It started as one of those “nothing is wrong… until it is” moments.
A line had been running fine for months. The motor was a standard 3-phase asynchronous, driven by a VFD—nothing exotic: steady production, predictable load, no drama. Then one afternoon, the plant took a brief power dip. The VFD rebooted, the control system did what it was supposed to do, and the motor went right back to work.
Ten minutes later, the overload warning lit up.
The maintenance team was puzzled. The motor hadn’t changed. The process hadn’t changed. The drive hadn’t been touched. So why did the VFD suddenly think the motor was in trouble?
The answer was sitting in a single parameter that most of us overlooked during commissioning:
Motor temperature at power up.
And behind that innocent label were three assumptions—three different narratives the drive can tell itself about the motor’s thermal state:
Cold start. Warm start. Retained temperature start.
The drive’s quiet little fiction: the thermal model
In many VFD installations, the motor is protected by a thermal model in the drive—often an I²t-based estimator. The VFD doesn’t measure winding temperature directly (unless you’ve integrated PTC/NTC/RTD feedback). Instead, it estimates the temperature rise from:
Electrical losses (primarily proportional to current²),
and an assumed cooling capability of the motor.
This model is usually accurate enough to protect the motor across normal operating conditions—but only if two things are right:
your motor data is correct (nameplate current, base speed/frequency, etc.), and
the VFD’s assumptions about cooling type and speed range match reality.
The model needs a starting point after a power event. That’s exactly what “motor temperature at power up” defines.
Three ways to start a story
1) Cold Start
The drive’s assumption:
“On power-up, the motor is at ambient temperature.”
In plain terms, a cold start resets the thermal memory. It’s telling the thermal model:
“We’re beginning from zero stored heat.”
When this storyline is true
The motor has been off long enough to cool down.
You have procedural confidence in the cool-down time.
You’re commissioning or performing baseline tests.
When it becomes a dangerous fairy tale
If the motor was actually hot—say after a short power dip—a cold start can overestimate safe thermal margin. The drive may allow overload that the real motor can’t safely tolerate.
2) Warm Start
The drive’s assumption:
“On power-up, the motor is partially warm.”
This is usually a fixed intermediate value (vendor-dependent). It’s a compromise between a hard reset and full memory.
Why do engineers like it
Warm start is the sensible default when:
you don’t know the motor’s real thermal state, and
you want a conservative approach without being overly restrictive.
3) Retained Temperature Value Start
The drive’s assumption:
“I remember where we left off.”
This is the most physically honest option. The VFD continues from the last estimated thermal state.
Where this shines
Short power interruptions are common.
The process is continuous and thermally demanding.
The asset is remote or unattended.
You rely heavily on the drive’s model (no winding sensors).
But retention is only as good as the model behind it. If the motor configuration or cooling assumptions are wrong, you’re just retaining a confident mistake.
The twist: cooling type changes everything
At this point in the investigation, the team tried toggling between cold, warm, and retained starts. The nuisance trips improved—but the bigger clue was still hiding in the motor’s cooling method.
Because the thermal model doesn’t just count current. It also assumes how the motor gets rid of heat.
Enter the two characters that matter most in VFD applications:
Self-cooled motor (auto-ventilé)
Forced-cooled motor (ventilation forcée)
Self-cooled motors: the fan that slows down with you
A self-cooled motor uses a fan attached to the shaft.
At line speed, cooling is fine.
At low speed, cooling falls dramatically.
That creates a classic VFD trap:
The process may demand near-constant torque.
The motor may draw significant current.
But the fan isn’t moving enough air.
So the motor heats faster even if the current looks “normal.”
This is why drives often apply:
derating at low frequency, or
thermal curves that tighten allowable current as speed falls.
Translation:
A self-cooled motor that lives at low RPM is a thermal risk unless your drive settings reflect that reality.
Forced-cooled motors: cooling independent of speed
Now meet the forced-cooled motor.
Here, the fan is driven separately from the motor shaft. The airflow is largely independent of RPM.
In VFD terms, this is gold.
It means:
better thermal stability at low speed,
more reliable continuous torque capability,
reduced need for aggressive low-speed derating—if configured correctly.
Where engineers get burned
This is the part of the story where the team realized the problem wasn’t just the “temperature at power up” parameter.
It was the combination of:
the starting assumption, and
the configured cooling type, and
the operating profile.
The most common failure modes look like this:
1) The drive thinks forced cooling exists… but the motor is self-cooled
This is the worst-case risk.
The model may allow current at low speed that is thermally unsafe.
2) The drive thinks the motor is self-cooled… but it’s actually forced-cooled
This usually won’t damage anything,
but you may get unnecessary thermal limiting and nuisance trips.
3) Cold start in a plant with frequent power dips
Each brief outage “resets” the model to cold,
even though the real motor keeps accumulating heat.
The model becomes optimistic—dangerously so.
4) Incorrect nameplate data
Thermal models depend on the same fundamentals you use for protection and torque limits:
rated current,
base frequency,
speed,
service factor/insulation details where applicable.
Bad data = bad model, no matter which start mode you choose.
Choosing the right start mode by real-world pattern
Once the team framed the problem as “what story does the drive tell itself at reboot?” the selection became easy.
Constant torque at low speed
Think conveyors, extruders, and mixers.
Self-cooled motor:
Prefer retained temperature start and ensure low-speed derating settings are correct.
Consider adding PTC/RTD sensors if headroom is tight.Forced-cooled motor:
Retained or warm start both work well.
Make sure the drive explicitly knows it’s forced-cooled.
Variable torque loads
Fans and pumps typically stress the motor less at low speed.
Warm start is often a safe, simple default.
Retained start still improves accuracy if dips are common.
High cycling, frequent stops
Indexing systems, packaging, stop-start processes.
Retained temperature start is strongly preferred.
A cold start can hide cumulative heating across rapid cycles.
The short version you can teach in two minutes
When the team briefed operations, they used a simple set of one-liners:
Cold start: assume the motor is at ambient; reset thermal memory.
Warm start: assume partial heat; conservative default.
Retained start: continue from last estimated temperature.
Self-cooled: cooling drops with speed; low-speed torque is thermally risky.
Forced-cooled: cooling independent of speed; improved low-speed capability.
The confusion vanished almost instantly.
How the story ended
The root cause turned out to be a mismatch:
The motor was operating for long periods at low speed under constant torque.
It was self-cooled.
The drive configuration effectively behaved like it had more low-speed cooling margin than reality justified.
On a brief power dip, cold start reset the model, granting fresh overload headroom to an already warm motor.
They changed two things:
configured the correct cooling assumption and low-speed derating behavior, and
switched to retained temperature start.
The nuisance trips stopped.
The thermal protection became consistent.
And the motor stopped living on borrowed time.
The takeaway
The parameter “motor temperature at power up” isn’t a control gimmick.
It’s a statement about protection credibility after a reset.
Your choice should reflect:
confidence in the motor’s real thermal state at restart, and
how cooling behaves across your speed range.
If you get those two right, the VFD’s thermal model becomes a reliable guardian.
If you get them wrong, it becomes either an alarmist… or worse, an optimist.
