Simple thermistor circuit

Thread Starter

jonescg

Joined Jul 24, 2011
16
Hi All,

New to the forum, so play nice :) My name is Chris and I know little about electronics. I have a sound knowledge of Kirchoff's rules and paralleled vs series circuits, so I'm probably not a complete nooob, but I am still struggling with things like transistors. Don't even start on ICs! Woah!

Anyway, I have built an electric motorcycle which I am racing in the Australian TTX-GP. It's great to be a part of electric motorsport, especially now when it's still a bit underground. Mission Motors has already blown the field away in the US, so it won't be long before the big boys muscle us back-yarders out of the way.

My bike is called Voltron, the Electric RG and it's pretty cool. Actually, it gets very hot, hence the need for the thermistor circuit... I managed to crack 180 km/h down the straight but the two DC motors I am running seemed to go out of balance and one took the full load. It overheated in no time, and melted, spraying molten tin over everything :eek:



So I think a thermistor circuit could have saved the motor. The Agni motors are brushed DC motors which have nominal power of 12 kW, and a peak of 30 kW. They get very hot if you abuse them, or try to dump 800 amps into it.

So what I would like is to set up a circuit that makes use of the thermistors which are mounted in the brush housings. These thermistors are about 40 kΩ at room temp, and 2kΩ at maximum acceptable temp. I found a circuit on the web which uses a couple of transistors to activate an LED when the thermistor is heated, presumably because 2kΩ is too high a resistance to light up an LED. The circuit is found here:
http://www.technologystudent.com/elec1/therm1.htm
But there are no values for each resistor. I figure this would be a neat little circuit that I can use, with a super bright LED staring me in the face so I can see it, even when I'm focussed on the next turn.

So how do you recommend I determine what each resistor needs to be? The supply voltage will be 12 V, not 9 V, so the resistor in series with the LED can be about 900Ω to prevent it from burning out. What should the preset resistor be?

Keen to hear some tips and advice! Thanks and hello!

Cheers,
CHRIS
 

Adjuster

Joined Dec 26, 2010
2,148
The circuit you are proposing to use will never be vary stable, because its turn-on temperature will depend on the battery voltage and on the temperature of the transistor - not just that of the thermistor. It really is too basic a circuit for a safety application.

More importantly, you say that you are trying to build an electric motorbike with limited electrical knowledge. You have already had a major burnout, and you say don't know how to deal with transistors or ICs. - I think therefore that your first priority would be to team up with somebody better informed. Unfortunately, that can't happen on this forum, as automotive topics are not allowed to be discussed.
 

#12

Joined Nov 30, 2010
18,224
You need 212 ohms, give or take 10%. Make it adjustable with a potentiometer. It won't be very dependable because the heat will make the transistors drift. Still, this is better than nothing. You really need a differential amplifier, but you won't get it here because the bosses will shut this down too quickly.
 

iONic

Joined Nov 16, 2007
1,662
Did you figure out at what temperature your motor blew up at? This and the proper choice of thermistor or temperature sensor and it's position will be important. It's only because what you hope to accomplish with such a circuit that I post anything at all. In the end safety should be enhanced greatly.

Perhaps a better safety mechanism would be to somehow cause a throttling down when the motor gets too hot, despite what you would like to do with respect to speed. If your reaching critical mass, then you are already "dead in the water".

Another idea is to incorporate some cooling mechanism with opening vents (or permanent vents) that draws air at the motor...I know...drag, but drag is better than explosions.

My best advise is to locate a seasoned mechanic, design engineer, or electrical engineer that can assist you with a safe and efficient fix to a very dangerous problem. Bottom line, I think you need more than a simple circuit.
 
Last edited:

Pencil

Joined Dec 8, 2009
272
If you are truly pursuing the circuit
as shown and you are not very experienced,

I found a circuit on the web which uses a couple of transistors...
Those transistors are in what is called a "Darlington pair"
configuration. You can make this from individual transistors
or buy them as a single component.

I would also seriously consider previous advice about stability.

Hope this helps you in some way.
 

wayneh

Joined Sep 9, 2010
17,496
I'd use an LM35 thermometer IC and a quad comparator. That would give you 4 "levels" in response to increasing temps. Maybe first light an LED, then 2, then start limiting current, then shut down, as temperature rises. You'll need supporting circuitry for establishing a voltage reference, dealing with transients, turning the comparator output into something useful (more than just the LEDs), etc. If this all sounds like too much, you need to find someone to help you.
 

Adjuster

Joined Dec 26, 2010
2,148
On re-reading this, I strongly feel that you should look for the services of a local expert, for more than advice on thermometers.
I managed to crack 180 km/h down the straight but the two DC motors I am running seemed to go out of balance and one took the full load. It overheated in no time, and melted, spraying molten tin over everything :eek:
This quote suggests that you don't know for sure what is going on with critical aspects of the design. If you seriously think that the load distribution can get out of control in the way you describe, you need to know exactly why this occurs, and how to stop it happening again.

A thermometer may help with simple overheating, but if there is some really nasty electrical instability going on you may find that things get damaged before your thermometer can possibly react.
 

Thread Starter

jonescg

Joined Jul 24, 2011
16
Hi everyone, thanks for your kind words, support and encouragement :rolleyes:

Building an electric motorbike is actually quite simple. You take the motors, controller, batteries and a motorbike, ensure all are within their rated capabilities, and assemble. Building an electric race bike means doing the above, except the motors will be pushed to their limits, and you never know where the limits are until you've crossed them.

I don't know how many here have built electric vehicles for racing applications, but to come out saying I'm under-qualified to be doing it, or that I need to find a more more experienced technician cause I blew up a motor is pretty funny. Ferrari blow up motors all the time. And they're at the top of their game!

Anyway, thanks for letting me know that the circuit above is not so hot. There are other kits around that involve ICs, and are a lot more stable. The thermistor is frequently wired into the throttle pot, so when it reaches it's maximum temperature it causes the controller to back off. This is great if you have one motor, but two motors is a bit more complicated.

Cheers,CHRIS
 

Thread Starter

jonescg

Joined Jul 24, 2011
16
Thanks Mr Wookie!

I'll have a go at this one and play around with it. I'll have the thermister circuit wired up so it only comes on when the bike is on, and it's never going to be on for more than an hour. I have a 5 Ah 13 V auxiliary battery for all the 12 V stuff.
 

mcasale

Joined Jul 18, 2011
210
Hello. I think the electric bike is really cool. I am involved with a start up doing electric vehicles in Delaware. Anyway, another suggestion might be to look at thermal cutouts. They are preset switches (NO or NC) that trigger at a specified temperature. It MAY be a lot simpler than thermistors and comparators. Make sure the cutout has some sort of safety rating (UL, CSA, etc.). Good luck.
 

#12

Joined Nov 30, 2010
18,224
I am so learning from this one. I see that motorcycles are not any kind of automobile. If they were, the moderators would have stopped this thread 2 days ago.

I learn something almost every day on this site.
 

wayneh

Joined Sep 9, 2010
17,496
Good idea. Consumer appliances (like the dehumidifier I just repaired) use these thermal fuses. It's an easy way to cut power if something is too hot. They can even be used to to cycle multiple times like a relay, but of course they're not mean to last forever when used that way.
 

iONic

Joined Nov 16, 2007
1,662
Honestly, I think this topic has been exhausted, and was a "banned" topic from the beginning. Other than suggesting to seek professional help, the thread should not have lasted this long. The more suggestion that come in, the more dangerous it can become.
If the OP seeks the proper professionals they will have all our suggestions and then some.
 

THE_RB

Joined Feb 11, 2008
5,438
Why would you ban this? The rules (I thought) were to stop people making dangerous modifications to road-going vehicles. That is a good rule.

But this is a completely custom built racetrack only vehicle. Surely adding a temperature sensor to tell the motor temp is not going to make this racing event (full of custom built vehicles) any more dangerous to the public?
 

Thread Starter

jonescg

Joined Jul 24, 2011
16
Hi again,

I noticed that Jaycar electronics sells those NO/NC temperature dependent switches. When the thing it is attached to exceeds 90'C it will pop open or closed, meaning a separate relay circuit can activate a light / cut power. I don't want it to immediately cut power, as that would be rather dangerous in a fast turn, so illuminating a big red LED might work nicely. It seems a lot easier than trying to utilise the thermistors inside the motors, especially since they were optimised for one motor at a time, not a paired set.

For those curious as to how I blew up the motor - there are two motors running in parallel from the one PWM-DC motor controller. This way, both motors experience the same voltage. The battery pack is a 32 cell LiFePO4 pack, with a nominal voltage of 105.6 V, and a charge voltage of 117 V. So what happens is one motor, due to reasons beyond my control, had a brush resistance higher than the other. This resulted in one motor working harder than the other; taking a much higher current. The motors are fixed on the one axis, so despite receiving the same voltage, and hence RPM, the currents were way out of balance. 800 amps was being drawn from the battery moments before the motor blew up. Turned out Agni, the manufacturer, had some QC issues meaning some motors were coming out with completely different kV values, so it was impossible to match them, even with advance brush timing.

I have two new motors on their way and they are gong to be vetted before I see them. I will also dyno-tune the bike so they share loads evenly. Should mean a first or second place come September!
 

Jaguarjoe

Joined Apr 7, 2010
767
I would be leery to put an automatic temperature shutdown unit on that bike. At 180km/h sudden upsets like loss of a motor could negatively impact the rider's well being.

Double the voltage and halve the battery amperage and put the motors in series, current will always be equal. Running at higher voltage allows you to use smaller wires and less bulky connections saving a lot of loss in the wiring, both money and powerwise.
If these are not permanent magnet motors, you may be able to tweak the field currents to balance your armature currents.
Rather than wait for a motor to overheat, sense armature current with Hall sensors and get your warning instantly.
Make a hefty MOSFET power splitter and use the Hall sensors to automatically balance the motors.
Most of the electric cars that I've read about have a motor at each wheel. Somehow power is equitably distributed among those wheels. They are probably not PM motors.
 

Adjuster

Joined Dec 26, 2010
2,148
Relying on the motors to be balanced "just so" and stay that way may be a really bad idea. They may go out of balance with use, or even worse there may be a tendency for motor imbalance to degrade further once it develops, e.g. due to reduced magnetisation during overheating leading to a sort of thermal runaway.
I noticed that Jaycar electronics sells those NO/NC temperature dependent switches. When the thing it is attached to exceeds 90'C it will pop open or closed, meaning a separate relay circuit can activate a light / cut power. I don't want it to immediately cut power, as that would be rather dangerous in a fast turn, so illuminating a big red LED might work nicely. It seems a lot easier than trying to utilise the thermistors inside the motors, especially since they were optimised for one motor at a time, not a paired set.
The placement of the thermistor inside your motor was purpose-dsigned, and the type of detector allows some flexibility in setting the threshold. You plan to fit devices preset to 90C, and believe that this will somehow be better optimised for a paired set???
 

Thread Starter

jonescg

Joined Jul 24, 2011
16
Ok so the Agni motor is a brushed, permanent magnet DC motor. So no way can I use speed sensing, hall sensors or anything like that. They are just kind of dumb; feed them a current and they turn. Running them in series would be fantastic, but it means I'd need to totally rearrange my battery pack to about 160 V, which is impractical, and secondly, MOSFETS don't behave properly at voltages above ~110 V.

Several teams around the world are using, and winning with paired Agni motors, running in parallel from one controller. Others use two controllers to run each motor, but that relies on having two controllers that put out the same power at the same rate.

The thermal switch / relay idea is to allow me to monitor each motor individually. Currently, there is a thermistor which can be wired into the throttle signal of the controller, such that when it reaches it's 2kΩ point it forces the controller to back off, no matter how hard you twist the throttle. I can't wire both thermistors in series and make use of this failsafe, as let's say one motor is hot <2kΩ, the other not (say, 10kΩ), it's still going to read >10 kΩ so nothing will shut down. If I wired both thermistors in parallel, and one was getting warm, at 5kΩ and the other at say, 10kΩ, then the total resistance will be about 3kΩ, which is better, but close to a premature shut-down.

Anyway, keep the ideas coming!

Cheers,
CHRIS
 
Top