Formula E

[autogyro]

Ok, I've found something, it's obviously a press-release inspired article so it only mentions the good parts. But I think it may help if Autogyro would point out the problems in this thing so we get the point a bit better:

The transmission has a "fairly conventional" twin shaft layout and uses two lubrication pumps - one mechanical, one electric - and a dual clutch for seamless gear changing. It could also be modified to become a 4-speed.

http://images.gizmag.com/inline/antonov ... sion-2.png
[..]

So why haven't we seen three (or more) speed transmissions in electric vehicles before? The idea goes against conventional thinking regarding electric drive systems and Paul believes that engineers engaged in EV development have thus far kept a narrow focus. Perhaps no-one realized the significant difference a multi-speed transmission can make ...

Source: http://www.gizmag.com/antonov-3-speed-t ... -ev/19088/

So they claim that a dual clutch setup allows them to do seamless gear changing.

15 percent increase in efficiency.
Here we go again.
2 percent is about the maximum increase using a multi speed conventional gearbox you can hope for over a single speed electric power train with the best motors.
This one uses energy to drive an electric oil pump and a hydraulic pump as well as two shafts.
I would sooner use a 1932 Wilson pre-selector box.
In direct top the Wilson had no losses.
That would give you much smoother shifts using band slip and would have less torque loss.
The Wilson would also last more than 2000 miles.
With modern materials it would be possible to reduce the weight of the Wilson cast iron casing to around the weight of this three speed sledgehammer attempt.
The Wilson won a few races pre-war in the ERA before they replaced it with a lay shaft box and started losing.
I hope Hewland have a better gear train in the FE, we shall see.
My guess is that the teams will go back to a single speed (no shift) gearbox after a few races.

[Andres125sx]

Somehow you have to slow it down with your electronic control system to match the output rpm.
Any ideas?

All of my brushless controllers for my rc planes have a built in brake to stop the prop when you use a folding prop, so it can fold instead of freewheeling. I think even RC cars have brakes without phisical brakes or generator mode, just the controller, so I don´t think that must be a problem for automotive industry

Brilliant if you want to stop the electric motor rotating.
No use at all if you want to accurately control the armature rpm just by using a variable electrical resistance.
Even a slight electrical load will stop the armature almost instantly.
If you do that to the input of a lay shaft or a epicyclic gearbox, when you connect the output it stands a very good chance of blowing the gearbox casing in half.

I´m not an expert here, I was just trying to say if a $20 controller can brake and also keep constant rpm (governor mode for rc helicopters), I can´t think managing motor rpm may be a problem for a car

It´s just a software problem, electric motors can manage rpm as they want, specially without load


Anycase, you could open a new thread to talk about gearboxes for EVs, this discussion has become quite off-topic

[Tim.Wright]

I think the discussion is relevant.

I'm still waiting on a decent answer, without handwaving, about why a layshaft gearbox is not suited to an electric drivetrain.

If anything, you have much more torque control possibility on an electric motor so the problems jerk should be able to be reduced to a level even better than what is achievable with an IC motor.

Though its not a trivial control problem, hence why it appears that it has not been solved yet.

[andylaurence]

If an electric motor really does have so little inertia that it's practically stopped by the time the gear is disengaged, then it's ideal for a dog 'box. They rely upon a difference in speed so that the dogs can locate and engage and lower inertia means the shock load is lower and the effect on the user much lower. My dog 'box is quite happy selecting gears all day just as long as the load is taken off the dogs during the shift. I suppose that's why the FE cars have a Hewland dog 'box rather than some vapourware supported by myths and anecdotes from the early 20th century

[NTS]

15 percent increase in efficiency.
Here we go again.
2 percent is about the maximum increase using a multi speed conventional gearbox you can hope for over a single speed electric power train with the best motors.

Sure, I immediately believe you, hence my comment about it being a press release. But my question was not whether they will gain 15% or 2% efficiency. My question was why you think this gearbox they've made would be impossible to make for an EV? Since this looks like a gearbox that works for EVs, claims to be smooth due to dual-clutch setup and does not explode immediately because otherwise they wouldn't sell it.

[NTS]

If an electric motor really does have so little inertia that it's practically stopped by the time the gear is disengaged, then it's ideal for a dog 'box. They rely upon a difference in speed so that the dogs can locate and engage and lower inertia means the shock load is lower and the effect on the user much lower.

If you operate a dog box the right way the driver would not feel much of a bump right? I've never driven a dog box, but I think a KZ shifter kart is similar? Gearchanges in that are quite smooth and very fast, without clutch, just by letting go of the accelerator a little bit to unload the box.

However the gearbox I linked to above claims to use a dual-clutch system for smoothness. Does that mean it allows the electric motor some time to adjust the RPM during the engaging and dis-engaging of the clutches?

[andylaurence]

Yep, a shifter kart is a dog 'box. They're great when you're going at a pace, but they're not so good pootling round town, although every biker since forever has coped just fine and there are ways to mitigate the problem with pneumatic actuation, which can avoid shifting off-throttle, cut the power for you and blip the throttle on the downshift.

The reason dual clutch gearboxes exist is because manufacturers didn't like the dog 'box because it is jerky as there's no smooth transition from one gear to the next thanks to inertia in the engine and an immediate change of gear ratio. They started by automating manual gearboxes, but they felt too slow as they were based on syncromesh systems to be smoother, so they switched to a dual clutch system that ensures the gear is pre-selected and the transition from one gear to another is purely by clutch actuation.

[autogyro]

If an electric motor really does have so little inertia that it's practically stopped by the time the gear is disengaged, then it's ideal for a dog 'box. They rely upon a difference in speed so that the dogs can locate and engage and lower inertia means the shock load is lower and the effect on the user much lower. My dog 'box is quite happy selecting gears all day just as long as the load is taken off the dogs during the shift. I suppose that's why the FE cars have a Hewland dog 'box rather than some vapourware supported by myths and anecdotes from the early 20th century

Nothing wrong with a manual dog box for IC engines Andy.
I have even said that they can be shifted faster and with better input to output matching than any other layshaft box if the driver knows what he is doing.
They also allow for the lightest and simplest single lay gear gear sets that saves weight and improves packaging.
The Wilson pre-selector however is not vapourware or a myth.
It is simply a mechanical manually selected epicyclic gearbox that won races.
It used friction bands in much the same although more efficient way to shift gears that were pre-selected than the dual clutch lay shaft system you all keep raving on about.
Let me know when you have driven an electric race car with a dog box Andy.
I would love to hear of how you get on with one.

[autogyro]

If an electric motor really does have so little inertia that it's practically stopped by the time the gear is disengaged, then it's ideal for a dog 'box. They rely upon a difference in speed so that the dogs can locate and engage and lower inertia means the shock load is lower and the effect on the user much lower. My dog 'box is quite happy selecting gears all day just as long as the load is taken off the dogs during the shift. I suppose that's why the FE cars have a Hewland dog 'box rather than some vapourware supported by myths and anecdotes from the early 20th century

Nothing wrong with a manual dog box for IC engines Andy.
I have even said that they can be shifted faster and with better input to output matching than any other layshaft box if the driver knows what he is doing.
They also allow for the lightest and simplest single lay gear gear sets that saves weight and improves packaging.
The Wilson pre-selector however is not vapourware or a myth.
It is simply a mechanical manually selected epicyclic gearbox that won races.
It used friction bands in much the same although more efficient way to shift gears that were pre-selected than the dual clutch lay shaft system you all keep raving on about.
Let me know when you have driven an electric race car with a dog box Andy.
I would love to hear of how you get on with one.
Oh yes and what do you think would be the result of connecting a stationary electric motor to a gear set rotating at road speed using a dog engagement?
Of course Hewland use an electronic system to control shaft speeds.
It remains to be seen how well they do it.
It is a lot more difficult than you think.
Even then the extra energy used plus the gear set weight and torque losses will negate the potential 2 percent efficiency saving over a single speed electric power train.
So where is the gain?

[andylaurence]

Oh yes and what do you think would be the result of connecting a stationary electric motor to a gear set rotating at road speed using a dog engagement?

My point was that if the electric motor slows so quickly that it stops during the shift, then its inertia must be practically nil and the impact of moving the dogs into place would be small, despite the large difference in rotational speed. In reality, wouldn't the inertia be roughly similar to that of a clutched gear change with a dog 'box? That's something that millions of bikers do every day and their dogs tend to last quite well.

[FW17]

The Tesla batteries can be could be charged in a hour with 120 KW supercharger

A Formula E could be charged in about 40 min with unknown power

Can the power be increased to such a level that the battery can be charged in a min with all sorts of gizmos like dry ice cooling of batteries etc?

120 kw sounds very unsafe considering an avg home is about 15 kw, what would be level required for a 1 min charge? what would be safe? considering racing is always about being insane as f1 refuel rig pumped 10 kgs/sec during regulated times and 30 kgs/sec during the unregulated times

[autogyro]

Oh yes and what do you think would be the result of connecting a stationary electric motor to a gear set rotating at road speed using a dog engagement?

My point was that if the electric motor slows so quickly that it stops during the shift, then its inertia must be practically nil and the impact of moving the dogs into place would be small, despite the large difference in rotational speed. In reality, wouldn't the inertia be roughly similar to that of a clutched gear change with a dog 'box? That's something that millions of bikers do every day and their dogs tend to last quite well.

I seem to remember that IC engines have a flywheel and a rotating clutch assembly.
Electric motors have an armature that is slowed by induction when under load.
It depends on how you electrically control the motor of course.

You are also well aware that the skill changing gear with manually operated dogs is to march the input and output rpm as near as possible, dogs only take up slight differences.
Your racing results show that you have mastered the skills needed.
A stationary motor and a gear set rotating at probably 2 to 3000 rpm is not a small difference.
Now try to reproduce those skills using electronic control as well as car control.
Let me know how you get on or more important, how long the gearbox lasts.

[andylaurence]

I seem to remember that IC engines have a flywheel and a rotating clutch assembly.

Isn't the flywheel attached to the engine and the clutch housing with only the clutch plate attached to the gearbox? Thus when the clutch pedal is pressed, there's not much on the end of the gearbox shaft.

You are also well aware that the skill changing gear with manually operated dogs is to march the input and output rpm as near as possible, dogs only take up slight differences.
Your racing results show that you have mastered the skills needed.
A stationary motor and a gear set rotating at probably 2 to 3000 rpm is not a small difference.

The dogs don't take up a difference in speed, they either lock the two together or they don't. When the dogs are engaged, they receive an impact, which is the inertia of the motor. The inertia increases with the speed delta. The skill in changing gear is to reduce that speed delta and to shift the dogs as quickly as possible so you don't chip the corners off the dogs by not mating cleanly. If you matched the speeds exactly, the dogs would be unlikely to engage as they would need perfect alignment. Dog engagement gearboxes therefore rely upon a difference in speed.

Now try to reproduce those skills using electronic control as well as car control.
Let me know how you get on or more important, how long the gearbox lasts.

Will do. I'm building a paddle-shift system at the moment. Hopefully in time for next season, so shall we revisit this in 12 months?

[autogyro]

The dogs do take up a slight difference in speed Andy but only a small difference.
You contradicted yourself when you said that if the input and output were at the same rpm the dogs would be unlikely to engage.
This only happens if the fingers on the dogs line up with the ridges on the gears that space the dog engagement grooves.
This is more likely to happen if the input and output are at exactly the same rpm but it can also happen if the speeds are not the same.
On the original dog engage systems you would then hold the gear lever against the detent in towards engagement and apply power or brake the dog would then engage with a bang.
Hewland call their box for FE a dog box but I am betting it uses some kind of sprung syncro equaliser in an attempt to smooth the shifts.
The problem with any EV gearbox based on lay shaft gear sets is the difficulty of controlling the motor rpm/torque during the gear shifts.
Using two clutches and lay shafts and slipping the friction material to smooth things simply wastes energy and negates the need for a gearbox. (such boxes are just acceptable in IC vehicles although not really for hybrids)
In FE the power trains will remain as they are with the Hewland gearbox so long as they have sufficient reliability.
There is no competition as yet because this year it is a controlled formula like F1.
If things open up as promised the teams have a year to contact me for a more efficient gear system.

[mzso]

Anyone knows of a good summary of the technical parameters of the cars?