Sudden reversal in a simple 3-phase motor's direction

[darth boozer]

Watching an old Royal Institution Christmas Lecture, given by Prof. Eric Laithwaite in 1974, there was a demonstration of a simple 3-phase induction motor. As the applied voltage was increased, the direction of spin suddenly reversed at about 203 volts. I have never found an explanation for this. Can anyone help?

http://www.rigb.org/christmas-lectu...h-the-looking-glass/tweedledum-and-tweedledee

from 21 minutes 21 seconds

 

[Guineafowl]

Watching an old Royal Institution Christmas Lecture, given by Prof. Eric Laithwaite in 1974, there was a demonstration of a simple 3-phase induction motor. As the applied voltage was increased, the direction of spin suddenly reversed at about 203 volts. I have never found an explanation for this. Can anyone help?

http://www.rigb.org/christmas-lectu...h-the-looking-glass/tweedledum-and-tweedledee

from 21 minutes 21 seconds

“It would take me two lectures to explain that to your satisfaction.”
With a bit of effort, you can explain anything complex in simple terms. That is the goal of science and the job of a popular science lecturer.

“You’ll forgive me if I [leave this effect entirely unexplained and] press on.”
I will not.

 

[jrmichler]

It's called aliasing. The ball was increasing in speed as the voltage increased, it just looked like it reversed direction. Aliasing was common in old movies where the stagecoach was moving forward while the wheels were apparently rotating backward. Some links:
http://www.physlink.com/education/askexperts/ae490.cfm
https://web.njit.edu/~joelsd/Fundamentals/coursework/BME310computingcw6.pdf

It's a common problem in movies and videos, where what appears to be smooth motion is really a series of still images projected at (usually) 30 frames per second. That room may have had fluorescent lights, which also flicker.

 

[Dr.D]

So if it is due to aliasing (and I don't doubt this), there is no true reversal in rotational direction, only an apparent reversal. Not quite the same thing!

 

[Windadct]

I thought Aliasing as well- but in this video I do believe the issue is the ball (rotor) physically reverses.

I went looking for info on this - the slip does decrease as V increases, and the curves all stop at 0 Slip - basically this would be Negative Slip - and beyond this operating point? -- I can only assume there is not much about this phenomena since we are talking about V applied that is much above the design voltage, and in any real or practical application we would already be talking about zero torque ( what is the point in that) and/or motor damage.

I am curious if the typical motor equivalent model will show this in simulation.

 

[Dr.D]

Has Old Jim (@jim hardy ) weighed in on this?

 

[jim hardy]

Has Old Jim (@jim hardy ) weighed in on this?

I'm puzzled too. My first reaction was aliasing as mentioned by @jrmichler but i watched it enough times to convince myself "it sure looks real " .

It's a spherical looking rotor and i'm looking for a practical treatment of spherical rotor machines.
This one would take several lectures...
http://www.compumag.org/jsite/images/stories/TEAM/problem34.pdf

I haven't given up yet.

old jim

 

[darth boozer]

My first thought was aliasing, but I doubt that it would be visible to the live audience at the time and the spin direction is also clearly visible during the final slowing of the rotor. Bearing in mind that the lectures were for young people, including pre-high school students, to be able to explain it in just two lectures would surely mean that the reason behind the effect cannot be too complex.

 

[jrmichler]

A single phase motor runs equally well in either direction, but he clearly states that it is a three phase induction motor. I know that a three phase motor will run on at reduced power when it loses one phase. When I look at wye and delta connections, it appears to me that losing one phase leaves a three phase motor running single phase? If so, it has zero starting torque. Is there some way that he could be applying full power to two phases, then varying the third phase that would make the motor reverse?

 

[jim hardy]

What was curious to me was it didn't reverse again when he returned below 203 volts.

Then i thought - "If that's just a hollow metal toilet tank float it's got no iron in it. How can those windings support 200 volts with basically an air core ? "

Then i looked at the machine - it looks like a donut surrounding the float .I can't tell which way the turns are oriented on it.

Does flux actually go horizontally through the hole in the donut, between same colored windings, as in a real motor ?
Or is this thing a toroid with six windings on it ?

If it's a toroid then flux stays mostly in the horizontal plane an just traverses the toroid?

I think the first step is to discard the notion it's a conventional motor.

old jim

 

[rokytnji]

From my practical experience. With conventional 3 phase motors. Not exotic stuff. Switching legs on phases is the only way to run a 3 phase motor in reverse. Single phase or lose one leg. Trips the fuse or burns up the windings. Going full speed and reversing phase feed. Probably would get ugly.

 

[anorlunda]

Where's James Randi when we need him?

 

[Mandy D]

I have seen this phenomenon when running a 37 kW squirrel cage motor from a 2.2 kW VSD! We needed to run up the motor to full speed but could not DOL start it from the supply we had, so we used the only VSD we had to hand and connected the motor in star (690V) and ran it from this 230V VSD, set a 3 minute acceleration ramp and got it up to speed with no load, obviously.

However, sometimes it would start in reverse and then just flip at about 2 Hz and run forwards for the rest of the acceleration ramp. We never bothered figuring it out as the tests we needed to complete were successful. I think it might be that there are several sequences going on simultaneously and that the motor picked up on a reverse sequence to start with and once the field became strong enough and dominated the flux it followed the proper sequence.

rockytnji is correct that reversing a motor running at full speed without stopping it first is a bad idea. We had this with a motor driving a synchronous generator that was synchronised to the grid. Unfortunately, on one occasion the extension lead I was using had been swapped for another one which had the phases crossed over. The synchroscope indicated that we were good to go, I hit the switch and the machine leapt into the air on my bench and stopped with an almighty bang. I've never seen a motor stop as fast as this one did! Fortunately, it was only a 7.5 kW machine, but it killed the d.c. PSU I was using to excite the generator! At first I could not understand what had happened, but it slowly dawned on me that phase reversal had to be the culprit. Yep, it certainly can get ugly. I am just glad it was such a small machine!

 

[Mandy D]

What was curious to me was it didn't reverse again when he returned below 203 volts.

Then i thought - "If that's just a hollow metal toilet tank float it's got no iron in it. How can those windings support 200 volts with basically an air core ? "

Then i looked at the machine - it looks like a donut surrounding the float .I can't tell which way the turns are oriented on it.

View attachment 226686
Does flux actually go horizontally through the hole in the donut, between same colored windings, as in a real motor ?
Or is this thing a toroid with six windings on it ?

If it's a toroid then flux stays mostly in the horizontal plane an just traverses the toroid?

I think the first step is to discard the notion it's a conventional motor.

old jim

Jim, I'm not surprised that it failed to reverse on the way back down. The motor had started backwards (see my earlier post), then it flipped, so was now going in the same direction as the phase sequence. This motor works because the windings have a lot of impedance (because it is small and has many turns) and are wound on an iron core. The rotor is basically anything that you can induce a current into (hence, induction motor) and does not technically need iron. The iron just concentrates the magnetic field and gives superior performance and efficiency, but is not essential. I cannot comment on the flux pattern, because I cannot see how the stator is wound.

I would not discard the idea that it is a conventional motor, because as I stated in my earlier post, this effect can happen in a regular cage motor.

 

[jim hardy]

I did notice a speed difference after it reversed. Fifth harmonic is reverse sequence, is it not ?

 

[Mandy D]

I did notice a speed difference after it reversed. Fifth harmonic is reverse sequence, is it not ?

Jim, yes fifth harmonic is definitely a negative (reverse) sequence. If the motor follows this with a lot of slip it will be running backwards and look like all is well, until you reach that magical point where the fundamental can pull it out and force forward rotation. You will then see the motor reverse, exactly as in the video and then behave normally under further acceleration, even during the run down. I think we can say this mystery is solved!

 

[jim hardy]

Partially solved , at the very least. I'd think raising voltage would increase harmonics.
But voltage induced in the rotor is dΦ/dt so 5th harmonic flux that reaches it gets a fivefold share in making torque ?..

 

[jim hardy]

If the motor follows this with a lot of slip it will be running backwards and look like all is well, until you reach that magical point where the fundamental can pull it out and force forward rotation. You will then see the motor reverse, exactly as in the video and then behave normally under further acceleration, even during the run down.

That's what i noticed, too.

Given who that guy is, inventor of the linear motor, it's no surprise he has uncanny skills with magnetic fields.
His "core" still looks to me like a toroid with three sets of windings. Look at their orientation , i think they encircle the core just as in a toroidal inductor.

A conventional motor would have its windings oriented to push flux diametrical through the middle of the torus not circumferentially around it.
That is the basis of my remark "not a conventional motor".
The video isn't quite clear enough to say for certain how they're oriented. So i could be all wet on this one...

We're used to thinking of a toroid as completely containing the flux.. So my mind immediately lept to the conclusion it was a conventional motor with diametric flux.
But if so, how does it support 200 volts with no iron in the rotor? Must have enough turns to support it with air core. I wish he'd shown current we'd know if he was saturating his core..
That the instantaneous current in the windings differs between the colored phases says the condition for total confinement of flux to a toroidal core is not met. So i think his "motor" is more a conducting sphere in a time varying field that's tangential to the windings of a leaky toroid, if you will. .


A search on "Conducting sphere in a magnetic field" produces scholarly articles . I didn't find one that i could apply, perhaps your Vector Calculus is fresher than mine.. But maybe that's why he says he can't explain it quickly ?

old jim

 

[anorlunda]

I hit the switch and the machine leapt into the air on my bench and stopped with an almighty bang.

Yikes! You're lucky nobody was hurt. The whole thing could have melted and flown apart. There are sync check relays to prevent mistakes, but those relays too can be wired incorrectly.

 

[jim hardy]

Wow 7.5kw sync'd out of sequence ? Lucky it didn't throw itself at you !

 

[anorlunda]

I don't have an answer. But three things occur to me:

1. A squirrel cage rotor has only one rotational degree of freedom. A sphere has two more rotational degrees of freedom.
2. We can only see the surface of the sphere. We don't know what's inside. The surface could be a non-conducting skin driven by (don't know what).
3. If we listened to the whole video, there may be an explanation. But I'm not willing to invest that much time.

 

[Guineafowl]

We can only see the surface of the sphere. We don't know what's inside. The surface could be a non-conducting skin driven by (don't know what).

It’s a float from a ballcock (stop giggling at the back). A thin, hollow brass sphere. He says it at the end of the demonstration.

 

[Baluncore]

Let me make some guesses and assumptions.
1. It is not a trick.

2. The supply frequency remains constant. The phase connections are not changed.

3. The motor reverses direction once when the voltage is increased through a critical value.

4. The rotor is a light weight, thin-walled, non-magnetic, metal sphere.

5. The spherical rotor will not saturate, but the surface currents may have a harmonic dependent phase delay.

6. The toroid has six windings which form a six pole field. The current in each coil creates a flux that is mismatched where two coils meet. The flux difference is released at the coil junctions as a magnetic pole. That leaked flux then travels out of the toroid, inducing a perpendicular current, (axial), in the surface of the sphere. The induced currents in the sphere are balanced so they form circumferential currents in the rotational axis regions away from the core. The induced currents sum to zero about the axis. By picking the phase, the step change in coil current, i, can be made to vary sinusoidally, with differential currents from +2i to -2i.

7. The toroid has a magnetic core. That could be an iron powder core made in 6 parts, one for each coil, each curved through 60°. They are then glued together, end to end, to make 360°. There may be something at the junction that aids the release of the magnetic field difference. An alternative core could be a bundled coil of one insulated iron wire, like a hedgehog transformer core, but much more orderly. That would give the phase windings something to magnetise, to reduce the current at higher voltages.

8. What will happen if the core saturates at higher voltages? Will odd harmonic distortion change to a reversed direction harmonic?

 

[anorlunda]

Let me make some guesses and assumptions.

Those sound pretty good.

But the bad part is that I don't think we'll ever get the definitive answer to know if our guesses are right or wrong.

 

[Baluncore]

But the bad part is that I don't think we'll ever get the definitive answer to know if our guesses are right or wrong.

Those points were the starting documentation for the crude simulation code I started work on, to sort out the phase needed in the coils. I am surprised by the number of different step patterns that can be generated with 3PH and reverse to six poles. When/if I discover the "biggest steps" sequence I will consider winding the coil. I can't trust a VFD, so the 3PH variac will come from the old power meter calibrator. The brass ball float turned up in a load of old plumbing scrap a couple of weeks ago. At the time I did not know why I needed to rescue it, maybe there is something in heart-strings-theory after all.
Too much to do in the time before the instant of the winter solstice here, Jun 21, 10:07 UTC.