Torque sensor conditioning EPS

hello I'm needing some help I'm wanting to increase the power steering assist in my mums car who's struggling. I need a electronic devices to modify the torque sensor, has anyone done this before? It's a 5v sensor with two analogue outputs. 2.v volts at no torque and goes higher or lower depending on left or right. the second output is the mirror image of first. any ideas?

Why? Does she have you listed as sole beneficiary in her will? You could be doing something very unsafe to her car, since you have to come here and ask the question. Some makes of cars EPS systems are adjustable, but must be done by dealer mechanics, you should go to the dealership and not mess with this yourself.

Often the EPS units are weighted according to the vehicle speed. We used to disconnect the speed sensor cable and use a 555 timer to manually vary the EPS assistance making it think it was driving slowly for maximum assistance or fast for low assistance (track days). Not sure if this would work on your car but worth a look.

lol I'm not going to kill her. lightened power steering for disability aid is a world practice.
picbits yup that's what they do when they add new eps systems to older vehicles. In this instance it can't increase the assist down low.

How about a larger diameter steering wheel?

A simple mechanical solution.

JimB

yeah i did think about that or even a planetary gear setup up. she does have a spinner... a knob on the wheel... like buses. but i'd like to keep wheel size and steering ratio the same if i can. i cant look past fact the little tiny assist i want is already right there. somewhat annoying. there are people that make professional signal conditioners listed for the torque sensor pacifically, I'm waiting to hear back from them as I'm not sure it's actually what i need? does anyone know about these eddy current sensors. they're some type of hall effect thats micro controlled at the sensor to produce a analogue signal thats then read by ecu?

jimmy898 said:

does anyone know about these eddy current sensors.

Click to expand...

That sounds like it could be an LVDT sensor? They have two sets of coils and a moving armature.

In the centre position the outputs balance out, with equal and opposite voltages induced, but as the core moves the coupling is stronger to one coil than the other and a signal is produced, with the phase dependant on which direction the movement is in & amplitude dependant on the amount of movement.

More info:

[As the others say, I'd not advise messing with the steering system; it's a legal can of worms even it functions correctly].

A couple of op-amps would do the job. You just need a 2 V reference and a non-inverting amplifier. You don't need to know what is inside the sensor.

It would be a good idea to have a completely separate amplifier for each output. The design would be the same for each. It would also mean that if one were to fail, the other would still give steering control, as I doubt your mother would be strong enough to control the steering if the power assist forced the steering one way as hard as it could.

The big danger is if you get a failure that makes the power steering think that full torque is called for.

A speed over-ride, to bypass your home made circuit as soon as the speed is above about 10 kph, might be a good idea as well. That would also reduce the risk if your home-made circuit failed, as it would only be in use at slow speeds.

If the circuit failed and gave no output at all, power steering would be lost. However, in normal driving in small cars at speeds above 30 kph power steering really isn't needed.

It would also be a good idea to drive the car with the sensor disabled so that you can see what it's like. Make sure you test any solution where it does not matter if the steering does something odd.

I've fitted ABS to four vehicles. One got fitted wrong, with two pipes swapped over. I tested in a deserted car park, going at about 20 kph. When I lbraked hard and locked up the front wheels, the brake pedal dropped to the floor and one front wheel stayed locked with the pedal released. I had to let the pressure off with the bleed screw to move the car.

My point is that I tested it where the lack of ABS and the fact that the car was stuck afterwards wasn't a problem. I strongly recommend you do something similar if you modify the steering feedback.

(My first four vehicles didn't have power steering. On the largest of those, a 3 tonne van, I couldn't turn the steering with it stationary, and I had to get it to at least creep along to get the steering to move. There was less problem with the two vehicles that were less than 500 kg and only steered one wheel)

For a completely non-electronic solution, depending on what car it is, you may be able to fit smaller wheels with narrower tyres, which can reduce the steering load.

If you know anyone with a similar car, for a quick test not on the road you could fit space-saver spares on both front tyres. If there's not change in how easy it to turn the steering, this suggestion isn't for you. I would expect a big change, and normal tyres will be harder to turn.

The big wheels and low profile tyres are fashionable and may give a small handling improvement if being driven really hard, but the lower profile tyres are more expensive, give a less comfortable ride, make it easier to bend or the wheels or damage them on kerbs, give more rolling resistance and increase the steering loads.

Many cars are offered with a range of tyre sizes, and you should stick to wheel and tyre sizes that are in the range that comes from the car manufacturer. The whole range will usually have nearly identical diameters.

My second-hand car came with 225/45R17 tyres, and I changed to 205/55R16. There is even a "sporty" option of 225/40R18. Those sizes are listed in the handbook. The diameters are 634, 632 and 637 mm, so <1% different so there's no speedometer recalibration or anything.

haha great story Diver300. I'm loving the feedback and ideas guys.

Quick note on the safety issue... i'm not going to do anything dodgy unless i fully believe it's safe. they do this in the UK ect but no one does it in NZ and I'll get it certed for lightened power steer. I know the system, I am just not great at electronics.
It has two signals for that exact reason of not getting bad sensor information. for anything to fail that perfectly is really beyond comprehension.
The ecu compares them against each other to validate authenticity. otherwise goes into a reduce mode.

Back in the day we use to machine down torsion bars, you could still do it with eps till later models.
I do have background in vehicle modifying.

Thanks again guys, anyone else welcome throw some ideas about.

jimmy898 said:

haha great story Diver300. I'm loving the feedback and ideas guys.

Quick note on the safety issue... i'm not going to do anything dodgy unless i fully believe it's safe. they do this in the UK ect but no one does it in NZ and I'll get it certed for lightened power steer. I know the system, I am just not great at electronics.
It has two signals for that exact reason of not getting bad sensor information. for anything to fail that perfectly is really beyond comprehension.
The ecu compares them against each other to validate authenticity. otherwise goes into a reduce mode.

Back in the day we use to machine down torsion bars, you could still do it with eps till later models.
I do have background in vehicle modifying.

Thanks again guys, anyone else welcome throw some ideas about.

Click to expand...

Hi Jimmey898, I wanted the same but for different reason. Possibly a turn know to either provide more or less steering assistance for track use. I hookup a multimeater and the power that goes to EPS rack is about 5v but polarity gets reversed. So for example where 0 is center if you turn left the positive cable is red and negative is green, when turning right the red becomes negative and green is positive. This is on the electric motor side. But I think the best would be to modify the input signal that comes from the tq sensor. It's 3 wire interference sensor. Inside it has two coils top and bottom and aluminum ring inside that is sliding up and down on the steering shaft. When you turn the wheel the friction of tires will cause the torsion bar twist on the shaft forcing the inside ring either move up or down to different coil.

I thought maybe it's possible to modify the tq sensor signal to either trix the EPS module that the torsion ring is farther in the interference coil. I think the more interference in the tq sensor the more assistance is added.






Here is tq sensor that I cut out to expose the two inside coils.


Here is readout from multimeater when turning left or right on the motor side.