Eventually, yes. You would have to work up to that frequency from zero, to allow the rotor to accelerate. The Hall sensor, correctly positioned, would trigger a microcontroller to switch polarity. The devil is probably in the control detail. I imagine that, because of the rotary inertia, timing may depend on rpm, much as ignition timing needs advancing in a combustion engine. All that electronics in the Dyson appliance must be doing something more though (unless it's just there for obfuscation ). Perhaps it's for regenerative braking and battery charging?

If my theory is correct about the angular offset of the rotor's rest position, I think applying say ~100mA DC in one polarity would be enough to cause the rotor to move just a few degrees in one direction; ~100mA DC in the other polarity (once the rotor had returned to its rest position) would move the rotor through nearly 180 degrees in the opposite direction.

Forget about regenerative braking, guys!

It require minimum 6.3A to pass from one position to the other, otherwise it is moving, but not able to complete the travel.
When connected to the current generator, I can see 0.52/0.63 V on the pins (at least these values was what appear in each channel of my power supply).

I am lucky that my power supply is capable of MAX 3.2A per channel.. Just at the limit to make the test using parallel mode.

With this milestone, we are already the first easily accessible resource that describes how to make it move!
As soon as I get equipped we will make it run!

Thanks everyone.
Ciao

PS: I forget to answer one point.
With 100mA, nothing happen; it's a hungry monster, I don't remember precisely (I didn't write it down) but it seems to me that it's already starting to move around 500-600mA - certainly with 1A.

PS2: now that we redefine it as BLDC, I found this that seems to me very similar looking, at https://etheses.whiterose.ac.uk/21881/1/719808.pdf, page 81.

 

In response to that question posed in post#60:
What else the controls would be doing???Aside from driving the motor and controlling the battery charger function and providing some speed control because not all vacuum cleaner tasks are the same. There may be a second motor driving brushes for different surfaces. There may also be battery state of charge monitoring. And I believe there is also a display, not sure what it was displaying.

 

With 100mA, nothing happen; it's a hungry monster, I don't remember precisely (I didn't write it down) but it seems to me that it's already starting to move around 500-600mA - certainly with 1A.

100mA was only a guess. The current has to provide enough torque to overcome the rotor magnet's reluctance (pun intended) to 'let go' of the stator's poles to move from the rest position.
That single-phase PM motor in the post #61 link certainly seems very similar from a mechanical perspective; even the stator pole asymmetry is there.
Good luck with your further experiments.

Edit:
From that linked document it seems the present Dyson motor would be categorised as an FRPM motor, i.e. "A single-phase flux reversal permanent magnet motor".

Edit2:
Figures 2.15, 2.16, 2.26 and 2.27 in that linked document are particularly interesting and give good coil-driving info.

 

100mA was only a guess. The current has to provide enough torque to overcome the rotor magnet's reluctance (pun intended) to 'let go' of the stator's poles to move from the rest position.
That single-phase PM motor in the post #61 link certainly seems very similar from a mechanical perspective; even the stator pole asymmetry is there.
Good luck with your further experiments.

Edit:
From that linked document it seems the present Dyson motor would be categorised as an FRPM motor, i.e. "A single-phase flux reversal permanent magnet motor".

Edit2:
Figures 2.15, 2.16, 2.26 and 2.27 in that linked document are particularly interesting and give good coil-driving info.

Thanks for your help.
Ciao

 

If it is 200W at 22.5V that is 8.9A if the current was continuous. But it is pulsed, so it could be 10 times that in operation.

 

With a 22V battery and the coil combo having a DC resistance of ~ 0.1 Ohms, clearly current limitation will be necessary, particularly at low speeds as the coil inductance is also low (but not specified). The linked document goes into detail about stator and rotor eddy current losses and, if I undersatand it correctly, says that these are better minimised by controlling the coil voltage to limit current indirectly (via the coil inductance), rather than by controlling current directly. So PWM of the voltage applied to the coils would need to be combined with the coil voltage reversal.
I visualise the resulting effective coil voltage being something like this as the rotor accelerates from rest :-

 

Hi all,
I have received this evening this H Bridge from Amazon.
Specifications tells: power supply range 5V-15V, 200 kHz max freq., 50 A max current (will it be true?)


Anyway, I took an SN74LS04N, put it on a breadboard and set it to invert the function a generator signal (square wave 0V-5V 50% duty cycle)



Here a test of the inversion at 50 kHz:


Then, I connect the H-Bridge to everything:




And I have tested:

Thanks to everyone who contributed!
AAC is now the first site where you can find complete instructions for running this engine!

Notes:
In the video you can see the motor running at 72 Hz.
I have started at 10 Hz increasing progressively - not all frequency works well - but for example the motor seems to be in sync at 42 Hz, then at 52 Hz, 62 Hz, 72 Hz... but after this value the motor vibrate and don't run.
If I try 1 kHz or 2 kHz it make "beeeep" like a speaker.
I think probably I need more current - my power supply with 6.4 A is at the limit and I ear the cooling fan starting.
I don't want to destroy it...
When I will have the possibility to a more powerful power supply, I will resume the experiment and eventually update here.

 

Good that you have it working!
What were the applied coil voltages and currents at those different frequencies? Those would give a clue as to the coil inductance, which could be useful for you (or others) to know.

 

I don't understand why, but if I put the oscilloscope in parallel on the coil stops.
I don't even have to plug in both probe connectors - just plug in the negative and the engine stops.

But if I measure across the coil connector with the multimeter while it was running at 72 Hz i get:



and on the screen of power supply I read this:


On the power supply I have setup for 6V, but you see, it switch in Constant Current mode and measure on the H bridge a much lower voltage - I am not able to understand exactly what this mean.

The same methodology samples for 44 Hz (today 42 Hz does not work... boh?!?) provide:




I will be soon equipped with an LCR meter.
When I will receive it, I will try to measure coil inductance.

Ciao
Francesco

 

I can't explain why the oscilloscope stops the motor working, unless it is somehow shorting the power supply output.**
If the pulse duty cycle is 50% then the average voltage across the coil will be 0V. That's what your DMM appears to be showing.
Since you are switching an inductive load, current and voltage will be 90° out of phase. Perhaps that accounts for unexpected readings on the power supply display?

Edit:
** See post #73.

 

Unfortunately I don't have the skills to argue with you at this level - I would say some nonsense.

I have applied the oscilloscope on the coil, in parallel.
Oscilloscope and power supply are connected to two different 220 V AC wall plugs - I hope there is no relationship between the two.

 

To observe power waveforms on a scope you can capacitively couple both the signal side and the common side and if the problem is in the common connection that should be a solution.

 

I have applied the oscilloscope on the coil, in parallel.

If the ground clip of the scope probe is connected to one of the motor terminals that would explain why the motor doesn't run and why you are getting unexpected displays. The ground clip would be shorting part of the H-bridge, so there would be no flux reversal. Current through the shorted part would be limited by the power supply, so the H-bridge should survive.
Monitoring the voltage (relative to circuit ground) on just one of the motor terminals should work, as should MrBill2's suggestion.
Motor vibrations and failure to run beyond a certain rpm will occur if the flux reversal isn't synchronised with the rotor angular position.

 

Are the readings on your power supply Average or RMS?

 

Hello all,

To observe power waveforms on a scope you can capacitively couple both the signal side and the common side and if the problem is in the common connection that should be a solution.

Monitoring the voltage (relative to circuit ground) on just one of the motor terminals should work, as should MrBill2's suggestion.

Could you please suggest me the correct type/capacity value for capacitors and a diagram for the connections to do?
Sorry but I am unable to understand it by myself.

If the ground clip of the scope probe is connected to one of the motor terminals that would explain why the motor doesn't run and why you are getting unexpected displays. The ground clip would be shorting part of the H-bridge, so there would be no flux reversal. Current through the shorted part would be limited by the power supply, so the H-bridge should survive.

But... just connecting the negative of the probe? Only one wire connection?
This mean there is to consider some "connection" between the oscilloscope and the power supply due to the wall plugs?

Motor vibrations and failure to run beyond a certain rpm will occur if the flux reversal isn't synchronised with the rotor angular position.

This seems to me a suggestion to use an hall sensor.
But in your opinion what can be an approach for do this correctly? I have no idea of what I can expect from the hall sensor, to read it and evaluate it.
And also supposing to get it, how to modulate the signal in order to obtain an improvement... i mean... varying the duty cicle? varying the frequency?
If I look to the architecture of the motor seems to me that original hall sensor can be something used to understand the initial position of the shaft and start with the correct polarity, but I can't understand another usage.

More about the fact that motor does not run > 72 Hz... it is not normal that a coil will increase impedance in response to an increase of frequency?
In my basic understanding a capacitor block DC and tend to allow increasing frequencies accordingly with its value - if an inductor (like also a motor coil, I think) have an opposite behaviour, this must mean that will allow DC and tend to block increasing frequencies... and so major current is needed. Am I wrong on this simple tinking?
In our first analysis of the power, we imagine values around 10A... I am able to provide only 6.4 A max. I think this should be a reason.

Are the readings on your power supply Average or RMS?

Ehm... I do not know how to answer you.
My power supply is a Siglent SPD3303X-E, and seems to me manual and datasheed do not tell it.
I think only info on display feedback reading are these:



But maybe this is not the correct parameter... if you want to take a look, maybe you can get some useful info:
Service Manual Data Sheet

Ciao & thanks

 

But... just connecting the negative of the probe? Only one wire connection?

Yes: this demonstrates the problem

It assumes your power supply shares a common ground with the scope (neither is floating). So if the scope probe negative clip is not capacitively coupled to the circuit it should be on the circuit ground, not the motor terminal.

I have no idea of what I can expect from the hall sensor

They come in various flavours: some analogue and some digital, with/without ouput pull-ups. I suggest you look at a few datasheets to get a feel for how they behave.

If I look to the architecture of the motor seems to me that original hall sensor can be something used to understand the initial position of the shaft and start with the correct polarity, but I can't understand another usage.

Each time the Hall sensor detects a passing magnet pole it needs to trigger a reversal of the stator flux.

Could you please suggest me the correct type/capacity value for capacitors and a diagram for the connections to do?

A 1uF non-polarised cap in series with the scope probe negative should do the job.

 

A Hall Effect current sensor is far more complex than what is required to see the voltage waveform. a simple capacitor, 0.01mFd, 400 volts rating, between one side of the drive coil and the scope connection will be adequate. The scope "ground" connection would be to the power supply negative terminal, using a capacitor of 0.1 mFd and 400 volt rating,

And I am amazed that an engineer of any sort of electronics, including computers, would require a drawing of the circuit suggested. Unless the word "Engineer" has a totally different meaning these days. Or perhaps................

 

In my basic understanding a capacitor block DC and tend to allow increasing frequencies accordingly with its value - if an inductor (like also a motor coil, I think) have an opposite behaviour, this must mean that will allow DC and tend to block increasing frequencies... and so major current is needed. Am I wrong on this simple tinking?

You understand correctly. So Dyson would likely drive the motor with an increasing voltage (achieved using PWM) and Hall-triggered increasing frequency as the speed ramps up, as shown in post #66.

 

A Hall Effect current sensor is far more complex than what is required to see the voltage waveform. a simple capacitor, 0.01mFd, 400 volts rating, between one side of the drive coil and the scope connection will be adequate. The scope "ground" connection would be to the power supply negative terminal, using a capacitor of 0.1 mFd and 400 volt rating,

I think you didn't understand why I mentioned the hall sensor.
It was referring to the motor control process, not as an alternative method to reading the voltage across the coil.
Thanks for the suggestion about the capacitor value: I will take it into consideration after evaluating any other answers.

And I am amazed that an engineer of any sort of electronics, including computers, would require a drawing of the circuit suggested. Unless the word "Engineer" has a totally different meaning these days. Or perhaps................

What a shot! I'm so happy that you consider me an engineer!
Quite a step forward, considering you didn't even think I'd be able to get the engine running.
But unfortunately I'm not an engineer... I'm just a curious person, who tries to spend the limited free time he has having fun and learning something, with the humility to ask when he doesn't know or understand something and who always thanks when someone helps him.

 

It assumes your power supply shares a common ground with the scope (neither is floating). So if the scope probe negative clip is not capacitively coupled to the circuit it should be on the circuit ground, not the motor terminal.

Thank you very much. I never thought that two separated instruments could influence each other in this way.

They come in various flavours: some analogue and some digital, with/without ouput pull-ups. I suggest you look at a few datasheets to get a feel for how they behave.

Maybe I'll try to recover the one on the burned controller and then try to see if I figure out how to use it.

Each time the Hall sensor detects a passing magnet pole it needs to trigger a reversal of the stator flux.

I understand and in truth I had thought about it too, but I had discarded the idea because I ask myself: how can I control the the speed in that way?
In this way, conceptually, it doesn't auto-tune to a specific frequency value on its own, making the presence of the function generator useless?