brushless-dc-motormotorwinding
I'm looking to rewind a brushless motor to run at maximum efficiency for a given voltage and RPM.
let's say given the supply voltage is 12v, I want it optimised to run at a constant RPM of 4000 and drawing no more than 0.5 amps.
how do I calculate the wire size and number of turns?
Rushing.
More later maybe ...
Note that "stall torque" is often used to mean locked rotor 0 RPM torque BUT you use it in the sense "dropout torque at a given speed". That's entirely fine as long as you note that some references will mean the former and not the latter.
Criticism (kind / constructive) welcome.
Written at a rush and unchecked. Better is possible.
See writer "Toper925" comment here
He notes:
There really is no single equation that fits all states of a PMSM but this one works in general:
Where:
p is the number of pole pairs
λ is the amplitude of the flux induced by the PMs in the stator phase
Lq and Ld are the q and d axis inductances
R is the resistance in the stator winding
iq and id are the q and d axis currents
I'd need to read more on what he said to make total sense.
Stall torque is when torque is not sufficient to "pull in" the next rotor pole piece using the available magnetic field.
SO, I'd expect
More pole pairs better. I'd expect better than linear gain as distance halves with doubled pairs BUT magnetic force at worst falls as distance cubed (only a considerable number of magnetic pole diameters away so not in most sensible motors), comes closer to falling with distance squared as gap falls to near pole width and at best can only approach linear at close proximity. SO more ples should give less interpole distance so ... (but pole sizes are down so ...).
Torque = power per rev. If the power falls faster than RPM your margin is dropping until you reach the point of no pull in. At a quick glance I think that this is what this man here is alluding to about half way down below the graph. Leading to ...
If you have a power curve you also have a torque curve as the two are related by motor rpm. (Torque = k x Power / RPM). If you have a speed-power plot for you load you should be able to overlay this on torque curve and see where load torque is > generated torque. This will be better than real world (probably).
Lowest R should help as it allows greatest I but this is really a secondary effect for two motors with the same power at the same RPM.
Induced flux should play an immense part. I'd expect non saturating magnetic (eg steel) cores to provide superior results EXCEPT if you can get all gaps so small that field is well maintained by magnet. Rule of thumb is you can get about 0.5 Tesla at an airgap of 1/2 a magnet diameter using a top class NdFeB magnet. Say N52? N45 won't be too bad.
Note that the US process NdFeB magnets are cast but ground and sintered subsequently and are inferior in max possible flux to the Japanese versions. This should all be covered in the flux spec.
Based on the datasheet, it seems to me the /DEC pin needs to be high (in addition to /ACC being low) for the motor to start. Anywhere between 2.2V and 5V seems fine for /DEC to register as high. Leaving it open-circuit won't do that though.
The reason why it does nothing with no voltage to /DEC is that this chip it has internal pull-down. Possibly other chips of this kind have internal pull-up instead.
And don't apply 24V, you'll fry it. The top limit voltage is given by the VREG pin (which you could measure, but 5V seems safe).
Best Answer
Without more information it is impossible to say.
'Iron' loss (magnetic and friction) increases as speed increases, but 'copper' loss (winding resistance x current2) increases as more current is drawn. At idle (no-load, minimum current draw) efficiency is zero. At maximum power output the winding resistance absorbs about half the input power and efficiency is 50% or less. Peak efficiency occurs somewhere between those two points, when iron loss equals copper loss.
Magnetic losses are determined by the physical design and construction of the magnetic path (stator shape, core material, air gaps etc.) Winding resistance is determined by wire diameter and length which is dependent on the space available, but the length required to get a particular rpm is determined by motor dimensions, winding pattern, magnet placement and strength. A large motor needs less turns of thicker wire than a small motor of the same design, and a motor with strong magnets need less turns than one with weak magnets or a larger airgap (but stronger magnets increase iron loss, so...).
If you have a motor that runs on a different voltage than you want, you can rewind it to run on a different voltage by changing the number of turns in proportion, eg. if it does 4000rpm on 6V now then double the number of turns to get 4000rpm at 12V. But that takes up twice as much space so you must reduce the diameter by 1.4 (to get half the area per wire).
However, rewinding will probably not produce maximum efficiency unless you are very lucky. For a given motor, peak efficiency occurs at a particular rpm and power no matter what voltage it is wound for. For example if your motor currently does 4000rpm at 1A on 6V with 70% efficiency, if you rewind it to do 4000rpm at 0.5A on 12V it will still get 70% efficiency.
If you rewind for lower rpm it will have higher resistance and copper loss, whereas if you rewind for higher rpm the magnetic loss will increase. If it is too small then copper loss will dominate due to the smaller amount of copper, but if it is too large then iron loss will dominate because the larger core has higher magnetic loss. So when you rewind it for different rpm the efficiency could get better or worse, depending on the particular motor and where you are operating it on the efficiency curve.