Fly By Wire - Page 2

 

If the dumbasses are using potentiometers, like we used to use for joysticks
on computers, there is little doubt that cheaper POTS as they are called,
soon send signal spikes not long after the first use.  Nowdays the "good'
computer joysticks (read that as flight controls or simulator wheels) use
high quality Optical sensors (like computer mice did before using infrared
light (the ones without balls in them) PLUS use a computer sampling circuit
board to 'normalize" the output from that before it goes into the receiving
device that is supposed to react to the signals.  I don't doubt these
sampling circuits could be under engineered for autos, heat or stress wise,
low voltage overvoltage etc.

In gaming, we use fly by wire.  so figures they can do it in real life too.

off throttle is probably 0 on throttle Full is probably anywhere from 25 to
1028 (or more I'm sure) steppings, the higher the more precise the pedal
movement to throttle movement reaction would be.  problem could be (my whole
post is somewhat speculation) that, like we have in "simulations or Gaming"
is that the older junker controller "spikes".  This is the untechnical term
that errant data is presented to the receiving unit...   thus at anywhere
except FULL throttle movement is sending a signal that the other end of the
device that applies throttle, sees as FULLthrottle.  this was a terrible
problem for many years gaming and Sim racing or flying with (sub 400.00
gaming controllers, and even high end ones have to be replaced).  A lot of
this was offset a little by slower computer computational times.  Meaning
the spike might last for say 1/2 second (extremely exaggerating the times so
it makes more sense here) but the computer didn't go "check" the input for
another 38/60ths of a second and the Spike was normalizing, then it didn't
affect as much.

I'm not sure how they have dealt with it in cars, I'm sure in jets and such
they have a processor that samples the signal many times a second and does
calculations based on an average for so many sampling times per second, like
an average, plus has to somehow fault tolerance the signal from "what it
should be, what it probably is, and what the hell is that?" then sends it to
the receiver.  I'd bet cars don't have all that built in as robustly.

the electronic controls (by wire controls) could & would fail safe some of
those days gone by antics we've all had, you know, when that tree grabbed
the throttle cable just right as you went by it, that wicked the throttle
wide open on you?  as it could check so see if it is even valid data to go
from half to full in under .002 seconds.  Plus the tree wouldn't really
directly affect the throttle in a mechanical sense like it did pulling the
cable.

 


... wondering if it would help to turn that around. Idle is 256 and full
throttle is ~0~... Kind of like air brakes, the default if the system
fails is to [stop]...

RMR...

 

default on airplanes is full throttle
cars "should" be closed throttle....
btw if you use an airplane motor on an
air boat then loose the throttle cable
things get exciting.... but that's another story

being a reformed jaguar driver I learned to downshift a lot
and plan on the xk120 brakes glowing & boiling fluid.

 

On 2/3/10 1:03 PM, PlowBoy wrote:

1st generation Prius sent to the USA used a pair of pots in tandem.
Don't have the specifics handy at the moment but was powered with
something like 5V but wired such that only 1V to 3V was expected output
from the wipers. Anything outside that range generated a fault. Safe
from shorts and opens which eventually occurred due to wear.

Current units use Hall-effect sensors which do not wear.


Don't know the resolution of the A/D converter which was used but if it
was 0 - 1023 (10 bits) then Toyota only allowed inputs from 100 - 700
and anything outside that generated a fault.

Don't know exactly how it handled the fault condition. Believe the
engine shut down during the fault condition but would run if the input
returned to legal range making the car difficult to drive. Plus a big
obnoxious flashing idiot light that required a scan tool to turn off.

 


C'mon Dean, you can't fix something that complex with hose clamps.
Sheesh.

Tim H
Trailside repairs designed while-u-wait