Q: Sensored controllers

E-HP

10 GW
Joined
Nov 1, 2018
Messages
6,201
Location
USA
I'm bouncing a few ideas around in my head, and had a couple things that I was wondering about.

  • With a sensored motor and regen capable sensored controller, does the controller need the hall sensor signals to provide regen?

Not necessarily related:
  • With a sensored motor and controller, if you disconnect the 5V wire between the controller and motor, will that stop any phase current going to the motor?
 
E-HP said:
With a sensored motor and regen capable sensored controller, does the controller need the hall sensor signals to provide regen?
In theory the answer is no. You have a magnetic field with a bunch of coils rotating through it. They are going to produce a voltage potential regardless of any outside influence. What the controller does with that potential and how it might use the sensors is another ball of wax.

E-HP said:
With a sensored motor and controller, if you disconnect the 5V wire between the controller and motor, will that stop any phase current going to the motor?
As I understand it, some controllers will start in sensor mode and then switch to sensor-less operation after they have the motor spinning. I have also been given to understand that with some controllers if sensor mode fails then they fall back to trying to run in sensor-less mode. They do not need the 5 volts to the sensors to power the phase coils however they may be designed to fail if the sensors do not operate as expected.

I am afraid that it going to be a more of a question(s) of the deign and operation of a specific controller.
 
LewTwo said:
E-HP said:
With a sensored motor and regen capable sensored controller, does the controller need the hall sensor signals to provide regen?
In theory the answer is no. You have a magnetic field with a bunch of coils rotating through it. They are going to produce a voltage potential regardless of any outside influence. What the controller does with that potential and how it might use the sensors is another ball of wax.

E-HP said:
With a sensored motor and controller, if you disconnect the 5V wire between the controller and motor, will that stop any phase current going to the motor?
As I understand it, some controllers will start in sensor mode and then switch to sensor-less operation after they have the motor spinning. I have also been given to understand that with some controllers if sensor mode fails then they fall back to trying to run in sensor-less mode. They do not need the 5 volts to the sensors to power the phase coils however they may be designed to fail if the sensors do not operate as expected.

I am afraid that it going to be a more of a question(s) of the deign and operation of a specific controller.

I guess I'll scrap my idea. I was trying to think of a way to disable regen temporarily on the fly, specifically for when the motor is hot already from climbing, and want to avoid additional heat from regen when descending. However, I just thought of something I can try today...maybe I just hit the power button. :eek:

Thanks for the info though!
 
Cutting 5v power to the halls means they shouldn't be able to provide readable signals to the controller, so after that, it depends on the specific controller.

If it's sensorless-capable, it will probably still do regen, though it may behave differently (for instance, the Grinfineons like this use trapezoidal drive instead of sinewave when the halls are not readable, so regen operates differently too).

If the controller is sensored-only, then if the halls are not readable, the controller will usually shutdown to prevent damage to motor or controller. This should also prevent regen from working, because the controller is in a fault state and shouldn't respond to any control input either. But it does depend on the specific controller design.


But if your whole purpose is to prevent regen, then simply don't engage the ebrake.

If you have ebrake switches in your brake handles and have to use the mechanical brakes, just put an override switch in series with this for any time you don't want regen to engage while you're braking mechanically. If your switches are the type that turn on (short) when the lever is pulled, use a NO (normally open) override switch, so that unless it's engaged, it is open, if it is a momentary type switch. If it's a toggle with only one contact set, then you don't need to worry about NC vs NO, only which way you mount the switch so the toggle is in the position you prefer for "on" vs "off".


Unless your controller is programmed to do "slip regen" where any motor speed in excess of that which should happen from the present throttle setting will cause the controller to try to brake the motor until it slows to the speed the throttle is demanding, then simply not engaging the ebrake will prevent the controller from performing regen.

If you are in a situation where a downhill speed is greater than that which the motor can create at the present battery voltage, then even if the controller is not active (or even "on"), but is still connected to the battery, then it will act as a rectifier and still perform regenerative braking--but it is no longer controlled braking, it is simply current flow from the higher voltage at the motor/controller to the lower voltage of the battery, which places drag on the motor and heats it and the entire current path from it to the cells in the battery up, by however much current flows.

The only way to prevent *that* situation is to electrically disconnect the controller from the battery.

Note that in this situation, there is no load on this generated voltage, and it can climb rapidly beyond the point the controller's FETs can handle it, unless they are rated far above the normal battery voltage. This will blow up the controller, or at best damage the FETs or the controller's LVPS (that makes the 5v, 12v, etc inside) so they no longer operate correctly.

You can calculate the approximate voltage the motor will generate if you know it's kV and the speed it will spin at, worst case. If you don't know the kV you can guesstimate it. Take the speed the motor spins at, unloaded, at max throttle on your full battery. Divide the volts by the speed. Then take the speed the motor will spin at at the maximum downhill speed you might ever achieve, and multiply that by the result. That gives you an approximation of the voltage the motor could output under that condition.

If the controller, and anything else connected to the controller's battery wires (other than the battery that isn't connected in this scenario), are not rated for *at least* that voltage, preferably higher, damage may occur.


In some scenarios it is much better to just let this kind of generator braking occur to the battery, because the consequences are easier to deal with than those without it connected. ;)



E-HP said:
I guess I'll scrap my idea. I was trying to think of a way to disable regen temporarily on the fly, specifically for when the motor is hot already from climbing, and want to avoid additional heat from regen when descending. However, I just thought of something I can try today...maybe I just hit the power button. :eek:
E-HP said:
I'm bouncing a few ideas around in my head, and had a couple things that I was wondering about.

  • With a sensored motor and regen capable sensored controller, does the controller need the hall sensor signals to provide regen?

Not necessarily related:
  • With a sensored motor and controller, if you disconnect the 5V wire between the controller and motor, will that stop any phase current going to the motor?
 
@E-HP, Assuming we're talking direct drive? If you flip your bike over on it's back and give that rear tire a spin with the power off, you're going to see that it doesn't take long to stop spinning, where the front tire goes on for quite a while. That's an issue you aren't going to eliminate unless you apply a small amount of power to it (like Grin does on the GMAC setup).

You likely know this, but that residual power being created while coasting is the reason for the clutches on the geared hub and mid drive designs - allowing them to coast more freely.
 
amberwolf said:
Unless your controller is programmed to do "slip regen" where any motor speed in excess of that which should happen from the present throttle setting will cause the controller to try to brake the motor until it slows to the speed the throttle is demanding, then simply not engaging the ebrake will prevent the controller from performing regen.

If you are in a situation where a downhill speed is greater than that which the motor can create at the present battery voltage, then even if the controller is not active (or even "on"), but is still connected to the battery, then it will act as a rectifier and still perform regenerative braking--but it is no longer controlled braking, it is simply current flow from the higher voltage at the motor/controller to the lower voltage of the battery, which places drag on the motor and heats it and the entire current path from it to the cells in the battery up, by however much current flows.
That's my situation, using slip regen controlling power and regen with the same throttle. So are you saying shutting off the controller and riding downhill can still damage the controller? I was going to go out and try it today, but if that's the case, I won't.
 
Yes, it is possible if the voltage climbs high enough from just the motor spinning (it is always a generator as long as it is spinning, whether or not you're applying power to it or tapping the generated power to cause braking).

If the battery is connected, then as long as the BMS does not shut down (disconnecting it from the controller and allowing sudden spike of voltage), the battery itself will load the output down and keep it from rising too high (since it will be drained down from the climb enough that it won't refill to full and overcharge).

If the battery is *not* connected, then there is a risk of damage.

You can make a good guesstimate if this will be a problem by determining the voltage the system will generate under those conditions, with the "math" I gave above. It is still just a guesstimate, of course. ;)


Some controllers have an "enable" pad for regen inside them, that when not grounded disables all regen. But there's no obvious consistently-used marking for this, so unless someone has already posted for your specific controller which pad it is to enable regen, it might be difficult to figure out (even if yours does have this pad)...but if it does, it should be connected to ground via a solder bridge, a jumper, or a wire. Placing a switch in this would let you disconnect regen when needed (just lke putting one in the ebrake line), if it works in your controller this way.






E-HP said:
That's my situation, using slip regen controlling power and regen with the same throttle. So are you saying shutting off the controller and riding downhill can still damage the controller? I was going to go out and try it today, but if that's the case, I won't.
 
amberwolf said:
Yes, it is possible if the voltage climbs high enough from just the motor spinning (it is always a generator as long as it is spinning, whether or not you're applying power to it or tapping the generated power to cause braking).

If the battery is connected, then as long as the BMS does not shut down (disconnecting it from the controller and allowing sudden spike of voltage), the battery itself will load the output down and keep it from rising too high (since it will be drained down from the climb enough that it won't refill to full and overcharge).

If the battery is *not* connected, then there is a risk of damage.

So, battery connected, controller switched off, battery less than full charge, and descending a hill. Risk of damage? But even if not, seems like it might still experience motor braking (and if so, creating the heat that I'm trying to avoid).

And, sensored controller, switching off the hall signals, may still potentially damage the controller.

Just want be sure before trying anything.
 
E-HP said:
So, battery connected, controller switched off, battery less than full charge, and descending a hill. Risk of damage? But even if not, seems like it might still experience motor braking (and if so, creating the heat that I'm trying to avoid).
Damage, probably not. But it will still generate voltage any time it is spinning, and if that's higher than the battery voltage, you will get braking, and heating. How much depends on the current flow.



And, sensored controller, switching off the hall signals, may still potentially damage the controller.
If it cause the controller to still operate but use the wrong timing, yes, it could happen. If the controller correctly shuts down due to the error condition, then that won't happen, but you still have the same effect of the controller rectifying the generated voltage from the motor back into the battery.
 
amberwolf said:
E-HP said:
So, battery connected, controller switched off, battery less than full charge, and descending a hill. Risk of damage? But even if not, seems like it might still experience motor braking (and if so, creating the heat that I'm trying to avoid).

Damage, probably not. But it will still generate voltage any time it is spinning, and if that's higher than the battery voltage, you will get braking, and heating. How much depends on the current flow.

OK, I'm going to go try descending a small hill, and shutting off the controller and see what happens. Battery connected of course. :thumb:
 
BTW, there have been people using a thermal switch in a motor to cut power to hall sensors to cause the sensored-only controller to fault out and shutdown, for some pretty old systems years back that didn't have any built in thermal monitoring, when "overclocking" controllers and motors. But not tested for the situation you're wanting to do it in.
 
amberwolf said:
BTW, there have been people using a thermal switch in a motor to cut power to hall sensors to cause the sensored-only controller to fault out and shutdown, for some pretty old systems years back that didn't have any built in thermal monitoring, when "overclocking" controllers and motors. But not tested for the situation you're wanting to do it in.

Actually pretty cool. Old school thermal rollback/shutdown.

EDIT: Shutting off the controller worked! The motor just coasted, as much as a direct drive hub can coast, with no braking. I guess I could still experiment with the 5V wire to the halls using a momentary switch, but this avoids having to add another button, for the rare occasion that I would use it.
Thanks for all of your guidance! :bigthumb:
 
Back
Top