Drive-by-wire for dual carbs/throttle bodies???, would you run them?? Options

[Mueller]

A lot of newer cars do not have the traditional cable that goes from the gas pedal to the throttle body...instead the pedal has a potentiometer*, this in wired to the ECU which tells the throttle body how much to open, just because you have your foot buried to the floor, the throttle body might only be open 50% since the ECU is dictating what is going on.

Now a drive-by-wire throttle body could be wired directly to the pot. on the gas pedal to be 1:1 at all times......with a setup like this on dual carbs or individual throttle bodies, no need to worry about thermal expansion of motor and how that effects the linkage.


* (similar to a variable resitor which varies the voltage output depending on foot placement)

[jsteele22]

QUOTE (lapuwali @ Feb 28 2006, 04:09 PM)

QUOTE (fiid @ Feb 28 2006, 02:33 PM) I'd be interested in how the actuation is actually done - I would think that there would be a return spring and perhaps a solenoid to pull the throttle open with some kind of pwm coming from the ECU. Anyways - I'd mess with it... sounds cool.  I'd make sure it's REALLY stable before going anywhere with it though. Just open the bonnet on your wife's Legacy. Soob's have had throttle by wire for awhile now. My wife's Forester has it, too. Pretty easy to do traction control this way, too. Despite my pro-EFI stance, I'm strangely against drive-by-wire. I also don't like traction control, stability control, ABS, or any of that. And yes, I just read the GRM article on the subject.

What was the GRM article ?

About the actuation, the easiest would be a servo motor. The most familiar example is probably the servos in a model airplane/boat/etc, but servo motors can also be quite large. You send them a a PWM signal and they rotate to the corresponding position. PWM stands for Pulse Width Modulation. You sen a signal which is either low (0 volts) or high (5V, say) and the width of the plse determines the position. PWM signals are very easy to generate, and very immune to noise. For this example, the input is the variable resistor driven by the pedal. The feedback is the variable resistor on the throttle body (the TPS). These signals are scaled (if neccesary), compared, and if say the TPS isn't opened enough, the output signal is increased. The output would be a voltage which is then converted to a pulse width. That's the basic idea. As with any negative-feedback control system you have to tune a few parameters to get acceptably fast response without running the risk of oscillations, overshoot, etc.

Less simple than a servo is a stepper motor. These are very common (take a look at Jameco.com) and popular with hobbyists : google it and you can learn a whole bunch. Basically the big advantage, and the big disadvantage, is that its all digital. The motor has several (usually 4 or 6) wires, and if you run current in one wire and out another, the motor will take one "step". Then you stop that, pick another pair of wires and run some current through, and it takes another step. And so on. A single step is typically 1.8 or 3.6 degrees and can be in either direction depending on the direction of the current. You can even be tricky and make it go half a step by driving two pairs of wire at the same time. So the whole problem comes down to choosing which pair of wires to drive and when, and for how long. All the kind of thing that microcontrollers are great at doing. The meager digital outputs of the microcontroller usually go to a chip called an "H-bridge" that can put out enough current to drive the motor. One possible disadvantage to stepper motors is that you don't know directly the position of the motor. But if you start from a known position (say, by trying to close the throttle plate by 360 degrees, so it is guaranteed to be closed) you can then count steps in each direction to know where you are. So in principle, you don't need any other kind of position sensor. But for a robust (safe) system, it's prolly a good idea. Even though its more work, having the system be all digital has its advantages. Any parameters that need to be tuned can be modified in software, rather than relying on physical components like resistors, which can change their value with variations in temperature, age, and grime. And you can email your parameters to someone across the country to try.