The Fully Computerized Electric Vehicle

[safe]

Computers are cheap and powerful. The fact that they are not part of the electric vehicle so far simply says that the industry is not mature yet, but eventually one would expect them to be used.

Here's how I would see them used:

First, the battey suppply is made up of cells. Depending on how you assemble the battery pack, in series or parallel, you can get any combination of volts and amps that you might need. Some battery chemistries are very sensitive to overdrain, low voltage, overcharge, and other various problems. If each cell were individually connected into a switching circuitboard and that board was then connected to a computer you could provide a whole range of services by exploiting the power of the computer.

1. Throttle control could be managed by the selection of how many and in what way the cells are combined to supply the needs of the motor.

2. All the critical cell management duties could be monitored by the computer.

3. Other things like measurement of the system (like all the usual amps / volts / watts) could be provided by the computer to a display.

The advantages would be:

1. 100% energy efficiency. (verses 95% for PWM controllers)

2. Programmable Powerband. (you could configure your electric vehicle for maximum range or power or something in between)

I could envison various preprogrammed powerband configurations that you could select from a menu before you rode. Maybe you might select "hillclimb" or "long range" or "top speed" or "fast cruise" or some other option.

 

[TylerDurden]

That would be one robust switching network.

How efficient could it possibly get?

 

[fechter]

I'm sure people have though about this before. From an engineering standpoint, you have to balance overall efficiency against cost and complexity. In order to have a battery switching scheme, the switches would need to be really big to handle the voltage and current. If you use FETs, they will have some resistance losses (and they are pricey). I think keeping the battey configuration fixed and just optimizing the controller is the most cost-effective approach.

The losses in a switcher will increase as the voltage difference between input and output increase. There is some merit to changing the battery configuration at low speeds, but I would guess a simple series/parallel (2:1) switch would be about as far as you would want to take it.

You could probably get more improvement by using a higher Q inductor, or better yet, use an ironless motor.

I like the computer idea. Look at RC controllers. Almost all of them are microcontroller based these days. The Tidal Force, Oxygen Lepton and Voloci controllers are also microcontroller based and very sophisticated.
Microcontrollers are pretty much dirt cheap these days, so it's only a matter of time before we see these used in most controllers.

With a microcontroller, you can eliminate a bunch of parts and have some very useful features, like automatic timing advance and sensorless control for brushless motors. Parameters like current limit, ramp up speed, throttle calibration, low voltage cutout, etc, can all be adjusted in software, rather than in hardware. All the functions of a DrainBrain and BMS could be integrated as well (the TF controller has features like this).

Too bad I suck at PIC programming....

 

[safe]

How efficient could it possibly get?

The only time a switch would need to trigger was when the throttle position changed. Previous postings that discussed relays suggest that a relay takes just a few watts a microsecond to trigger the switch.

The computer might draw some current... but a typical laptop battery can last for a few hours, so I don't think that amount of drain is significant.

Maybe 99.5%?

Mass production brings down the price, but you need a "universal solution" to get the quantities up high enough.

 

[safe]

Voloci controllers are also microcontroller based and very sophisticated.

Where they PC / Mac based?

That's what needs to happen... no more "custom computer solutions", but instead they need to link a typical computer to the system. A laptop PC seems like the ideal because you can just plug it in and then ride. Even take the laptop out and sit under a tree and play PC games! :wink:

 

[fechter]

These things are not computers (computers suck too much power anyway). I'm talking about a Atmel PIC or something like a Basic Stamp. More like what's in a Palm Pilot.

 

[TylerDurden]

How efficient could it possibly get?

The only time a switch would need to trigger was when the throttle position changed. Previous postings that discussed relays suggest that a relay takes just a few watts a microsecond to trigger the switch.

The computer might draw some current... but a typical laptop battery can last for a few hours, so I don't think that amount of drain is significant.

Maybe 99.5%?

Mass production brings down the price, but you need a "universal solution" to get the quantities up high enough.

You might need four relays per cell... more?

 

[BiGH]

I think that you could combine the two - much like car aftermarket car ECU's - have a custom setup which datalogs (as they are often easier to interface and less trouble with drivers, operating systems etc) and have a usb connection to either a pocket pc or a laptop - which could allow you to see historical logs and adjust parameters. much like when car's get dyno tuned with aftermarket fuel management.

ever since i found ebikes- i've been itching for a project like this. I'm not great with electronics, but good with software.

 

[Toorbough ULL-Zeveigh]

Yes, by all means look at the Tidalforce example. How everyone despises the fact that it's computerized & can't even so much as rebuild the internal battery pack with lithium cells as a consequence. You are correct, computerizing EV's is the way of the future, but it's prime benefit to the manufacturer is to lock out the end user.

http://endless-sphere.com/forums/viewtopic.php?t=586


Does this description from Teslamotors sound like what you're trying to emulate?

Power Electronics Module (PEM)
Most of the subsystems in the Tesla Roadster are completely electronic and under direct software control. But unlike all other cars, these systems are not a hodge-podge of independent systems — instead, they are designed as an integrated system, the way complex network and computer systems are designed today.

You’ll see the hub of this network every time you pop the trunk — the Power Electronics Module. When you shift gears or accelerate in the Tesla Roadster, the PEM translates your commands into precisely timed voltages, telling the motor to respond with the proper speed and direction of rotation. The PEM also controls motor torque, charging, and regenerative braking, and it monitors things like the voltage delivered by the ESS, the speed of rotation of the motor, and the temperatures of the motor and power electronics.


Oh, it also only has two gears, which is all any e-vehicle really needs because of the flatter power band of an electric motor.

 

[TylerDurden]

Give it all to research...

 

[Leeps]

I really really would like to build a motor controller with an imbedded microcontroller, i have a ton of pics lying around with a compiler and programmer ready to go. I just dont trust my programming skills when it comes to something that could end up killing me.

"Hmm the throttle doesnt respond anymore, must be a divide by 0, oh well theres only twenty miles of battery left. road trip, out of my way people"

Joe

 

[BiGH]

kill switch is ur friend! esp during testing.

 

[Miles]

I came across this:
http://www.embedded-control-europe.com/c_ece_knowhow/871/ECEfeb07p14.pdf

 

[Mathurin]

The advantages would be:

1. 100% energy efficiency. (verses 95% for PWM controllers)

Oh man, I LOL'ed...
The best humour is unintentional.

But, don't you recall this topic having been beaten to death not long ago?


Ah well, let's remember to not divide by zero while fishing!