TON, TOF and TONR: PLC Timer Types Explained With Real Machine Examples
The three timers, in one table
| Timer | Starts when | Output goes true | What happens when the input drops |
|---|---|---|---|
| TON (on-delay) | Input goes true | After the preset time has elapsed while the input stayed true | Output off, elapsed time resets to 0 |
| TOF (off-delay) | Input goes false | Immediately when the input goes true | Output stays true for the preset, then drops |
| TONR (retentive on-delay) | Input goes true | After total accumulated true-time reaches preset | Output holds, elapsed time is kept; needs an explicit reset |
Everything else about timers follows from that last column.
TON: the one you will use most
An on-delay timer says "this must be true continuously for N seconds before I act".
Real use: conveyor start warning. Press start, a siren sounds, and the motor only runs after five seconds. TON with a 5 s preset, input from the start latch, output enabling the motor. If somebody hits stop during those five seconds, the timer resets and the count starts again from zero next time. That reset-on-drop behaviour is exactly what you want for a safety warning.
Real use: filter blocked alarm. Differential pressure above setpoint for 30 continuous seconds means a blocked filter. Without the timer, every pressure spike raises an alarm. With a TON, transients are ignored and only a sustained condition alarms. This is how you stop an operator drowning in nuisance alarms.
TOF: the run-on timer
An off-delay timer says "keep this on for N seconds after the trigger goes away".
Real use: extractor fan run-on. The welding booth fan should keep running for two minutes after the process stops, to clear fumes. TOF with a 120 s preset. The output is true the instant the process starts and stays true for two minutes after it ends.
Real use: hydraulic cooling. Pump stops, cooling fan continues until the oil temperature settles. Same shape.
Engineers new to ladder often try to build this with a TON and get into a mess with latches. If the requirement contains the words "after it stops", reach for TOF.
TONR: accumulated time, and be careful
A retentive timer keeps its accumulated value when the input drops. It only resets when you explicitly reset it.
Real use: running-hours counter for maintenance. A motor runs in bursts all day. TONR accumulates total run time across all those bursts, and at 500 hours you raise a service alarm. Maintenance resets it after the service.
Where it goes wrong: people use TONR for machine sequences because it "remembers", then a power cycle or an emergency stop leaves the machine believing it is halfway through a step it never finished. A sequence should generally restart from a known safe state, not resume from wherever it was interrupted. Use TON for sequences and TONR only when accumulation is genuinely the requirement.
Preset, elapsed and the two ways to read them
Every timer gives you the preset (PT, what you set) and the elapsed value (ET, how far it has got). Two things worth knowing:
- ET keeps counting past PT on some platforms. Do not rely on ET stopping at the preset unless you have checked on your hardware.
- Time types are not integers. In TIA Portal a preset is a TIME value, written
T#5sorT#2m30s, not the number 5. Passing an integer where a TIME is expected is a compile error that confuses people for far longer than it should.
Timers across the platforms
The IEC 61131-3 standard defines TON, TOF and TP (pulse). Vendors then add their own:
| Platform | On-delay | Off-delay | Retentive |
|---|---|---|---|
| Siemens S7-1200 / 1500 (TIA Portal) | TON | TOF | TONR |
| Allen-Bradley (Studio 5000) | TON | TOF | RTO |
| Mitsubishi (GX Works) | OUT T | via logic | OUT ST (retentive) |
| Delta / Omron / Schneider | TON | TOF | varies |
Rockwell calls the retentive one RTO and its reset RES, which trips people moving from Siemens. The behaviour is identical.
The pulse timer nobody teaches
TP, the pulse timer, gives you a fixed-length output pulse regardless of how long the input stays true. It is perfect for triggering a solenoid a fixed duration when a sensor sees a part, no matter how long the part sits in front of the sensor. It is in the standard, it is on every platform, and it is missing from most training courses. Knowing it exists will save you building a latch and a TON to do one instruction's job.
The interview question
"What is the difference between TON and TONR?" is asked constantly. A complete answer is not the definition, it is the consequence: TON resets its elapsed value when the input drops, TONR keeps it, so TON is for conditions that must be continuous and TONR is for totalising run time. Then give one example of each. That structure, definition plus consequence plus example, is what an interviewer is actually listening for.
More interview preparation: our free interview question prep tool covers the PLC fundamentals section with what a good answer contains for each.
Practise on real hardware
Timers are easy to read about and easy to get wrong on a live machine, because the failure mode is usually a sequence that half-works. On the Automation Engineer Program you build and commission sequences on real panels rather than a simulator, which is where the difference between TON and TONR stops being academic.
Related: Seal-in and latching circuits, PLC scan cycle explained.
