During my research on the power steering sysytem aboard our toys I came across this little tidbit and thought I might share.

http://www.nsxprime.com/FAQ/Technical/eps.htm

 

Nice info.

 

Thanks. The pictures explain a lot. It's basically a feedback servo loop. The fact that it has a load cell and a rotary encoder has given me an idea of why my steering wheel spazzs out from time to time. I plan on tearing down my rack, if I ever get a chance, so I can troubleshoot the sensor.

 

If you look in the EPS section of the helm you can see a schematic of the torque sensor and how it integrates into the whole system. I don't see a rotary encoder - only a pair of coils and what I'm going to assume is a pin/slug/solenoid (not actually doing any work) that moves around with steering input and perturbs the inductance of both coils. There are 3 wires on the torque sensor connector which the schematic shows going straight to the coils.

My leading hypothesis is that the shaft running through the torque sensor is a torsion bar which the EPS system measures deflection on to read driver input torque. The sensor does double duty as a steering center indicator when the two inductance levels match the center programmed into the EPS.

For the curious: fooling an inductance based signal is quite complicated but likely possible. When looking up systems that can adjust the feel of the S2000/NSX EPS I found one NSX race car that had a two position adjustment. If the EPS is using a torsion bar to measure torque then disabling EPS to "increase feedback or feel" may not be productive. You'll increase steering effort but introduce slop with the torsion bar which is now deflecting more than normal operation because there is no correction on the other side of it to take that deflection out as it shows up.

 

Great info. Maybe popping he cover off the eps module and checking a few datasheets will narrow down exactly how it works.

 

I'm trying to get my hands on a broken or used EPS torque sensor so I can see what's going on inside; please PM if anyone has one for cheap. I'd really like to figure out if does use a torsion bar or not - this is pretty easy to tell. Hold one end of the shaft static and put a torque wrench on the other side. There should be a very high torque with very low deflection if it is not a torsion bar. Be very careful not to screw up the splines or you'll be buying a new part.

If the sensor does use a torsion bar there might be a "cheap" and "easy" way to increase steering effort and road feel while removing slop: heat treat the shaft to increase stiffness. A decrease in deflection should be interpreted by the control module as a reduced steering effort with reduced assist to go with it. Cheap and easy are in quotes because a detailed and proper procedure has to be made for this - DO NOT take a torch to the shaft and then quench it - you could turn it into a piece of metal that's so brittle it'll break apart while installed in the car.

 

I'm trying to figure out exactly what causes my steering wheel to vibrate back and forth 1/4" rapidly while I'm sitting still.

 

My first thought for something like that is a bad signal to the torque sensor and in my mind the most likely suspect would be bad connector or internally loose connection or something like that. Does it happen when you drive? Does it happen at idle? Does it happen when the system is engaged but the engine is off? You can get the EPS to come on with the engine off if you roll the car around 100 feet with the ignition on; it worked on my 2002 a while ago because that's as far as I got down the hill when seeing what the effort was like with it off.

Do you have assist on the steering? Bad power drivers can twitch too but I would expect if the drivers were damaged it would have drastically reduced assist. Power input problems could cause that too but I would again expect drastically reduced assist.

Have you tried sharing the sensor/controllers with other cars to see if it fixes the problem or follows them? That'd be a good way to isolate out which component it is (control box, sensor, or vehicle wiring).

 

It only happens first start sitting still and more often near full lock, and only once, if at all. I think something is corroded or loose as you said, and I'm leaning more towards a contact in a sensor, or a loose sensor part of some kind. It seems like the feed back loop gets into a mechanical resonance loop and it just shakes....maybe it thinks it should be returning to center, but can't because I'm not helping. hmmmmm. I'll have to post a video I have. I've only had one other person with the same problem. It works perfectly after that and never acts up while moving.

 

Is it normal for the S to "road steer"

I've noticed this since i've had the car from the very beginning. I feel that it's normal. any differences in the road and the car likes to follow them very hard.