Motor temp meaurement from Hall sensors?

[WheezyRider]

Don't want to re-invent the wheel here, but has anyone ever used the Hall sensors as a means of measuring motor temperature?

Hall sensors are semiconductors, so hole/electron mobility should change significantly with temp. Can this be monitored and calibrated to temperature?

 

[Woosh]

No. The surrounding temperature has little effect on the Hall current.
You use a thermistor for temperature sensing.

The Hall probe consists of a thin rectangular plate made of a semiconductor material. The plate has four electrical contacts with ohmic contact. Two out of them are called biasing contacts, through which the biasing current flows. The other two contacts are called sense contacts at right angle to the biaising contacts; their function is to measure the Hall voltage.
The magnetic field traverses the plate - the biasing current is deflected by the magnetic field causing a Hall voltage to appear on the sensing contacts.
You only see the Hall current which is proportional to the strength of the magnetic field.

 

[WheezyRider]

No. The surrounding temperature has little effect on the Hall current.
You use a thermistor for temperature sensing.

The Hall probe consists of a thin rectangular plate made of a semiconductor material. The plate has four electrical contacts with ohmic contact. Two out of them are called biasing contacts, through which the biasing current flows. The other two contacts are called sense contacts at right angle to the biaising contacts; their function is to measure the Hall voltage.
The magnetic field traverses the plate - the biasing current is deflected by the magnetic field causing a Hall voltage to appear on the sensing contacts.
You only see the Hall current which is proportional to the strength of the magnetic field.

Yes I know, you could use a thermistor, a thermocouple or even a platinum resistor. However, this involves trying to thread yet more wires into the hub and then trying to find a good place to attach the sensor and then hope it doesn't fall off in use.

So, if we can utilise what is already inside the motor, it's a significant advantage.

I wasn't specifically thinking about measuring a Hall current.

The Hall voltage is = IB/nte

Where I is the current passing through the sensor, B is the magnetic field, n is the charge carrier density, t is the thickness of the material and e is the atomic unit of charge. The situation with semiconductors is slightly more complicated as you have holes and electrons, and these have different mobilities.

B is dependent on temperature, as increased temperature reduces the strength of the magnetic field.

I will change with temperature, as in semiconductors, resistance tends to drop with increasing temperature.

t will change slightly with temperature, as materials often expand when heated, but this may be too small to be significant.

n will change with temperature, as in semiconductors there are usually more charge carriers available at a higher temperature.

e is constant.

So you can see, there are several parameters related to the Hall voltage that have a temperature dependency.

Now, the Hall "sensors" we have in our motors are not really sensors, they are Hall switches. Inside they will have a Hall sensor which triggers another circuit to activate a "switch". That "switch" will be some form of transistor, i.e. another semiconductor, whose electrical properties are dependent on temperature.

I can imagine it may be possible to measure temperature with these Hall switches using different strategies.

One could be just monitoring the current flowing in the Hall switch circuit and seeing how this varies with motor temperature. Higher temperature should lead to a slight increase in current in this circuit.

Another method may be looking at the rise time of the switch with temperature. Since the Hall voltage is temperature dependent, the speed at which the charges can separate and reach the opposing sides of the sensor will vary accordingly. This means the switch will take longer or shorter to transition from conducting to non conducting and back again according to the temperature.

As you know the motor RPM, you could even look at how long the switch should be on and off and then measure the on/off pulse width.

These are just a few thoughts I had going through my mind cycling home. I'm sure there are probably many more ways of doing this.

 

[Woosh]

So you can see, there are several parameters related to the Hall voltage that have a temperature dependency.

how are you going to extract the temperature related change?
You'll usually need a reference to measure any quantity. Where does the reference come from?

 

[RossG]

I can't help thinking your solution to temp monitoring is a bit like using the old sledge hammer on it Wheezy.
What happens when you notice the heat rising, pedal slower or is that faster with an ebike ?

When I built my last Computer I wanted mega power but with a machine that ran absolutely quiet, so I fitted temp sensors on every component that saw rises in temperature when the Computer worked harder.
There're all connected to a separate micro computer that constantly monitors the state of the system and increases cooling when and where necessary , hence a totally silent machine except when driven hard.
Unless you do something with that temp information on your bike like say trigger an alarm or visual indicator, you'll be riding along looking at a display instead of the road ahead !

 

[Woosh]

Now, the Hall "sensors" we have in our motors are not really sensors, they are Hall switches.

you are wrong.
The Hall sensors in our motors are usually Honeywell SS41 or similar.
The output is connected to a comparator in the controller.
Although the biaising current varies with temperature, the sensors' positive supply is connected to the 5V rail, there is no provision for using Hall sensors as a temperature probes.

As RossG said, you tried to crack a nut with a sledge hammer.

https://sensing.honeywell.com/honeywell-sensing-bipolar-hall-effect-digital-position-sensor-IC-ss41-l-t2-t3-s-sp-datasheet-32312814-b-en.pdf

Lots of direct drive motors have a temperature sensing thermistor fitted.
Look at the MAC motors and kits sold by Justin @ ebikes.ca

 

[WheezyRider]

I can't help thinking your solution to temp monitoring is a bit like using the old sledge hammer on it Wheezy.
What happens when you notice the heat rising, pedal slower or is that faster with an ebike ?

When I built my last Computer I wanted mega power but with a machine that ran absolutely quiet, so I fitted temp sensors on every component that saw rises in temperature when the Computer worked harder.
There're all connected to a separate micro computer that constantly monitors the state of the system and increases cooling when and where necessary , hence a totally silent machine except when driven hard.
Unless you do something with that temp information on your bike like say trigger an alarm or visual indicator, you'll be riding along looking at a display instead of the road ahead !

Why should anyone care about motor temperature?

If you had to climb a 10% hill with a small hub motor, the simulation shows it getting to 250deg C in just 4 mins, which would be bad news for the motor.

Motor Simulator - Tools

Our ebike motor simulator allows you to easily simulate the different performance characteristics of different ebike setups - with a wide selection of hub motors modeled, and the ability to add custom batteries and controllers and set a wide variety of vehicle parameters you'll be able to see...

www.ebikes.ca

Even a 5% incline would cause serious overheat in 13 min at 77% throttle. An example would be if you were climbing a long incline, up in the hills.

Motor Simulator - Tools

Our ebike motor simulator allows you to easily simulate the different performance characteristics of different ebike setups - with a wide selection of hub motors modeled, and the ability to add custom batteries and controllers and set a wide variety of vehicle parameters you'll be able to see...

www.ebikes.ca

The vast majority of hub motors seem to have no provision for temperature monitoring, even though it is not difficult for them to be put in a situation where they could overheat.

So if there is a way of adapting an existing motor to give temperature readings, without modifying it then I think that is worthwhile.

It doesn't have to be extremely accurate, just a warning to the rider that things are getting too warm and they need to stop.

 

[vfr400]

Why should anyone care about motor temperature?

If you had to climb a 10% hill with a small hub motor, the simulation shows it getting to 250deg C in just 4 mins, which would be bad news for the motor.

Motor Simulator - Tools

Our ebike motor simulator allows you to easily simulate the different performance characteristics of different ebike setups - with a wide selection of hub motors modeled, and the ability to add custom batteries and controllers and set a wide variety of vehicle parameters you'll be able to see...

www.ebikes.ca

Even a 5% incline would cause serious overheat in 13 min at 77% throttle. An example would be if you were climbing a long incline, up in the hills.

Motor Simulator - Tools

Our ebike motor simulator allows you to easily simulate the different performance characteristics of different ebike setups - with a wide selection of hub motors modeled, and the ability to add custom batteries and controllers and set a wide variety of vehicle parameters you'll be able to see...

www.ebikes.ca

The vast majority of hub motors seem to have no provision for temperature monitoring, even though it is not difficult for them to be put in a situation where they could overheat.

So if there is a way of adapting an existing motor to give temperature readings, without modifying it then I think that is worthwhile.

It doesn't have to be extremely accurate, just a warning to the rider that things are getting too warm and they need to stop.

The simulator doesn't take into account cooling affects when riding your bike. The data is based on extrapolated static tests. The overheat times are wildly underestimated. In real life, who has ever burnt a small hub-motor when going up long steep hills? I can't remember it ever happening in the 10 years I've been on this forum.

 

[WheezyRider]

The simulator doesn't take into account cooling affects when riding your bike. The data is based on extrapolated static tests. The overheat times are wildly underestimated. In real life, who has ever burnt a small hub-motor when going up long steep hills? I can't remember it ever happening in the 10 years I've been on this forum.

Ok, that's good to hear.

It's still of interest to me as I like to experiment and push things beyond spec.

The other aspect of heating I worry about is permanent magnetic field loss due to high temperatures. So even if the motor doesn't completely burn out, does a bad heating event cause a permanent loss of power?

 

[WheezyRider]

you are wrong.
The Hall sensors in our motors are usually Honeywell SS41 or similar.
The output is connected to a comparator in the controller.
Although the biaising current varies with temperature, the sensors' positive supply is connected to the 5V rail, there is no provision for using Hall sensors as a temperature probes.

As RossG said, you tried to crack a nut with a sledge hammer.

https://sensing.honeywell.com/honeywell-sensing-bipolar-hall-effect-digital-position-sensor-IC-ss41-l-t2-t3-s-sp-datasheet-32312814-b-en.pdf

Lots of direct drive motors have a temperature sensing thermistor fitted.
Look at the MAC motors and kits sold by Justin @ ebikes.ca

Thanks for the data sheet Woosh, it confirms a lot of what I've said.

The devices are digital, not analogue, inside these Hall devices there is circuitry to make it operate as a digital switch:

"The built-in regulator provides enhanced stability of operation from 4.5 Vdc to 24 Vdc supply voltage range, and internal circuitry is designed to prevent sensor damage in case the supply voltage polarity is accidentally reversed. The open-collector sinking output voltage is easily interfaced with a wide variety of electronic circuits."

As you can see from Fig 2. the current varies by a factor of 2 between 0 and 150 Deg C, so this could be measured and used as an indicator of motor temperature.

It would be a case of measuring the current at different known temperatures and using this to calibrate measurements.

DD motors may already have embedded temp sensors, but I don't have any, I have geared hub motors and trying to fit and wire an additional sensor is not straightforward.

 

[vfr400]

The other aspect of heating I worry about is permanent magnetic field loss due to high temperatures. So even if the motor doesn't completely burn out, does a bad heating event cause a permanent loss of power?

It's not possible to say without objective test data; however, I can't ever remember someone complaining that their motor had gone weak after climbing a steep hill. Maybe if they lost 5% they wouldn't notice.

 

[scott gaza]

Standard grade neodymium magnets have a maximum operating temperature of 80 degrees Celsius. When heated above this, they will experience irrecoverable losses in performance. High temperature grades of neodymium magnets with higher maximum operating temperatures are available but are they used ? Good luck Wheezy Neodymium magnets will lose a certain amount of performance for every degree rise in temperature even if the temperature is below their maximum operating temperature. In fact, depending on size, shape, grade and how it is used, a neodymium magnet will lose 0.08%-0.12% of its magnetic strength for every degree Celsius rise in temperature.
https://www.amazon.co.uk/Digitemp-Level-Horizontal-Scale-Thermometer/dp/B00SHCK15C/ref=sr_1_95?dchild=1&keywords=stick+on+thermometer+strip&qid=1594469401&sr=8-95
£2:50 Well spent.

 

[vfr400]

In that diagram the output is digital - high or low.

 

[vfr400]

Standard grade neodymium magnets have a maximum operating temperature of 80 degrees Celsius. When heated above this, they will experience irrecoverable losses in performance. High temperature grades of neodymium magnets with higher maximum operating temperatures are available but are they used ? Good luck Wheezy

Internet theory, not borne out in practice, otherwise we'd be all pushing our bikes instead of riding. Like a lot of stuff, probably taken out of context or from some special case completely irrelevant to what we have.

How much irrecoverable loss - 0.00001%? It's not quantified.

 

[joelectric]

Possible to mount a thermocouple to the outside of the hub and measure this way from the thermal transfer? Obviously it wouldn't be completely accurate and would be influenced by ambent Conditions but it might be useful enough to monitor the change in temperature.

 

[Woosh]

if you want to read the temperature of the motor, there is a standard solution.
In the 9-pin motor plug, you have 3 phase wires, 5 Hall wires (yellow, blue, green, +5C and ground) and a white wire for the speed sensor.

The output of the LM35 is 10mV per degree C.
Your problem would be how to use the temperature signal.
I guess you can connect the output of the LM35 to a LED to see how bright it glows.
You can get an external speed sensor and use the white wire for your temperature probe. Wire an LM35 temperature sensor to +5V and ground, wire the output of the LM35 to the white wire.
Some guy connects it to an Arduino and insert the temperature into the datastream between the controller and the LCD.

 

[flecc]

[QUOTE="WheezyRider, post: 569530, member: 29386"
If you had to climb a 10% hill with a small hub motor, the simulation shows it getting to 250deg C in just 4 mins, which would be bad news for the motor.

Even a 5% incline would cause serious overheat in 13 min at 77% throttle. An example would be if you were climbing a long incline, up in the hills.
[/QUOTE]

Forget simulation, this absolutely does not happen due to the cooling vfr400 mentions. I've carried out a test 12 years ago and you can read about it on the link below, showing a slight temperature drop!

LINK
.

 

[vfr400]

This has been an interesting thread. Lot's of information and discussion is always good.

I quote Flecc from that 2017 thread in the link above.

All this proves nothing other than motor overheating affecting reliability is nothing but theory, in practice it's not an issue with decent motors.

This is it in a nutshell. For most of us, overheating is just not a problem. As I said above, nobody is burning motors, so it's not a problem we need to solve. It's only when pushing your system to well beyond it's normal limits that temperature might be an issue, and I don't mean running a 36v motor at 48v.

If you want to measure temperature anywhere, you'd be better off measuring the controller's temperature. That's the thing that goes first.

 

[Woosh]

Lishui controllers have already got thermal cutout.
Same as most batteries.

 

[RossG]

Why should anyone care about motor temperature?

If you had to climb a 10% hill with a small hub motor, the simulation shows it getting to 250deg C in just 4 mins, which would be bad news for the motor.

Motor Simulator - Tools

Our ebike motor simulator allows you to easily simulate the different performance characteristics of different ebike setups - with a wide selection of hub motors modeled, and the ability to add custom batteries and controllers and set a wide variety of vehicle parameters you'll be able to see...

www.ebikes.ca

Even a 5% incline would cause serious overheat in 13 min at 77% throttle. An example would be if you were climbing a long incline, up in the hills.

Motor Simulator - Tools

Our ebike motor simulator allows you to easily simulate the different performance characteristics of different ebike setups - with a wide selection of hub motors modeled, and the ability to add custom batteries and controllers and set a wide variety of vehicle parameters you'll be able to see...

www.ebikes.ca

The vast majority of hub motors seem to have no provision for temperature monitoring, even though it is not difficult for them to be put in a situation where they could overheat.

So if there is a way of adapting an existing motor to give temperature readings, without modifying it then I think that is worthwhile.

It doesn't have to be extremely accurate, just a warning to the rider that things are getting too warm and they need to stop.

Yes of course you should care about the temperature but it's an affect you are aware of anyway, when you stress your motor it gets warmer. A temperature readout while interesting to observe wouldn't correct that effect, that's down to you. Unless you have some sort of automatic control system that cuts power at a set temp level like my computer that initiates an action when required, I can't really see the point.