[rejuvenile]

I wonder what the efficiency numbers are. The retrofit will add weight to the car, but won't remove any in the form of a lighter engine. There is also no regenerative breaking, and presumably no way to disengage the engine to run purely on electric power. I also wonder, speculatively, if the power assist will interact degenerately with the existing automatic transmission--perhaps running the car in a less than optimally efficient gear.

[tocomment]

Actually I don't see why they can't put regenerative braking in. The motors double as generators and they already have a connection running to the battery. And the motors are right next to the brakes so they could easily detect when the brakes come on.

Anyone know?

[pdonis]

To do regenerative braking the motors have to know when you're braking. That means hooking into the car's controls to detect when the brake pedal is pressed. It also means preventing the normal mechanical brakes from engaging while regenerative braking is being done. And it means being able to engage the mechanical brakes if the batteries can't accept incoming energy fast enough. All this is a highly non-trivial controls problem, which requires close integration with the car's control computers. I don't see how that can be done with a retrofit.

[rcvassallo]

You're making the problem harder than it needs to be.

Knowing when the car is braking is trivial - the vast majority have a dedicated brake switch wire.

Why do you want to prevent the normal brakes from engaging when regen braking is happening? Since the electric motors are on the rear wheels, braking traction will be severely limited. The majority of braking needs to happen up front due to dynamic weight transfer.

Instead, use a simple brake balance controllers to reduce the mechanic braking effort at the rear and make up for it with the regenerative braking. Brake balance controllers are readily available in the aftermarket, so this is a solved problem.

[pdonis]

> Knowing when the car is braking is trivial - the vast majority have a dedicated brake switch wire.

Yes, that would work, provided it didn't affect the signal going to the vehicle's own computer. (It doesn't seem like it should, but I spent some time as an automotive engineer, and one thing I observed is that the car will do lots of things you don't expect.)

[Edit: after seeing another comment below I'm not sure the switch would be enough; you need to know how hard the brake pedal is being pressed so you know how much braking the driver is requesting and can adjust the amount of regen accordingly. But there are ways to do that as well, as the commenter notes.]

> Why do you want to prevent the normal brakes from engaging when regen braking is happening?

Because the normal brakes prevent regenerative braking (regen) from happening. Both processes convert the vehicle's kinetic energy into something else--regen into electricity, normal braking into heat. You can't convert the same kinetic energy into two things at once, so any energy converted into heat by normal braking is unavailable for regen.

> The majority of braking needs to happen up front due to dynamic weight transfer.

Which will limit how much regen you can get, since it can only withdraw energy if it is actually slowing the wheel down. It may still be worthwhile to do regen, but not nearly as much as if you could do it on all four wheels.

> Instead, use a simple brake balance controllers to reduce the mechanic braking effort at the rear and make up for it with the regenerative braking.

Are these controllers dynamic? That is, can they adjust the balance depending on whether regen is happening or not? There will be circumstances when regen can't happen, or can't take all of the energy available (for example, if the battery is close to full or unable to accept charge fast enough).

If the controllers are dynamic, then I agree you could put together a solution this way. But you would still need to do a lot of tuning of the controls. One thing I would be worried about if I were doing this is being able to get everything tuned smoothly without having access to the code and calibrations in the vehicle's own computers.

[tocomment]

Yeah, or if nothing else install a button for the driver to engage regenerative braking when he knows he's going down a hill or coasting to a stop. Have the regenerative braking turn off if the driver touches the real brakes.

It seems like you'd still get 50% of the benefit and have no potential to interfere with the brakes.

[pdonis]

> Yeah, or if nothing else install a button for the driver to engage regenerative braking when he knows he's going down a hill or coasting to a stop. Have the regenerative braking turn off if the driver touches the real brakes.

Just a button wouldn't work; how would the system know how much braking to apply? You would need something with variable input. Even then it would require considerable practice, I think, on the driver's part; it would be something like trying to control braking with the parking brake while the vehicle is in motion (which I have had to do on one occasion when my regular brakes failed--it's not at all easy).

[jack-r-abbit]

The way the video describes the controller in the trunk, I believe it is not interfaced directly to the car. I kind of assumed it uses an accelerometer to notice the car is moving forward and engages the DC motors. If this is true, I would imagine that the same accelerometer could notice that the car is braking.

[rcvassallo]

Not sure why the trunk mount is relevant for interfacing with the car. The brake switch signal activates your brake lights, those are typically at the rear of the car :)

Admittedly, for a universal application you want to interface with as little of the car as possible.

That said, in the US all cars made after 1995 comply with certain On-Board Diagnostics protocols (OBD-II) so you can interface with several standard systems.

One such sensor available in OBDII is the VSS or the Vehicle Speed Sensor. Check out a device like the Scan-Gauge[1] to get an idea of other sensor that can be interfaced with universally. It would be trivial to run some wire under the carpet and plug into the OBDII port to give the hybrid system access to all kinds of engine & sensor info.

[1] http://www.scangauge.com/products/scangaugeii/

[jack-r-abbit]

> Not sure why the trunk mount is relevant for interfacing with the car.

I wasn't saying that. Let me rephrase: The way the video describes the controller (the one in the trunk)...

[ars]

> I would imagine that the same accelerometer could notice that the car is braking.

It would not be able to tell the difference between braking and simply slowing down, for example going uphill - or worse, driving in water - hitting the brakes when going through standing water can be disastrous.

As for the accelerometer detecting moving forward it has some nasty feedback going on there, since it detects its own motion. Plus imagine going downhill and the read engines think you hit the gas so they cause you to go even faster.

That said, it should be possible to interface with the engine there are plenty of signals that can be used (throttle, rpm, etc).

[jack-r-abbit]

> it should be possible to interface with the engine

Agreed. I was simply saying that from what I got from the video, it is not interfacing with the vehicle. Maybe it should be but it seems that is not what they are building. I'll leave it to the people that have been working on it for 4+ years to make that decision as they certainly know a whole lot more than I do.

[ars]

The switch wire is hardly enough unless you want to suddenly slow down for a light press of the brakes.

But teeing off of the brake line with the pressure sensor is hardly impossible.

[rocky1138]

We don't have numbers, so this is all speculation. Still, I think we can agree it's a neat idea that definitely deserves exploration.

With regards tot he transmission question: my car (and many others) allow me to select the gear even though it has an automatic transmission. In my car it's called Autostick, other cars it will be something else; sportshift maybe.

In this way, I can ensure the gear will be optimal for gasoline consumption.

Also, perhaps in the future, a model can be made to incorporate regenerative breaking.

[polshaw]

>perhaps running the car in a less than optimally efficient gear.

A properly designed solution would enable the opposite; keeping the car in an optimal gear for longer. A hybrid works best when the batteries run where they have the greatest advantage against the fuel.

[tocomment]

Also I don't see why you can't put the engine in neutral and drive with these motors (I guess you'd need a separate throttle though?)

[adjwilli]

Yeah, they say it gets between 50 and 100% better gas milage, but are those "savings" spent at the electrical outlet?

[altcognito]

http://www.scientificamerican.com/article.cfm?id=electric-ca...

Depending on the efficiency of the implementation I'm sure (as well whether or not this is a valid study at all), but it seems that electric cars are able to do it significantly cheaper than gas. So those savings are probably not spend at the electrical outlet.