Stall detection - How quickly do mains powered induction motors startup?

[e100]

As in approximately how many milliseconds does it take for the first revolution to occur when starting from rest, assuming a minimal load and a 1kw single phase motor.
A Google search didn't produce any concrete numbers, just various graphs showing the high current draw at startup without mentioning the angular velocity with respect to time.

Visually I'm guessing that the first revolution takes something like 200 milliseconds. Does that sound about right?

[T3sl4co1l]

What type?  Average ~1750 RPM induction motor with a starter winding and cap?  Yeah, probably that.  The first revolution probably even less, maybe a couple (line) cycles?

Note that it scales with size, probably linearly?  Because more power is larger rotor is quadratically more inertia.  Or something like that.  You gave a scale reference so this isn't exactly relevant here, but may be useful to keep in mind.

The startup current is due to slip, which drops off as rotor speed approaches synchronous speed.  So, the fact that it's decaying at some point, is proof that it's already moving at some speed, which should be calculable.  Though I'm not sure offhand the slip to current transfer curve, and whether that includes starter current (which I think stays pretty high until the centrifugal switch opens).  But anyway if you're looking for actual stall (or at least, not nearly enough RPM to open the starter switch), that should be fine too, it'll just sit there cooking at LRA.  Indeed, ye olde fashioned "motor starter" is just that, a slow-blow resettable breaker (usually with some resistance to limit inrush, which is bypassed once running).

Tim

[filssavi]

There is no fixed answer, it depends on many factors:

- machine size: the bigger the machine, the more inertia it has the slower it will start up
- machine speed-torque characteristic: the more torque at low speed the faster it will accelerate
- upstream power delivery: while there is no converter to limit the current, there might be situations where your startup is limited by resistance of an inadequately sized power network (long cables that might be undersized for the full startup current of the machine)
- load moment of inertia: if the load does not require large amounts of torque at low speed, or because it can be turned off (one example would be a variable pitch fan, a pump for a unloaded hydraulic system, etc) when running at steady state, it still might have a large moment of inertia, slowing down startup

[Siwastaja]

Simple answer: It's all a crappy compromise.

Surge current is massive, there is quite some torque but it's nowhere near what the current would suggest, so the iron is basically saturating and efficiency is maybe some 10%.

"Stall detection" is usually based on a motor protector switch which is basically a fancy circuit breaker with adjustable tripping current. This is because if the motor spins up properly, the ACIM slip decreases to maybe not optimal but better range, improving the efficiency and providing the torque without such massive overcurrent.

How long does this take? It depends on motor, load, it can be 200ms, it can be many seconds, even tens of seconds for large masses which store a lot of energy in inertia.

Such motors are often a compromise in design so that giving up efficiency at normal operating RPM, they widen the "sweet spot" of slip making the totally crappy efficiency at low RPM a tiny bit less crappy but still quite horrible. The efficiency starts plummeting somewhere below 50% to 80% of rated RPM so you need to bring it up to quite some serious RPM before you can say it's "up to speed" and out of the high-current region.

If at all possible for anything larger than trivially small motors (fans, small pumps...), always consider the possibility of a true 3-phase machine driven by a VFD. Then this complete set of rules totally change, as the combination of VFD+motor is able to avoid running at the totally non-optimal slip, always provide maximum torque with current related to that torque, no more, and motor efficiency always around 80-95%, of course depending on a motor.

[Doctorandus_P]

The time for the first revolution is unimportant.
It may also vary wildly on the exact geometry with which the motor is built.
If you stay within the parameters you've given, (Single phase, 1kW) there still are a bunch of other factors that are important:
1. How much poles does the motor have 1 and two pole motors (3000rpm and 1500 rpm for 50Hz, minus some slip)
2. 120V or 230V?
3. Start capacitor with centrifugal switch (I think these have higher starting torque) or Run Capacitor?
4. Geometry of the motor, especially the length to diameter ratio.

When the motor is starting, the inertia of the total system is an important factor, you specified "no load", but that is not realistic. I once was fiddling with a circular saw (3-phase 5.5kW 400Vac) The goal was to minimise the time for the Star / Triangle starting circuit, and I discovered that it can change between about 2 and 6 seconds depending on the size of the sawblade.

So if this is important, then just measure it for the motor you are using, and make it adjustable if you want to build some universal circuit.
But it all probably is either overkill or just not very useful.
Even if the motor axle is fully blocked, it can survive a few seconds of startup current without damage.

The amount of slip, and therefore RPM also changes with the load of the motor. You could measure RPM during runtime, and calculate motor load from that, but the standard way for determining motor overload is with a switch with bimetallic elements, which is good enough. (Frequency inverters have similar function built into their software, and complexity of that can vary of course).

[rstofer]

Since you are talking about a very small single phase motor, we'll leave out wye/delta or resistance starters.  Motors that use these types of starters may take 10s of seconds to get up to speed.  But even then, 'something' starts to move fairly quickly.

If you have a specific application in mind, come up with a way to measure angular velocity versus time.  If there is a flat surface perpendicular to the shaft, paint half black and the other half white.  Use a retroreflective IR sensor to sense the rotation.  Grab up an Arduino and figure out how to grab RPM from the sensor pulses.

Here are several approaches: