Electronics > Projects, Designs, and Technical Stuff

Finding a controller to maximize torque in salvaged washing machine motor?

(1/3) > >>

I have a huge 310v, 2.8A, 0.92hp COM2000 direct drive motor from a Samsung washing machine, which I am trying to salvage in the process of turning the washing machine into a cannon.

The goal is to have the motor pull back the drum (which is attached to springs at the front of the gutted washing machine) like a sling shot, with some load in the drum (watermelon, lego millenium falcon, basketball, etc.), and then release it with a trigger mechanism. It functions like a giant sling shot. Everything is already built and works, I am just trying to make it more powerful, and figuring out how to access the beautiful torque in this motor is the bottleneck for me (previously I have just been turning the motor with my hand). I also have very redundant safety precautions, and a huge private property to do this on.

My goal is to find a motor controller that can maximize the torque of the motor, to maximize elastic potential energy in the springs.

This challenge (with almost an identical motor) has been attempted to be solved before on this blog, but their solution was to just use the original controller from the washing machine: https://www.eevblog.com/forum/projects/building-a-motor-controller-for-a-samsung-brushless-direct-drive-washer-motor/
^(This is a pretty long thread, but towards the end you can find the important detail- the hall effect sensors of these motors are different from most.)

I stupidly threw away the original controller when I first took it apart, so the only option here is to build/buy a new one. I know that people have successfully used e-bike ESC's like this one https://www.amazon.com/dp/B08HWF7MZ7?psc=1&ref=ppx_yo2ov_dt_b_product_details
to drive motors exactly like this. So it is definitely possible.

I am trying to figure out:
1. The optimal specs of such a motor controller in the interest of maximizing torque (there are no limitations on available power, rpm requirements, etc.- I only care about torque)
2. Would rewiring the stator like this unlock more possible torque?
My understanding is that this lowers the required voltage and increases possible current, and more current equals more torque. Is this correct? I am trying to educate myself on what rewiring the stator would accomplish, and other general theoretical info about this problem, but all online sources about it seem to just give instructions without diving into the theory.


Rewiring/rewinding does not unlock any more torque because torque is finally limited by the iron saturating. The only thing you could do by rewiring is change the voltage/current relationship (i.e., change the motor from low-voltage high-current into high-voltage low-current or vice versa, with power and torque characteristics kept the same); this is useful if you have a power supply / battery pack of certain voltage too far away from what the motor was designed for. One could also increase the efficiency and maximum continuous power rating by rewiring if they are able to squeeze in more copper (factory-made motors have some empty room to allow easier machine insertion of windings), but this effect is small and it seems long-term power rating (in minutes, limited by heating of the windings) is not that important for you. Therefore, no rewinding needed.

The key to maximum torque is to forget sensorless drivers and sense the rotor position (hall sensors, encoder or something else, the exact type doesn't matter). Then any controller with suitable current/voltage ratings pretty much perform the same.

You can exceed the rated motor current by maybe 2-3 times, but after some point, the iron saturates hard enough that returns in torque are diminishing and any extra current is just increasing resistive losses without generating much incremental torque. Maximum current is best found out by testing. Keep the testing short (a few seconds) to avoid burning the windings at highest currents.

Your thinking mistake in rewiring business is that while you can indeed increase the current by lowering the number of turns, being able to wind with thicker wire, then again magnetic flux (and thus torque) is really not defined by current, but current * number of turns. This works out to a constant, meaning that given certain amount of copper and iron, motors can be wound for any voltage / current relationship, within some practical limits, and there is very little if any difference otherwise; a 1A, 100V motor with 100 meters of winding wire is very close in all performance parameters to a 10A, 10V motor with 10 parallel strands 10 meters in length each.

Thank you very much for the response!

I ordered this controller a few days ago and it just came in:

Here is the problem I am trying to solve. The motor has built-in hall effect sensors, but not in the traditional way. It has 4 pins- 2 for U and V, and one for voltage, one for Gnd.
This (and seemingly every other consumer ESC that would fit this application) has a 5-pin sensor, 3 for hall & 1 for voltage, 1 for GND.

This post on this forum from about a year ago talks about this exact problem with this exact type of motor, but the OP had the original driver to use. I don't.
If you scroll to the bottom you can see them talk about these things.

Do you think there is any way I could connect this 4 pin hall effect sensor to a 5 pin one? Is my best option to just try to replace it with a 5 pin one? I am trying to see if there is a simple, less intrusive solution to this.

Thanks again

Maybe it only has two hall sensors like Benta speculated on the linked thread. Converting that to simulate the expected three would be quite complicated, at that point I would consider rather designing your own controller or picking one of the "open source" ESC projects and modify their code.

Low-speed high-torque application would benefit from more accurate rotor angle sensing than achievable with just two hall sensors which will require interpolation. In the intended use of washing machine, motor always rotates at a significant enough speed so that even sensorless is an option. If I understood correctly, you want instantly full torque from nearly standstill, in which case not only sensorless goes out of window but 2-hall sensing is a bit iffy, too. You could always consider retrofitting more accurate angular sensing.

The 3 hall sensors give 6-step output per electrical revolution, meaning electrical angle will be at most +/-30 deg off from any instantaneous optimum value. With 2 sensors that would drop to 4 steps and +/- 45 deg respectively which is a big difference as 45 degree error from the optimum 90deg phase shift means significant drop in torque and increase in reactive current. So while 6-step sensing benefits from angular interpolation, it becomes nearly mandatory with 4-step sensing.

I think it definitely is the case that it has 2 hall sensors.

It doesn't need to necessarily go straight to instantly full torque, it would fit the application needs if it started lower and worked its way up as it turned. Does this change the feasibility of trying to do this with 2 hall sensors? Also, I could easily implement a gear-lock system on the motor to prevent it from turning in the wrong direction if the motor stalls. Does having a motor that is allowed to stall mean it is ok to run it sensorless?

Could you elaborate on what you mean by retrofitting more accurate angular sensing? I don't have much experience with using/installing hall effect sensors, but from some research on it, it seems like this can be pretty tough to pull off right, especially without an oscilloscope. Am I wrong in perceiving that as a very complicated task?


[0] Message Index

[#] Next page

There was an error while thanking
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod