Adding a Sin/Cos sensor to our motors?

At work I deal with many FOC and BLDC controllers (industial grade), which offer very reliable performance and control, at least up to a VESC, but a fraction of the cost.
Despite my “side enquiring” software and hardware developers, got no luck for use with sensorless motors, so far. So the idea: why don’t just add a cheap and simple SinCos sensor on the back of the motor? I know it’s another cable to route up to the mast, but a simple 4 x 0.25 signal wire could be used… it won’t be a big deal.
In the next weeks I’ll try to get one of the sensor we usually mount on my new TP 320 motor, wish it can work, would be nice to upgrade from RC toys or skateboards into more serious industrial grade “no bullshit” products

Nice idea, we could also have a real time view of the rpm.
Could we integrate this encoding device in the propeller fairing between motor and propeller. Are those devices waterproof ?
Wouldn’t it be cheaper to buy the (hall) sensored waterproof version of the motor (Maytech sells both versions of their 65151) and connect it to your new ESC?

Can it handle 100A continuous ?
How much cheaper are we talking about ?
Also, apart from the price, what would be the advantage ?

The controllers of which I’m thinking about for efoil (can’t disclose the brand yet but I will as soon as I asked to the Company) are rated 140A with no cooling, we watercooled this on an outboard motor, 70-80A for 2 ours did not even raised the temp above 35 C.
Advantages are the huge level of controls that they offer, any kind of protection, CanBus if needed and lots of additional input and outputs signals that you could utilize for safety. Being devices that equip forklifts, truck, industrial pumps etc, can’t fool around with fake numbers and poor quality like the most of the Chinese rc toys or clones.
Drawback but maybe solvable, when you build a lift truck or a chemical plant pump the sensorless option it;s not even considered, for 10euro you rather go safe with a sin cos or resolver sensor. It must be in the back of the motor, magnet on the shaft and a pcb on top.
Will post photo as I get one disassembled

Thanks a lot.
I’m pretty interested in this new ESC. Is it possible to setup limits as easily as on Vesc Tools ?
Is it as fool proof too ?

Regarding the sensors, there are plenty cables with many wires embedded inside.
We could even find one cable with 3 power wires for motor phases and 4 data wires for the sensor.
Then waterproof this one cable with a quality cable gland.

Hi Mauro,

Just wondering : why sincos sensor type over conventional hall sensor ?

I am developing a motor, so including a 4th cable is not a big deal, but is it worth it ? What’s the point reading the exact position of the rotor on efoil ? I wasn’t planning to release the motor with sensors (my test motor will still have a PT1000 sensor on the stator - for test purpose-) might be an available option to choose between PT1000/hall sensors.

https://www.amazon.de/Digital-Drehzahlmesser-Hall-Sensor-Näherungsschalter-blau/dp/B07G2HPK62 possible?

SinCos gives resolution of the sensor, so e.g. 8bit Encoder gives you 1/256 = ~1,4Deg accuracy
And 14bit Encoder gives you 1/16k = ~0,2Deg accuracy
Hall sensor is always 1/3 of electric revolution. So accuracy depending on the motor, but always worse than encoder.

For Industry robot you want high accuracy. And Hall sensors can be as worse as 120Deg resolution.

But for me any sensor other than temperature makes absolutely no sense.
I don’t need to position my Prop in 0.x Deg step and other than that sensors only help the ESC for startup. Startup torque at Props in water is pretty low (compared to a Bike/Skateboard/Car at a hill). And beyond ~2000ERPM the ESC is more accurate by BEMF and FOC model than from any sensor.
TL;DR (in our application) Sensors are only for ESCs who can’t handle startup.

Can you tell the measurements of said ESC?
Dont have to be accurately, I dont want to search for it, just to get an idea of it compared to RC ESC.
I could imagine the industry rated ESC is much bigger and really made for 140A…like if it was speced by hobbyking it would be more like 500A super duper best ESC in the world.

Thanks for the explanation :wink:

Agree with you. And even the temperature sensor would be only useful to quantify the motor’s max power & current during the first tests. No commercial motor worth >1k€ would be available with sensors. It has already been tested, simulated and re tested to avoid the end user to worry about that.

Why sin/cos over hall? The way it’s been sold to me was that it make no sense to work with sinusoidal output and square sensor signal, sin/cos usually requires just a N/S magnet on the shaft and simple pcb, it’s just as cheap as hall sensors. Don’t know if this is the reason but the fact is that sensorless controllers are not common. Most of them handles temp sensor (PT1000 is one among them, but there are others) and as far as I saw, professional (or industrial) application have such, no matter if tested and within safe range etc.
It just doesn’t make sense to save on this kind of safety, given the numbers and the money involved!
For what concern software and foolproofness … I do feel like a jerk many times in the process of program functionalities and sort out the outstanding number or parameters, everything is settable, the current, throttle, torque curve can be adjusted at different values. Anyway, as usual, shared experience will be the key. Will keep you posted hopefully soon