Inductive Torque sensing demystified

[fusername]

So I got bit by the "how the hell does this work" bug. An internet friend of mine has a prius steering column, which has a DC motor and a torque sensor in it. the ECU is AWOL. while he could find one in ajunkyard, it expects sensors his car he wants to put it in (early 70s bug) doesn't. WHat I would love to do is reverse engineer my own ECU for him that takes a reading from the built in torque sensor, and feeds a PWM motor controll to the assist motor.

so I expected some sort of strain guage, but nope, that's only early models (looks like a strain gauge in the pics atleast). the later model (his is 08) uses a flexing torque shaft with two inductive sensors on it ot see how hard he is flexing the steering column and controll the DC motor to assist that. THe attached image shows the sensor, there is no further information on the sensor in the PDF i took that shot from, but a google search for "prius power steering" will turn hte pdf up as the first result. For some reason I can't make a link to it, my computer is silly like that sometimes.

the inductive sensor is not layed out in any way I am familiar with and need someone here smarter than me to help explain it to me. I am a ME by schooling, but have been playing with electronics as a hobby for about 8 years now (not as impressive as it sounds) and am trying to mix the two more and more. So tell me, what does that sensor rings output look like? I am used to a sensor "probe" with a toothed ring spinning near it, not an sensor "ring" around a toothed piece. I assume the phsycis are quite different between the two, so I have no idea where to start googling for info.

IN SUMMARY, my question is: what do you call the type of inductive sensor shown in the attached file, and how does it work? Thanks a ton folks!

 

[fusername]

I found a better pdf, and I think I understand a bit better now, does not at all work like I expected, I'll post up some specific questions if no one here jumps in and explains it before my head is fully wrapped around it. It sees to be some kind of transformer with a variable core that causes the output voltage to change, but I don't quite get it just yet

I can't post links yet, so format this by adding the obvious missing w's
scribd.com/doc/83157930/EPS-Toyota-Background

 

[jim hardy]

Here's my best guess:

Think of that "transformer" as a variable inductor with a flux sensing coil.

If flux sensing coil is not allowed to carry significant current it reports what flux is present without affecting it. An unloaded transformer does the same thing if you think aboout it. Primary current establishes flux, the secondary is in presence of flux so gets a voltage induced in it.

The inductance is changed by the alignment of the teeth. Recall flux moves easily through iron but not so through air.
Aligning the teeth gives more iron area for flux to pass through, misaligning them gives less iron area.
So when they're aligned flux is higher.
Higher flux gives more induced voltage in detector coil.

Inductance is flux linkages per ampere
so they're indeed varying inductance by alignment of the teeth.
You would call that a variable reluctance sensor.
That's my take on it.

OPINION that you didnt ask for -
Personally i would not ride in and would certainly never own an automobile with such a contraption attached to my steering.
I think it's asking for trouble.
Be sure that gear is flimsy enoough that a ninety pound girl can break it when the motor controller goes berserk or the motor locks up. It's a computer so is capable of anything.

I don't ride Airbuses either.

My advice - get a '62 Chrysler dash and steering gear it's sooo StarWars.

photo credit to this guy
http://www.californiaclassix.com/archive/62_Newport_62C.html

 

[fusername]

that (allmost) makes sense. i think i get how it works. the only thing that still confuses me is why are there teeth under the secondary winding, the correction coil. wouldn't a solid core do the same thing?

for added info, the wiring harness for the sensor only has 4 conductors, so I assume it is two independent coils with their own windings. which is weird because one of diagrams implied 3 coils total. one primary and two separate secondarys, onw of which may even be center tapped giving 7 wires, but they are not there.

 

[jim hardy]

Indeed detection ring 2 appears to have teeth on left end only.
I think they intentionally leave out a lot of details.

fifth "wire" could be the metal frame.
With your ohm-meter check and see if you have two isolated coils, or if they're tied together, and whether either has a connection to metal frame.
It's helpful to make a truth-table of every combination of wires with ohm readings. One wire could serve double duty, acting as both return line for excitating coil and centertap of detecting coil.Often they'll use a center-tapped coil as shown in that training diagram as a differential signal, comparing one half to other. But if you don't find a centertap they did something else.
I haven't yet figured out what they mean by "VT1 and VT2" shown at right angles on that chart of volts vs torsion.

We need to read their patent.

uspto.gov and you'll need a TIFF file reader. I use the Alterna , USPTO links to it and it installs easily.

 

[fusername]

got two nice pics from a toyota info pdf.
I took a shot at the patents, oof hard to search. going to have my friend see if there isn't a patent number on it somewhere. let's see what he gets back to me and what the multimeter tells.

 

[jim hardy]

check
US6536293
US6000491
US6843142

and for some shop instructions on calibrating the darn thing:
http://www.google.com/url?sa=t&rct=...c4-JirhL5jZ9T_WDg&sig2=cS3CR_ZSi_kQKe2jHHUGVQ

good luck !

old jim

 

[jim hardy]

That third one is the torque sensor. Leave off the "US" when you search.

It's more complex than i had imagined. They measure phase shift of both coils and take difference. It is not clear to me what part is done in the steering column and what part is done in a computer someplace else. Or how they do it with so few wires.

Maybe I'm on wrong torque sensor patent ? US6843142 was applied in 2004 but references a 2001 foreign patent.

It's going to be a challenge but it can work if kept simple.

old jim

 

[jim hardy]

USPTO dot gov is a "busy" presentation, not too easy to find your way around.

on main page about center just below the US Flag-draped lightbulb is a link "Patent Search", click that

about halfway down that page, below this line "Searching Full Text Patents (Since 1976)""
is a link "Patent Number Search" , click that
and enter the patent number ... without the US prefix.

If you have a TIFF reader installed in your browser, the patent appears.

old jim

 

[Mike_In_Plano]

Not having put more than a moments thought into this...

I'm wondering if this is like the classic LVDT sensors. If one coil can be drive such that the phase / magnatude induced in the second varies with direction / deflectection, than this is pretty straightforward device (electrically)

It must have been an interesting hazard analysis when designing this in...

Mike