On gearboxes, electric motors and ICE's.

[Wyvern_64]

I'm going to college for engineering and one thing I may focus on for my career is automotive engineering, either for performance/sport, or for efficiency/alternative fuel. This summer I've been thinking of ways to possibly make internal combustion engines (ICE) or electric motors more efficient.

It was actually GT5 that made me think of this, in the tuning menu you see a power curve, like this:

So at a certain speed, any given engine has an optimal power output, and then it drops off on either side. So I was wondering if it would increase overall performance if an engine could stay at the revs that it gets peak power output (~5000 RPM in the picture) and let the transmission deal with changing speed. That way, instead of the engine revving up and down that power curve, it stays at the top all the time. Would this also increase efficiency for regular daily driving? If an engine is running at peak power output, it would require less gas to do whatever the driver wants. Or if that's not as efficient, maybe the car could calculate what speed the engine needs to be going to be most efficient, taking into consideration it's power curve, what the driver is telling it to do, and how fast it uses fuel at different revs.

This could be done through a gearbox with, instead of 6 or 7 set forward speeds or gear ratios, but with every gear ratio within a certain range. What I mean is, if a regular transmission has first gear ratio of 1:1 and seventh gear ratio of 1:2, then make this one have everything between 1:1 and 1:2, so 1:1.5, 1:1.6732, etc. It seems weird but I think it could be done. Maybe not infinite possible ratios in a certain range, but at least hundreds of smaller increments.

Back to the ICE power curve, what about electric motors? I read about the new R8 e-tron, how it has tons of torque at low speed but can't get over 100mph, and this being caused by the fact that electric motors are much more efficient at lower speeds, that they produce loads of low speed torque but almost no high speed power. So could this idea work with an electric car as well? Hold the motors at the low speed at which they're most efficient and slide the gear ratio up smoothly to accelerate the car.

I know this is a weird idea so if you're having trouble visualizing it, think of this: In a car with a normal manual transmission accelerating from a stop, the engine starts in first gear with low revs. Then it revs up as the car gains speed, then you gear up to second and the engine restarts at low revs again. This repeats until you get to speed. Say the car is normally in 5th gear at 100mph. What if it had 10 or 20 gears to get up to 100mph? Again, from a stop, the car would start at low revs on first gear, then rev up until it got to a suitable speed to shift into second. But with more gears closer together, the engine would rev up less between gears. As the number of gears increases, the less the engine has to change in speed. Infinite number of gears, engine can stay at a constant speed while the gearbox does all the work in accelerating the car.

I'm not posting this asking whether or not this is actually possible (I have a few ideas I have yet to build). I'd like to know what you think about the idea of it: if it could actually increase both performance on the track and efficiency in daily driving, for both ICE's and electric motors. If anyone has a more knowledgeable or experienced opinion than mine, please share! 💡 I'd love to hear what everyone thinks about this!

 

[Bopop4]

They already have that, it's called CVT. (Continuously Variable Transmission.)

Image 2 cones as the gears.

As the position of the belt changes on the two cones, so does the transmission ratio.

small to big = slow speed.
medium to medium, medium speed.
big to small = fast speed.

 

[Wyvern_64]

Right. Yeah. Not sure why I've never heard of that before. Why aren't these more widely used?

That 2 cones and a belt was exactly what I was thinking. Would a system with no belts possibly be better? I had a different idea, like a cone shaped gear with teeth spiraling up like a screw.

 

[Bopop4]

Well they can be noisy, and I'm totally sure about this, but I also think that they can be more inefficient in certain applications.

The Nissan Altima for example, has a CVT.
But the computer is programmed to make you feel like you're driving a 6-speed car.
So in semi-manual mode, instead of having the rpm stay more or less the same, the belt is moved into preset postions to simulate gears. Pretty cool huh?

And I don't see how a no-belt system would work, you can't have the cones touching.

 

[Wyvern_64]

One cone, and another regular gear that hovers on and between the rows spiraling teeth, with a controlled height.

I don't really see the point of simulating gears when the point of a CVT is to get rid of them altogether.

What kind of transmission are they using in the Volt, or Tesla models?

 

[Bopop4]

Since the little gear would be moving, it would need a system of pulleys and tensioners.
It would be more complicated than the CVT.
And yeah, but I think the car runs a normal CVT program for normal driving, and then for sport driving you can change it with the semi-manual mode.

And I don't actually know, google could probably answer that though. 👍

 

[Wyvern_64]

The Tesla Roadster and Model S use a single speed fixed gear, Chevy Volt has a CVT.

I'm reading about all these benefits of a CVT but other than the noise and friction on the belt I can't see why this isn't more widely used. Especially with electric cars.

 

[Beeblebrox237]

I think it's not widely used because people don't like it. It's not that it's a bad idea, it's brilliant, but it is sometimes prone to motorboating and doesn't feel as fun. Motorboating is when the engine stays at a constant rpm for a long period of time, and can get annoying. Even though it is more efficient to do it that way, comfort comes first. Also, shifting gears is fun. With a CVT, that element is lost, so you don't have those surges of power. Again, it's not as fast or efficient, but it's more fun.

 

[Wyvern_64]

Ah, makes sense. What about motorsport? Comfort isn't exactly the number one concern in Le Mans.

 

[Bopop4]

I don't think that the current CVT technology could stand the power of those cars.

 

[Keef]

CVTs don't seem to be very shock-resistant which is definitely a problem in racing. They're exceedingly common on machinery like mills and lathes for precise speed control and they work well because the spindle load is one-directional and stays pretty constant while the machine is running.

 

[hawkeye122]

I don't think that the current CVT technology could stand the power of those cars.

Exactly. In an Altima, the belt would only have to change speeds from 0-120(ish).

In something like a Rebellion-Toyota-Lotus, you'd have to have that belt stretch and contract from 0-215+.

 

[Wyvern_64]

So a CVT-type transmission that doesn't use belts could be good, if it has low friction even at high torque... that would also solve the shock resistance problem.

Too bad I don't have Solidworks and a CNC mill here. Make a CVT using mill that uses a CVT. 💡

 

[hawkeye122]

Make a belt that barely accommodates full speed, and then have a empty gear that stretches the belt at lower speeds... I wish I could describe what i'm seeing... Have a standard CVT, but have the belt also go through a wheel that isnt connected to anything, above and inbetween the cones. Have it able to change altitudes, thus making sure you aren't stretching the belt, but just maintaining ideal tension.

 

[homeforsummer]

It doesn't really need to be said as the discussion has gone in the correct direction anyway, but strictly you're speaking about gearing rather than efficiency.

In essence, what you're trying to do is make an inefficient device (an internal combustion engine) use the efficiency it has more wisely, and make an efficient device (an electric motor) spread that efficiency over a greater area by giving it gears.

Efficiency

Here is a good list of the efficiency of various devices. You'll see that combustion engines are between 10-50% efficient (in motor vehicles, 50% is very optimistic indeed - figures I've heard in the past are between 20-30%) and electric motors from 70-90% (again, figures I've heard for cars seem to be around 90%).

That means in a combustion engine, only 20-30% of the energy put into combusting the fuel is actually used for pushing pistons around. The rest is wasted as heat and noise. In an electric motor, the vast majority of the electricity put into the motor results in motion.

Gearing

Aside from the relative irrelevance of top speed in day-to-day driving (other than in combustion engine cars, where it's an indication of how low the engine can rev - and therefore how fuel-efficient and quiet it is at a given speed), the reason infinite gearing won't work isn't as simple as you're making it sound.

It'd be great to keep an engine in its power and torque curves all day long... but that's exactly what CVTs (as mentioned already) and to some extent, regular gears do already.

The issue is that aerodynamic resistance quadrouples every time you double the speed (i.e. at 80mph there's 4x the aero drag there is at 40mph), so even if you could keep an engine spinning right where it was developing the most power and torque, there would always come a point where the engine could no longer overcome the bodywork's aerodynamic resistance (not to mention rolling resistance of the tyres).

That's also detrimental to fuel efficiency, so even if you're using the power and torque to achieve a top speed, the engine's actual ability to move you down the road for a given amount of fuel is diminishing because you have to work it harder to maintain a certain velocity.

Gearing in EVs

The same applies to electric cars, whatever you do with the gearing. Most EVs are limited electronically (rather than limited by the motor's maximum RPM) because at high speeds the aero drag and current going to the motor rapidly drains the battery. EV top speeds therefore are often limited by range, not by gearing.

That said, some companies are working on manual transmission EVs. The original Tesla Roadster was a two-speed car, but they found it unreliable so went back to a single-speed reduction gear. Morgan is working on the "Plus E", which has a five-speed manual gearbox.

The real reason we don't have manual EVs though is that it's largely pointless. One of the benefits of EVs is a huge reduction in moving parts and a seamless drivetrain, compared to combustion vehicles. Throwing in a gearbox immediately makes it more complicated (i.e. more prone to failure) and removes that seamless power...

 

[Wyvern_64]

...and electric motors from 70-90% [efficiency]

Gearing in EVs

So electric motors are just as energy efficient at all speeds?

Most EVs are limited electronically (rather than limited by the motor's maximum RPM) because at high speeds the aero drag and current going to the motor rapidly drains the battery.

After reading this, I think I spoke wrong about increasing 'efficiency' with a CVT-like transmission in an EV. In an electric motor (correct me if I'm wrong), it can deliver a constant power. So at low revs there is lower current through the battery so using P=IV, it can have a higher voltage which translates to more torque. But at higher revs, the current is higher so the voltage must be proportionally lower, therefore less torque at high speeds, and since the aero drag is higher, the lower torque on the wheels means it has a lower top speed.

So instead of this way, put in a CVT-like transmission. This will keep the motors running at a constant speed, and therefore a constant current and voltage. Lower current drains the batteries slower. Even at higher speeds, the torque on the wheels from the motor will be the same. The only difference is higher aero drag which just takes away from the torque generated by the motors and reduces acceleration, until a top speed where aero drag and torque from the motor are equal. So the point of introducing a more complicated transmission into an EV would be to keep the current through the batteries low at all speeds.

 

[homeforsummer]

So electric motors are just as energy efficient at all speeds?

Ish. They do have pick-up and drop-off like any device with an energy input and output, but it's completely different from that of an internal combustion engine, and that relationship between energy input and rate of work is far more efficient with an electric motor than it is with internal combustion.

With internal combustion, you're putting in fuel, oxygen and igniting it with a spark (or compression, as in a diesel). Ideally, you'd want the energy contained in that fuel to result in work, i.e. pushing a piston down, and it does, but with surprising inefficiency.

The vast majority of the energy produced by combustion is wasted as heat and noise, and only a relatively small proportion does the task you want of it. Engines are certainly getting better with modern levels of tolerances, low-viscosity oils, better combustion etc, but there's still a lot of waste energy.

With an electric motor, you put electricity in, and it does work. The vast majority of the electricity you're using goes towards making the car move.

It's why, although a 100-mile range may sound unimpressive from an electric car, it's actually very impressive - energy stored in a lithium-ion battery isn't anywhere near as energy dense as gasoline (hence the need for big, heavy batteries), but it still manages to move a car 100 miles down the road. That's why energy density is such a big deal for batteries of the future. Give a battery the energy density of gasoline (between 6-10 times more than its current level) and you have an electric car that does 6-10 times greater range than it does at the moment...

After reading this, I think I spoke wrong about increasing 'efficiency' with a CVT-like transmission in an EV. In an electric motor (correct me if I'm wrong), it can deliver a constant power. So at low revs there is lower current through the battery so using P=IV, it can have a higher voltage which translates to more torque. But at higher revs, the current is higher so the voltage must be proportionally lower, therefore less torque at high speeds, and since the aero drag is higher, the lower torque on the wheels means it has a lower top speed.

This graph is particularly fascinating - it's the power curve of a Nissan Leaf. You'll note in its single gear, there's a gradual build-up* of power to around 30mph, at which point the motor generates essentially maximum power in a plateau all the way to its 90mph limited top speed, at which point it drops off completely.

The only EV I've driven to its top speed is a Renault Twizy, which is limited to 50mph. It's not a powerful car - 17bhp - but like the Leaf it seems to develop its power and torque fairly consistently. Acceleration dies away the faster you go, but that's a side-effect of rolling resistance, air resistance and the weight of the vehicle with only 17bhp to push it along. When it gets to 50mph, the power dies, but a bit like a rev-limiter you can keep your foot pinned and it'll just (quietly and smoothly) butt up against the limiter. I suspect the Twizy's power curve is similar to the Leaf's, only lower down...

This, I think, is a typical torque curve of an AC motor as found on the Leaf (couldn't find an actual dyno torque curve). I'm not an engineer so I don't understand it completely, but it appears to show a peak, trough, another peak and then a sharp cliff-face drop-off.

So instead of this way, put in a CVT-like transmission. This will keep the motors running at a constant speed, and therefore a constant current and voltage. Lower current drains the batteries slower. Even at higher speeds, the torque on the wheels from the motor will be the same. The only difference is higher aero drag which just takes away from the torque generated by the motors and reduces acceleration, until a top speed where aero drag and torque from the motor are equal. So the point of introducing a more complicated transmission into an EV would be to keep the current through the batteries low at all speeds.

Yeah, I know what you're getting at, and although I'm not clued-up enough to know how viable it would be, it does sound like something that could work.

That it's not been done yet suggests there could be a problem with it - perhaps weight, perhaps complexity, perhaps it wouldn't get on with an electric motor's regenerative capabilities - i.e. with a non-direct drive to the wheels, the electric motor could no longer function as a generator when there's no current running to it and the wheels are driving the motor instead.

I wonder if the benefits would be minimal, but the negatives (i.e. no regen) would be significant. Regen doesn't put much energy back into a battery, but it's generally enough to supply several extra miles over the course of a journey.

*Two things to note on this. First, we've all heard stories about electric motors producing maximum torque from zero revs. This is still the case, but in many electric cars that torque is artificially limited to help prevent the car frying its tyres everywhere from a standstill. That said, the incredibly smooth nature of electric motors means that tyre-frying is less likely than with the sharp, uneven bursts of power you get from a combustion engine.

The second thing to note: In the forum thread where I found this, the owner suspects that the power and torque up to 30mph are conservative, as he thinks the wheels were spinning slightly on the dyno (lots of torque + low rolling resistance tyres) so the dyno wasn't registering the whole hit until 30mph.

 

[hawkeye122]

The only EV I've driven to its top speed is a Renault Twizy, which is limited to 50mph. It's not a powerful car - 17bhp - but like the Leaf it seems to develop its power and torque fairly consistently. Acceleration dies away the faster you go, but that's a side-effect of rolling resistance, air resistance and the weight of the vehicle with only 17bhp to push it along. When it gets to 50mph, the power dies, but a bit like a rev-limiter you can keep your foot pinned and it'll just (quietly and smoothly) butt up against the limiter. I suspect the Twizy's power curve is similar to the Leaf's, only lower down...

Why not have some sort of computer that, rather than having it bounce off the limiter (and waste energy) have a system where at full throttle, the engine gets to pick how it maintains it's top speed. Going up a hill? It gives enough power. Downhill? Less power. That would seem vastly more efficient than me trying to maintain 50, without hitting the limiter.

 

[homeforsummer]

Why not have some sort of computer that, rather than having it bounce off the limiter (and waste energy) have a system where at full throttle, the engine gets to pick how it maintains it's top speed. Going up a hill? It gives enough power. Downhill? Less power. That would seem vastly more efficient than me trying to maintain 50, without hitting the limiter.

Sorry, I didn't really make it clear. It doesn't just hold the car at 50, the current literally dies. Essentially it does what you're suggesting - down a hill it'll cruise down a few mph over 50 using no power (I assume motor resistance stops it doing more, but of course if gravity is driving the wheels then it tops up the battery, so it's win-win) and uphill it'll try and maintain 50 with the gas pedal pinned, though obviously there's a limit to how fast you can tackle steep hills in something with 17hp.