact.

← All articles