Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Multi-gear EVs are probably the future

This site may earn commission on affiliate links.
I'm a big fan of the single-speed gearbox that Tesla and other vehicles have. But, costs will have to be cut. Manufacturing a motor to run at over 15,000RPM is expensive as the rotor requires improved balancing and cooling. A lower-cost design would have a two-speed manual gearbox with a less powerful, slower, cheaper motor. The first speed would cover almost all vehicle speeds up to 70mph improving 0-60 acceleration further. The second speed would cover speeds up to 155mph or beyond, as necessary.

For the vast majority of driving the vehicle would remain in the 1st speed. No more city shifting, eliminating one of the complaints around manual boxes. Another bonus will be the car can be locked in the 2nd gear all the way down to a stop as the motor RPM can drop completely to zero. The disadvantage of leaving the motor in the 2nd gear is that acceleration would be reduced from a standstill and efficiency would also likely be reduced.

I know the prototype Roadster had difficulties with a two-speed gearbox system but that was due to poor engineering from Borg-Werner. I don't think it's impossible to design a gearbox for an electric motor, but engineers need to think differently.

Manual cars use a friction clutch to buffer the connection between the gearbox and wheels. The clutch acts as a way to slip the engine power, which allows gears to be engaged easily. Electric motors don't require a clutch - at all. By simply open-circuiting the motor (turning off all power switches in the motor driver) the motor can be made to free wheel ("virtual" clutch down.) The gearbox can then switch gears, using dog teeth as in a standard manual gearbox to synchronise the now open circuit motor and connected wheel speed. Finally, the motor circuit is re-energised with the motor controller matching wheel RPM quickly. (The effect on the power consumption/meter would be to see a slight regen of power while the motor wants to lag, then power would increase to the original level, plus/minus efficiency changes. Improvements in motor firmware could reduce this change to no noticeable drop in power by matching the speed quicker.)

Shifting would probably be similar to a DSG gearbox. Automatic mode would almost certainly be available (D mode.) Budget EVs could even have a "clutch" pedal and manual shifter... the clutch being an electronic input only.

Thoughts? I doubt Tesla will do this for Model 3 as they'd like to keep complexity down. But perhaps we'll see this for even further cost reduced models.
 
Last edited:
  • Helpful
Reactions: Texas
I'm a big fan of the single-speed gearbox that Tesla and other vehicles have. But, costs will have to be cut. Manufacturing a motor to run at over 15,000RPM is expensive as the rotor requires improved balancing and cooling. A lower-cost design would have a two-speed manual gearbox with a less powerful, slower, cheaper motor. The first speed would cover almost all vehicle speeds up to 70mph improving 0-60 acceleration further. The second speed would cover speeds up to 155mph or beyond, as necessary.

For the vast majority of driving the vehicle would remain in the 1st speed. No more city shifting, eliminating one of the complaints around manual boxes. Another bonus will be the car can be locked in the 2nd gear all the way down to a stop as the motor RPM can drop completely to zero. The disadvantage of leaving the motor in the 2nd gear is that acceleration would be reduced from a standstill and efficiency would also likely be reduced.

I know the prototype Roadster had difficulties with a two-speed gearbox system but that was due to poor engineering from Borg-Werner. I don't think it's impossible to design a gearbox for an electric motor, but engineers need to think differently.

Manual cars use a friction clutch to buffer the connection between the gearbox and wheels. The clutch acts as a way to slip the engine power, which allows gears to be engaged easily. Electric motors don't require a clutch - at all. By simply open-circuiting the motor (turning off all power switches in the motor) the motor can be made to free wheel ("virtual" clutch down.) The gearbox can then switch gears, using dog teeth as in a standard manual gearbox to synchronise the now open circuit motor and connected wheel speed. Finally, the motor circuit is re-energised with the motor controller matching wheel RPM quickly. (The effect on the power consumption/meter would be to see a slight regen of power while the motor wants to lag, then power would increase to the original level, plus/minus efficiency changes. Improvements in motor firmware could reduce this change to no noticeable drop in power by matching the speed quicker.)

Shifting would probably be similar to a DSG gearbox. Budget EVs could even have a clutch pedal and manual shifter...

Thoughts? I doubt Tesla will do this for Model 3 as they'd like to keep complexity down. But perhaps we'll see this for even further cost reduced models.

My first thought is when you're thinking mass volume, lower prices cars: Why would a cheaper vehicle have to have a top speed of in excess of 120 or even 110 mph?

My next though is why a two- og three-geared classic gearbox. Why not a variable transmission (Continuously variable transmission - Wikipedia, the free encyclopedia) like in the production model Lexus 450h and in some Audis?
 
In this case, the speeds were an example. Adjust to taste; perhaps one gear for 0-50, and additional gears to fill in the rest, giving you city driving and motorway driving with a single gear shift only. The high speed motor used in Model S costs thousands of dollars to make. You only need to look in a motor manufacturers' catalogue to look for a 240kW (~320hp) continuous motor (that's P85D continuous power limit):
300 HP Electric Motors (AC Motors)

The water cooling of the rotor does help Tesla lower costs. And of course the motor is not an ultra-high reliability motor (as much as Tesla go for quality, it only needs to last ~500,000 miles @ ~60mph average = ~8,300 hours operation.) So costs will be lower. But the rear motor easily costs $5,000 to make. It's not a cheap part.
 
Last edited:
If anything I see a system similar to the ring gear setup with multiple traction motors of varying windings similar to how the Volt (Gen 1 car, Gen 2 is different) is setup. It's a complicated system to explain but think of it in the same way the D cars utilize two electric motors of varying size to increase efficiency. Several videos on the Voltec power train attempt to explain this.

One thing I wish Tesla would apply is the ability to regen brake even with a full battery. The current cars will only use full friction brakes until the SOC drops low enough for the power to have somewhere to go. GM counters this by separating the two motors and having one regenerate while the other uses that power to apply force in a braking direction. Power is effectively used to slow the car, that power coming from the regenerating motor. A while back I had asked the SC guys if changing the charge level limit before departing would then allow the car to regen immediately as my commute starts with several long downhill sections. They had no information as they'd never been asked the question.
 
My next though is why a two- og three-geared classic gearbox. Why not a variable transmission (Continuously variable transmission - Wikipedia, the free encyclopedia) like in the production model Lexus 450h and in some Audis?

Well, I've never driven a CVT so can't really compare. But, I thought that the complaints with CVTs were that they were "laggy" in performance, taking time to change ratio limiting acceleration, and that they had poorer reliability because most of them use a rubber belt to transfer power. Not to mention, the cost is much higher, and reducing cost will be important. There's no real need to use a CVT as the torque band of an electric motor is flat, unlike an ICE; the two-speed setup would only cause a slight momentary dip in power when switching gears. The CVT wouldn't have the dip in power, but would bring other problems with it.

- - - Updated - - -

One thing I wish Tesla would apply is the ability to regen brake even with a full battery. The current cars will only use full friction brakes until the SOC drops low enough for the power to have somewhere to go. GM counters this by separating the two motors and having one regenerate while the other uses that power to apply force in a braking direction. Power is effectively used to slow the car, that power coming from the regenerating motor. A while back I had asked the SC guys if changing the charge level limit before departing would then allow the car to regen immediately as my commute starts with several long downhill sections. They had no information as they'd never been asked the question.

I had a thread on an idea using a resistor to dump the brake power when the battery is full. But now that since Tesla uses electromechanical braking with full authority control on all newer cars they could just simply create a fake regen system which used the brakes to slow the vehicle when the battery is full. I'd be surprised if we don't see this in a year or so. They would only need to match the deceleration to create the equivalent to regen. It would be nice to have the power meter go grey when using virtual regen, to let the driver know that the car is still decelerating.
 
The high speed motor used in Model S costs thousands of dollars to make. You only need to look in a motor manufacturers' catalogue to look for a 240kW (~320hp) continuous motor (that's P85D continuous power limit):
300 HP Electric Motors (AC Motors)

That's not even close to a valid comparison, those motors are heavy duty industrial motors over built to run for years on end, and weigh thousands of pounds. The fact remains that all current EV's, including low cost models, use a single speed gear reduction, and they don't do it because it costs more than using a multi speed gear box.
 
That's not even close to a valid comparison, those motors are heavy duty industrial motors over built to run for years on end, and weigh thousands of pounds. The fact remains that all current EV's, including low cost models, use a single speed gear reduction, and they don't do it because it costs more than using a multi speed gear box.

Here's an interesting document from the US government produced in '99.
http://www.transportation.anl.gov/pdfs/HV/14.pdf

See page 35. The cost of an AC-induction motor in a 40kW cont/67kW peak application is estimated at approximately $500 of which appoximately $300 is material cost. Material cost is likely to scale linearly with motor output power, so we can easily arrive at a material cost of $1,800 for a 240kW motor plus $200 in assembly costs. My $5,000 estimate was a wild-ass guess but was within an order of magnitude. (That web store link was, admittedly, too far off for any reasonable estimates.) Let's not forget since that document was produced the price of copper has almost quadrupled and AC induction motors are >80% copper:

GraphEngine.ashx?z=f&gf=110563.USD.lb&dr=max.png


The cost of a conventional transmission for an ICE car is given as 5% of the cost of a $21k car. Thats roughly $1,000 for a standard automatic transmission. A manual transmission would cost even less. DSG-style computer controlled shifting would cost a little more though the lack of a true clutch would reduce costs somewhat.
 
Here's an interesting document from the US government produced in '99.
http://www.transportation.anl.gov/pdfs/HV/14.pdf

See page 35. The cost of an AC-induction motor in a 40kW cont/67kW peak application is estimated at approximately $500 of which appoximately $300 is material cost. Material cost is likely to scale linearly with motor output power, so we can easily arrive at a material cost of $1,800 for a 240kW motor plus $200 in assembly costs. My $5,000 estimate was a wild-ass guess but was within an order of magnitude. (That web link was, admittedly, too far off for any reasonable estimates.) Let's not forget since that document was produced the price of copper has almost quadrupled and AC induction motors are 80% copper:

GraphEngine.ashx?z=f&gf=110563.USD.lb&dr=max.png


The cost of a conventional transmission for an ICE car is given as 5% of the cost of a $21k car. Thats roughly $1,000 for a standard automatic transmission. A manual transmission would cost even less. DSG-style computer controlled shifting would cost a little more though the lack of a true clutch would reduce costs somewhat.

And that standard automatic transmission is capable of handling 600 Nm of torque?
(That's a rhetorical question, with the answer "No").

1) Performance transmissions are pricey.
2) Multi-speed transmissions add complexity
3) Cost of power electronics is gradually decreasing
4) Cost of batteries is decreasing
5) Density of batteries is increasing

Basically, it would be a lot of trouble for little value.

Maybe with a next-gen trackable Roadster, or an electric pick-up, Tesla will _need_ the performance-efficiency that a multi-geared transmission would provide, but until then, they stick with a simple reduction gear.
 
...fake regen system which used the brakes to slow the vehicle when the battery is full. I'd be surprised if we don't see this in a year or so. They would only need to match the deceleration to create the equivalent to regen. It would be nice to have the power meter go grey when using virtual regen, to let the driver know that the car is still decelerating.

Ok but what would be the point of that as it does nothing to decrease wear on braking components. It's not the perception of regenerative braking I'm after its the lack of any brake dust or brake wear except under hard braking. Neither of our Volts have any brake dust whatsoever on their wheels. The pads look like they haven't been touched despite over 40,000 on one of them.
 
Tesla has achieved a two speed transmission by having dual motors with different gear ratios. The rear is geared low for acceleration and the front is geared for highway and above speeds. They can dynamically switch between either or both motors.
 
If anything I see a system similar to the ring gear setup with multiple traction motors of varying windings similar to how the Volt (Gen 1 car, Gen 2 is different) is setup. It's a complicated system to explain but think of it in the same way the D cars utilize two electric motors of varying size to increase efficiency. Several videos on the Voltec power train attempt to explain this.

One thing I wish Tesla would apply is the ability to regen brake even with a full battery. The current cars will only use full friction brakes until the SOC drops low enough for the power to have somewhere to go. GM counters this by separating the two motors and having one regenerate while the other uses that power to apply force in a braking direction. Power is effectively used to slow the car, that power coming from the regenerating motor. A while back I had asked the SC guys if changing the charge level limit before departing would then allow the car to regen immediately as my commute starts with several long downhill sections. They had no information as they'd never been asked the question.

Since you can set the slider in 1% charge increments, you should be able to set it to something eg. 96% that allows you to capture all or most of the regen on the way down. I can't understand how they would not be able to figure this out on the spot.
 
Tesla has achieved a two speed transmission by having dual motors with different gear ratios. The rear is geared low for acceleration and the front is geared for highway and above speeds. They can dynamically switch between either or both motors.

Although it makes a lot of sense, Tesla haven't made this statement as far as I am aware. Has this been proven by anyone yet?
 
Thoughts? I doubt Tesla will do this for Model 3 as they'd like to keep complexity down. But perhaps we'll see this for even further cost reduced models.
Tesla may do it with a truck and not with a 'car'. They might be thinking differently, but they have a history and transmission almost ended them.

What a gearbox can do for an EV is increase low-speed power by reducing maximum speed. Not just torque but actual delivered power to the wheels.
Whatever motor you have, you will increase low-speed power by higher reduction ratio.
And you will also increase highway efficiency by having a reduction ratio that keeps motor in maximum power window at 'standard highway' speeds.

CVT is a bad idea because it offers no gains over fixed gear setup, but it reduces efficiency.

Multispeed transmision can be had also in other ways - P85D as a whole has multispeed gearbox already as front motor spins at different RPM as the rear motor.
When power is needed at low speeds, higher-geard rear motor takes the lead. When power is needed at higher speeds, lower-gear front motor takes the lead.
It is just that gearboxes in P85D are very similar i.e. reduction ratios are close. In a truck they would be further apart to improve continous low-speed power (i.e. same power at lower currents and less overheating).
 
Nope, material cost scales with material. Tesla has found ways to get very high power to weight ratios, higher than any other OEM, so they are using less material per kW.

Tesla has been able to do what no other car manufacturer has done and has made a great electric car with long range and high performance.

What they have not done is innovated on individual parts - the innovation has been putting all the pieces together in a manner not done before. They are using commodity battery cells, commodity chassis design (albeit aluminium, but standard technology nonetheless) and, importantly, their namesake's motor design conceived nearly 130 years ago. Of course the motor is quite compact but you can only make motors so compact and you can only remove so much material.

Here's a picture of the rotor test/assembly line: http://www.trbimg.com/img-54dc1c1b/...ont-model-s-production-pict-003/1150/1150x647

The entire rotor is a near-solid core of copper! How much does that cost to make?

Also note the notches taken out of each rotor to balance them. It looks like the rotors are balanced by hand.

I realise I made an error previously. The P85D's peak continuous power output is 240kW after a few laps around the track. This power limit is actually a function of the front and rear motors and the battery pack cooling so the motor continuous power limit will be closer to 180kW for the rear and 60kW for the front (the distribution may be different, but it is more likely the power will be biased to the rear.)

Tesla will want to keep Model 3 complexity down to a minimum so it will most likely be single speed w/ possibly AWD being an option.
 
Personally, I think it'll go the other way - I think the typical EV of the future will follow in the footsteps of the concept cars and supercars of today - with four separate motors each with fixed gearing (though possibly with two different ratios like Tesla told us about and probably implemented.)

The benefits of multispeed gearing seem unlikely to balance the cost and weight penalty - especially if motor technology continues to expand the efficiency range as it has in the past, while the benefits of going to four motors are substantial - better traction in starting and stopping, active handling, etc.
Walter
 
Tesla has achieved a two speed transmission by having dual motors with different gear ratios. The rear is geared low for acceleration and the front is geared for highway and above speeds. They can dynamically switch between either or both motors.

I'm not certain about this, but I don't know enough about the physics behind it to address this. Still, this does mean unless you have some kind of clutching mechanism, each motor has to be designed for the highest speed the vehicle will encounter, which makes the motors more expensive.

- - - Updated - - -

Ok but what would be the point of that as it does nothing to decrease wear on braking components. It's not the perception of regenerative braking I'm after its the lack of any brake dust or brake wear except under hard braking. Neither of our Volts have any brake dust whatsoever on their wheels. The pads look like they haven't been touched despite over 40,000 on one of them.
It would only be used when the pack was full or in cold weather when regen is unavailable or limited. 95% of the time it would not be used so you'd still get good brake life, but no more surprises when regen doesn't work on a cold morning.