Apparently the EPS Torque sensor in the S2000 is the same as in the Insight. It is a non-contacting sensor. This is very cool since most EPS designs use contacting sensors (potentiometers) for cost reasons. It is not clear from the reported source below if the sensor is inductive, capacitive, or some sort of RF feedback but it is definitely not a strain type; There would be no torsion bar (T-bar). From the drawings in the link below (not copied here) the only visible concern I can see is wear of a sliding diagonal sleeve.

I can also see adjusting the basic amount of power assist by increasing or decreasing T-bar rigidity if you didn't want to mess with replacing or modifying the controller (which is still a black box to me, aside from known error codes).

From http://www.insightcentral.net/eneps :

Electronic Construction

The EPS control unit is mounted inside the car on the right side bulkhead, underneath the dash. It receives input from
the vehicle speed sensor and torque sensor mounted on the steering pinion shaft.

The torque sensor is identical in construction to the unit on the S2000. The pinion shaft engages the pinion gear via a
torsion bar, which twists slightly when there is a high amount of steering resistance. The amount of twist is in
proportion to both the amount of resistance to wheel turning, and to the steering force applied. A pin on the torsion bar
engages a diagonal slot in the sensor core, which moves up or down depending on the amount of torsion bar twist, and
the direction of rotation. Two coils surrounding the core detect both the amount, and the direction of movement.

Using this information, the EPS control unit determines both the amount of steering assist required, and the direction. It
then supplies current to the motor for steering assist. The amount of assist is also modified in proportion to vehicle
speed to maintain good steering feel.

Torque Sensor

The torque sensor is a device to detect steering turning direction and read resistance. The sensing section of the
torque sensor consists of two coils and a core (slider). The steering input shaft and pinion gear are connected via a
torsion bar. The slider is engaged with the pinion gear in a way that it turns together with the pinion gear but can move
vertically. A guide pin is provided on the input shaft and the pin is in a slant groove on the slider.

When road resistance is low, the steering input shaft, pinion gear and slider turn together without the slider's vertical
movement.

When road resistance is high, the torsion bar twists and causes a difference of steering angle between the input shaft
and pinion gear. In other words, the turning angle of the guide pin and slider differ, and the guide pin forces the slider to
move upward or downward.

 

Interesting... Is there interest in reducing the amount of steering assistance to get a 'heavier' feel?

Ted

 

Very much so. I was going to crack open the controller and lookinto modifying the feedback/control loop, though this might be much easier.

Anyone know how easy it would be to modify this? I would think it would be a sealed unit, and that finding another torsion bar would be a LOT easier than it sounds. It would be relatively easy (electronically) to modify the signal, though.

BTW, that link is bad. Here's the right one:

http://www.insightcentral.net/eneps.html

 

I'll pay for it .

 

Bump.

 

I wonder if one coil is energized and the other is for pickup. The rods movement could alter the eddy (sp?) Currents through the bar between the two, resulting in the pickup coil sensing the change. Just an idea.