CVT for a low power electric engine

[Artur Yukio]

Hello guys, I`m a new member here and also a member of a FSAE team here in Brazil. We are a new team and we are going to design and fabricate an electric formula style car.

The thing is, due to the lack of options and resources we had to stick with a donated electric engine. This engine has a rated torque and power of 35 Nm and 11 kW, respectively, and a peak power and torque of 100 Nm and 22kW. These values ??are considerably lower than the average electric fsae cars and seeing that we estimate the weight of our car to about 320 kg without pilot.

Because of that we have a great challenge when talking about transmission. If we assemble a single speed transmission we will have to choose between having high acceleration and low top speed, low acceleration and higher top speed or medium acceleration and top speed. So we thought about a multi speed gearbox, but beacause of the complexity to design and fabricate or suit that kind of gearbox to an electric engine we started to think about other options. That`s when we started to study the possibility to implement a cvt.

As I am responsible for the project of our transmission system I started to search about the possibility of application of a cvt plus a planetary gear reduction on our electric engine thinking that the almost infinite gamma of reductions that it has could resolve our problem. But that was when I came across several divergent opinions and theories.

Most of the arguments against the application of a cvt for electric engines were based on the fact that the electric engines usually can deliver high and constant torque for a long time or because usually they have a good torque curve, which I dont think is our case. I`d like to know if you guys could help us giving your opinion about our situation and possibility of this application and excuse me for any mistakes I may have made in grammar. Thanks!


Artur, Cheetah E-Racing Unifei.

[Adam Farabaugh]

What is going to cause your car to be 700lbs w/out driver?

I am the lead engineer of Penn Electric Racing. We're still in the design phase for our car but with dual motors / controllers and ~150lbs of batteries we are on track to hit 525lbs or better (no driver).

Do you have a torque/speed graph for the motor? What about a torque/current graph? If you do, then it would be fairly easy to write a simulation of the acceleration event in Matlab, given the basic parameters of your car.. weight, battery SOC, etc. If you aren't going to score any points at all in acceleration then you might want to rethink your drivetrain.. the motor may not be at all appropriate.

Care to share any more info on the motor? It would be easier to help if so.

[MCoach]

I'd like to see a break out of where you think the weight is in the car. I've done that the past several years to focus on what can be better designed for the project. For the sake of "trade secrets" or whatever people are concerned about these days, I'm not looking for like "wheel bearing: 0.25kg" but more so something like this:

Motor: 45kg
Frame: 40kg
Batteries: 65kg
Unsprung mass (entire corner assembly): 14.2kg

etc., etc.

This will help narrow down where some work could be used to reduce that. My guess is the frame, as that seems to be the most common component that causes weight gains.

[Francis Gagné]

Originally posted by Adam Farabaugh:

Do you have a torque/speed graph for the motor? What about a torque/current graph? If you do, then it would be fairly easy to write a simulation of the acceleration event in Matlab, given the basic parameters of your car.. weight, battery SOC, etc. If you aren't going to score any points at all in acceleration then you might want to rethink your drivetrain.. the motor may not be at all appropriate.

Care to share any more info on the motor? It would be easier to help if so.

I agree, but even simpler, use OptimumLap for the first iterations, at least you won't have to question too much the validity of the results. You won't waste time programming and debugging, and there is already some FSAE tracks ready. (If you use OptimumLap, don't forget to turn on Advanced tire parameters in the advanced option pane! It changes results in term of mass a lot on most of FSAE tires)

Compare the two concepts, take your "base car" and add the weight of the CVT. Use the peak values of the motor for now in the engine data. Adjust the drivetrain efficiency to reflect the CVT. Run an endurance run. Then you can take the value of energy used (and multiply by number of endurance laps) and compare it to your "base car". You will have to add more batteries to compensate for the CVT loss and the added mass. Then add the mass of the additional battery series and run again. Verify that you have more than enough battery energy to run the endurance.

With OptimumLap you will be able to select a final drive ratio rapidly for the direct drive "base car". The software will also help you to find a good approximation of the needed top speed.
If there is not a lot of difference in concepts (1 or 2 seconds per lap (maybe more, it's your call)). Take the simpler route, finish before and test more!


The CVT losses will also mean that you will have less power going to the wheels, and the engine will probably be used more, this means cooling will be more of an issue.
Is your engine air or water cooled? Cooling will be the difference between running at "constant" or "peak" values of the motor.


What is your motor max RPM? And max Voltage? If you are going the Lincoln, max voltage is 300V. This mean the the nominal voltage of the battery will be lower when running, in the range of 260V maybe (depending on battery technology), and even lower when the battery will be nearly out at end of endurance. Take that into account when calculating the max RPM of your motor (if it can take a higher voltage than 250V don't worry about it now).

[Artur Yukio]

Thanks, guys, for the tips. It was very helpfull I appreciate it. With respect to the high weight of our car, we estimated it taking as a basis the ICE car that a co-team built last year, their first car. It weighed 280 kg, so we calculate: 280 kg - 60 kg (IC Engine and components) + 60 kg (batteries) + 25 kg (motor and controller) + 10 kg (Wiring, relay, and other components) = 315 kg. But we are working on a new estimative with lighter components such as frame, wheels and careen.

The motor is air cooled, but we are studying and working hard on a efficient and effective cooling system.

I have the torque/power x rotation speed curves, which I`m going to post here. http://imageshack.us/a/img27/2166/nxav.jpg

It is in portuguese but here are the translations:

Motor de indução trifásico = three phase induction motor
Torque = Rated Torque (blue line)
Torque Máximo = Peak Torque (Yellow line)
Potência = Rated Power (Red line)
Potência Máxima = Peak Power (Green line)

Rated Power : 11kW (15 HP-cv)
Rated Rotation Speed : 2925 rpm
Rated Voltage : 60 V
Max RPM : 9000 rpm
Max Voltage : 72 V
Max Power : 22 kW
Rated torque : 3.66 kgfm (35,8 Nm)
Max torque : 280%
Max current : 300 A

The curve does not agree with the value that i said about peak torque, because the manufacturer missed it and we are waiting for new curves.

I just got the Licence key for the OptimumLap, and starting to learn how to use it. Thanks again for the help, guys!

[Kirk Feldkamp]

There have been a few short discussions of "tractive effort curves" in the past on this forum. Check them out for some more info. Basically you need to create a vehicle speed vs. thrust curve for the selected motor and gear reduction. Gearing will shift the thrust curve along 1/x "iso-power" lines. In other words, gearing will not change how much power output is generated at any given operating condition, only how much torque is generated at the output. This means that multi or infinite gearing will only provide you with a few benefits with a motor with curves like you've posted above:

(1) Improved thrust the lower end of your vehicle speed range while also extending upper operating speed (assuming you have the power to overcome the vehicle drag at the higher vehicle speeds).
(2) Depending on your mechanical efficiency (this is CVT-type dependent) and control system, you can potentially get some minor gains in efficiency under low throttle and cruise conditions. The reality of a racing drive cycle, however, means that these effects probably won't be seen.

All of that said, the motor shown in your post is pretty darn small in terms of power for FSAE. Like the other guys have said, take a look at either an accel sim or a lapsim like OptimumG to come up with vehicle targets for your design. It's usually a poor design path to start with "what's available" instead of "what's right" for the job.

-Kirk

[Jay Lawrence]

Artur,

If you are stuck with this engine type, are you able to use more than one and do a hub drive setup? I imagine, with such low power, that this motor is quite small and may be able to be packaged neatly. That would give you 22kW (RWD) or 44kW (4WD), which is plenty. Obviously that's not the end of the story, but it's something to consider.

[Z]

Originally posted by arturyukio:
We are a new team and we are going to design and fabricate an electric formula style car.

Artur, Cheetah E-Racing Unifei.
Artur,

Your problem is not power.

Your problem is not torque.

Your problem is not even CVT vs gearbox ...

Your real problem is that you are a first year team ... of engineering students!!!

I imagine that some 17 year old boys who have left school and are working in the local garage could knock up a simple electric car (10" wheels, simple frame and suspension, your batteries and motor...), weighing about 200kg, all in a matter of a few weeks. This car would have a motor-to-wheel reduction of ~8:1 (possibly single stage chain, or maybe double stage gear and/or chain).

Thus equipped, it could squeal its rear tyres from 0 to ~30kph (work it out), and then go on to 90+kph with gradually reducing acceleration. Finished early so lots of time for testing and driver training. More than enough performance to do really well in any FSAE comp.

But, because you are engineering students, it seems you want to design an unnecessarily complicated transmission, sitting in a grossly overweight car (why!? *), that probably won't be completed until just before the competition.

So, no testing. No driver training. Lots of DNFs at comp...

Your choice ...

Z

( * BTW, Cheetahs are the lightweights of the big cats, so maybe look at the Uni-Cal-Berkeley car (at recent Lincoln comp, IIRC) for inspiration... http://fsae.com/groupee_common/emoticons/icon_smile.gif)

[Artur Yukio]

...All of that said, the motor shown in your post is pretty darn small in terms of power for FSAE. Like the other guys have said, take a look at either an accel sim or a lapsim like OptimumG to come up with vehicle targets for your design. It's usually a poor design path to start with "what's available" instead of "what's right" for the job.

-Kirk

Thanks for the reply, Kirk.
We are doing what we can do to make this project possible, but there is the point that you said.




If you are stuck with this engine type, are you able to use more than one and do a hub drive setup? I imagine, with such low power, that this motor is quite small and may be able to be packaged neatly. That would give you 22kW (RWD) or 44kW (4WD), which is plenty. Obviously that's not the end of the story, but it's something to consider.

Jay

UoW FSAE '07-'09

Thanks for the reply, Jay.

We already e-mailed our sponsor asking about the possibility of sending one more motor. But I think they will send more than one. It would really help us in terms of power. Thanks for the tip though!




Artur,

Your problem is not power.

Your problem is not torque.

Your problem is not even CVT vs gearbox ...

Your real problem is that you are a first year team ... of engineering students!!!

I imagine that some 17 year old boys who have left school and are working in the local garage could knock up a simple electric car (10" wheels, simple frame and suspension, your batteries and motor...), weighing about 200kg, all in a matter of a few weeks. This car would have a motor-to-wheel reduction of ~8:1 (possibly single stage chain, or maybe double stage gear and/or chain).

Thus equipped, it could squeal its rear tyres from 0 to ~30kph (work it out), and then go on to 90+kph with gradually reducing acceleration. Finished early so lots of time for testing and driver training. More than enough performance to do really well in any FSAE comp.

But, because you are engineering students, it seems you want to design an unnecessarily complicated transmission, sitting in a grossly overweight car (why!? *), that probably won't be completed until just before the competition.

So, no testing. No driver training. Lots of DNFs at comp...

Your choice ...

Z

( * BTW, Cheetahs are the lightweights of the big cats, so maybe look at the Uni-Cal-Berkeley car (at recent Lincoln comp, IIRC) for inspiration... Smile)

Thanks for the reply, Z.

That may be our problem, I don't know. What I know is that our car is wheighing too much (estimative). You know, yesterday I made another estimative of our car based only on our project and it wheighed about 230 kg, thats a lot lighter than our first guess. With that wheight we could keep it simple and implement a single stage chain with a motor-to-wheel reduction of ~9,7 to make the car accelerating through the 75m of the track and get a ~89 km/h top speed. That would be reasonable for our first car.

The big thing is, we depend mostly on sponsorship and partnership programs to get our resources. Just an example: we project our frame to have some tubes with a diameter of 1,0" and than we send the project to our sponsor and they just say "No can do, take these 1,5" and deal with it", hypothetically. Stuff like that add a lot of wheight to the car and thats why I can't neglect the fact that if I stick to the single stage chain with a motor-to-wheel reduction of ~9,7 the car may not even move due to the extra wheight. Like Kirk pointed a few posts above:


It's usually a poor design path to start with "what's available" instead of "what's right" for the job.

-Kirk


About the car beeing ready just before the competition is our biggest fear, thanks for the heads up. We are working hard to not let this happen. I just took a look at the Uni-Cal-Berkeley car, and if you said about the same that I saw they nailed it, only 218 kg, it really inspirates us, thanks again! http://fsae.com/groupee_common/emoticons/icon_smile.gif

[Kirk Feldkamp]

Originally posted by Artur Yukio:
I just took a look at the Uni-Cal-Berkeley car, and if you said about the same that I saw they nailed it, only 218 kg, it really inspirates us, thanks again! http://fsae.com/groupee_common/emoticons/icon_smile.gif

No no, 125kg. http://fsae.com/groupee_common/emoticons/icon_wink.gif

-Kirk

[Artur Yukio]

Originally posted by Kirk Feldkamp:

No no, 125kg. http://fsae.com/groupee_common/emoticons/icon_wink.gif

-Kirk

Just saw it, the B13. I was looking at their last year car haha, that`s awesome. 125 kg is pretty light. http://fsae.com/groupee_common/emoticons/icon_smile.gif

[Warpspeed]

This is interesting...
http://imageshack.us/a/img27/2166/nxav.jpg

Now those curves are for a three phase induction motor that is rated to run at 60 volts nominal through some kind of variable frequency drive.

The motor designers have published these performance curves which are no doubt valid.
But you don't have to run the motor that way if it is safe at 9,000 rpm as they claim that it is.

The thing to understand here, is what those curves are actually telling us.
Torque is dependent on the current (ampere turns) through the windings, which is limited by safe temperature rise.
Up to 3,000 rpm we are current and temperature rise limited, and the ac voltage fed to the motor will be kept roughly proportional to the motor drive frequency.
So as the motor speed increases from 0 to 3000 rpm, the motor ac drive voltage, and the motor ac drive frequency are increased together to produce constant (max) current, and constant (max) torque.

What happens at 3,000 rpm is that the motor voltage reaches 60 volts and cannot rise any further, because that is all the voltage there is to run from.
So we get constant motor current and constant torque (temperature limited) up to 3,000 rpm.

Above 3,000 rpm, the motor drive frequency can be increased much further, but motor drive current falls off, because of the increasing inductive reactance of the motor windings.
So above 3,000 rpm we have constant voltage and a constant power output.

That is what determines the shape of the above published motor curves, and all three phase induction motors have that same characteristic.

The interesting thing about it is that you can quite safely keep increasing the motor voltage above 60 volts, above 3,000 rpm, as long as you don't exceed the maximum rated motor current.
If the motor is truly safe to operate at 9,000 rpm, there is no reason why you cannot run the motor up to 180 volts, still at full rated current and get THREE times the rated power from it without overheating.

Basically the flat part of the torque curve will then extend to 9,000 rpm, and the power keeps increasing linearly up to 9,000 rpm. You just need a 180 volt power source, and a correctly programmed VFD capable of 300 amps to run the motor which is all quite straightforward.

It is probably an ideal motor, because you can run it from an off the shelf commercial VFD, and stay well below the 300 volt limit, and you can probably coax 66 Kw out of it for short bursts if your cooling system is really well developed and you have some reliable system to monitor motor temperature.

That should relieve some of the anxiety you have about the transmission and the available power.
I suggest you do a bit of research into induction motors and VFDs it will all become much clearer.

[Z]

Artur,

I agree with what Tony (Warpspeed) said above. More power from the motor is simply a matter of pushing the electrons in harder at higher revs (ie. use more volts).

But, as has been pointed out countless times on this forum, FSAE cars only have the right pedal flat to the floor for ~20% of the lap, at most. Most of the lap is done at quite low power. (Hint: corner speed wins the races, and high corner speed means less braking before the corner, which means less energy wasted through the race, which means smaller battery pack needed, etc., etc...)

The Cal-Bekeley car was about 125kg mainly because it was of very simple design (ie. no carbonfibre or titanium needed, just steel tube frame, small 8" wheels, etc.). Its IC engine weighed ~25kg, so swap with your 25kg motor, add your 60kg batteries + 10kg controller, etc., and you are at ~195kg. Fitting 10" wheels, because of better availability (?), might then push you slightly above 200kg.

Then with 70Nm at the motor and, say, 8:1 reduction, you have 560Nm at the wheels (less some small losses). For wheel radius = ~0.225m (typical of 10" tyres) this gives you 2.5kN (~250kg) thrust. With 50:50 weight distribution this will probably spin the rear tyres. But with a bit more rear weight (say ~60%) the rear tyres should stick (so less chance of spinning out on corners). So, with a lightweight driver you have about 1G forward acceleration whenever below ~30 kph and the right foot is on the floor.

IMO this sort of simple and light car with your existing motor, built early, is a better all round option than chasing more power from bigger motor, or using multiple motors, which inevitably leads to an upward spiral of more weight, bigger battery pack, bigger brakes, +++ http://fsae.com/groupee_common/emoticons/icon_frown.gif

Once you have such a car built and running, then you can try Phase 2 modifications, like more volts to push those electrons through faster... http://fsae.com/groupee_common/emoticons/icon_smile.gif

Z