Found out quite a bit re the EPS Torque / position sensor.
As far as I can tell, this is the model from which it is designed.

http://www.bitechnologies.com/pdfs/sx4300.pdf

Pinout is same. (in reference to eLearn wiring spec for EPS)
There does seem to be some variation with a 6 wire and 7 wire unit (both 8 pin) but that can be sorted - I have a unit on the way courtesy of a kind gentleman here and Paypal.

The info and diagrams in the datasheet should enable a reasonably comprehensive series of tests to be done on any suspect unit. AN oscilloscope would be the best (I will use that to start with) but a good quality Multimeter with analogue bargraph would possibly be ok.
It looks simple enough to build a small hand held sensor / tester and aligner.
I work with embedded microprocessors for a hobby so may build a small test unit.
If I build one, would it be useful if it was sent to the club / forum folk to lend it out to forum members ? - (not much use for me to do it as I am in portugal.) ?
It wont be for a few months but I can get something together no worries. It would be up to the club to sort out the loan of the tester though. !
If I could get hold of torque sensors for other models (and the sockets from the ecu !) - I could add a range of units to one model..
I suppose the thing is is it worth it to the members - as in, would it, in itself, be informative or helpful ? - and also if a 'club' unit would be workable ?
Let me know what you think.

Below is pinout of torque sensor to eps node ecu according to fiat... (Note it is identical) - new sensors can be had for 100 euros inc vat. - can be changed in the car without removing steering column..... (edit - I originally put without removing wheel doh ! )
Just pondering......
1 P3
2 P1
3 Vref
4 T2
5 NC
6 P2
7 GND
8 T1

captainslarty said:

Found out quite a bit re the EPS Torque / position sensor.
As far as I can tell, this is the model from which it is designed.

http://www.bitechnologies.com/pdfs/sx4300.pdf

Pinout is same. (in reference to eLearn wiring spec for EPS)
There does seem to be some variation with a 6 wire and 7 wire unit (both 8 pin) but that can be sorted - I have a unit on the way courtesy of a kind gentleman here and Paypal.

The info and diagrams in the datasheet should enable a reasonably comprehensive series of tests to be done on any suspect unit. AN oscilloscope would be the best (I will use that to start with) but a good quality Multimeter with analogue bargraph would possibly be ok.
It looks simple enough to build a small hand held sensor / tester and aligner.
I work with embedded microprocessors for a hobby so may build a small test unit.
If I build one, would it be useful if it was sent to the club / forum folk to lend it out to forum members ? - (not much use for me to do it as I am in portugal.) ?
It wont be for a few months but I can get something together no worries. It would be up to the club to sort out the loan of the tester though. !
If I could get hold of torque sensors for other models (and the sockets from the ecu !) - I could add a range of units to one model..
I suppose the thing is is it worth it to the members - as in, would it, in itself, be informative or helpful ? - and also if a 'club' unit would be workable ?
Let me know what you think.

Below is pinout of torque sensor to eps node ecu according to fiat... (Note it is identical) - new sensors can be had for 100 euros inc vat. - can be changed in the car without removing steering column..... (edit - I originally put without removing wheel doh ! )
Just pondering......
1 P3
2 P1
3 Vref
4 T2
5 NC
6 P2
7 GND
8 T1

Click to expand...

Sounds like an interesting hobby for you though and I'm sure you'd have fans on here and probably a nice little earner on ebay if you can provide fixes:


1. Ultimately the question is, what is failing? Is it always the same component? Is this something worn or some electronic glitch - resistor/diode/processor failing of some sort?



2. Depending on the answer to the above - how realistic is it for a layman to fix, and if it isn't who can do the work cheaper than what's currently charged?


Last time I saw an oscilloscope it was in the school lab in 1982. It was the size of small TV and the only use was changing the frequency of the sine-wave. At that age though it was seriously cool. I expect there's an Android app for that now.


Have fun.

I think you are possibly missing what I am suggesting,
I am saying I produce a small handheld box for testing the torque sensor, which is, afaik, the only thing that fails in 99% of case, the rest of the time it is poor connections and supply voltage.
The tester would be given, free, to the forum owners, (if they wanted to manage the use), and a user with a genuinely suspected issues may 'borrow' the test unit from the forum, by post, deposit, at a meet-up - I dont know really - I dont want to 'produce' them. I have no need for the agro of business or working, or indeed the money, but, I would be happy to produce plans for someone competent in that area to build one. Once built, it can be used by someone not as competent in electronics, but yes, it is not for someone who does not understand basic concepts, or even what torque is - or position sensing, or the general layout of a modern vehicle.
It would simply plug in to the sensor with ignition off - they turn the wheel, it tells them if it's ok.
I am hoping as well that further investigation on an actual unit can result in a test using a multimeter only (possibly difficult, but not impossible). there is absolutely NO information out here regarding testing your own sensor, for beginners or electronic experts alike, so I hope even if it results in purely having the information of HOW to test it and HOW it works, that is enough for some.
Hope that makes sense.
The test unit I am contemplating would be a simple 2 line lcd readout, Torque, position and sweep scale for testing the axis. or to put it simply, wiggle it a bit and see if it says continue, then turn slowly until zero on both readings. thats it.

Just as an edit to the above, re what is failing -
When one looks at the multitude of 're-conditioners' out there, there are LOADS ! - whatever is causing issues is minor and relatively constant. They could not do it for the price otherwise !. the sensor, if they are buying in bulk, will be costing them as little as 25 euros, they can clean a unit and refit a sensor in about 1 hour, that is a LOT of profit. !
I do not believe for one minute they are going into ECU and changing drive components, or,motors (a 3 phase unit) or in fact, anything else apart from a bit of gunk, switch cleaner, and POSSIBLY a new sensor. money for old rope, but cheap if one has no alternative.

Point taken about diagnosis, and you're probably right that, low-current aside, it probably is that sensor 99% of the time. Presumably they knew this and the part was quietly upgraded later on.


Ultimately the unfortunate victim will need to replace/recondition something - having a meter to confirm the rogue part at least avoids a trip to see the teenagers at your local dealer. STE - straight to ebay after analysis confirmed.


Keep it up.

Thanks sir,
To also clarify the type of failures, I believe that in 99% of cases of GENUINE issues it is the sensor, but I strongly believe that in perhaps 75% of presumed problems, there is NO FAILURE, it is simply poor contacts / connections. the act of removing and refitting the unit 'cures' most of these so the re-conditioners are, to a degree, getting away with money for nothing.
The sensors ARE extremely well built and reliable. I have not seen a case of a reported sensor failure here yet (I havent searched ALL the posts) what I have seen is many signs of simple battery and connection issues.

Good work Captainslarty.

This is exactly the sort of thing I'd like to see more of on the forum. Finding cost effective ways of fixing the common (and the not-so common) issues will be of great benefit to many owners.

I'd agree that the reconditioners are most likely only changing the sensor 99% of the time.

Keep up the good work!

A brief idea how the sensor works....
First the torque, later position.

Ok, what you have in the sensor are two small splined location that can move in relation to each other (these are what you can see in the centre. of the unit.
The main steering column attached to the lower end passes through the lower splined fitting and locates there, it then is about 8 inches long forming the INNER part of the upper column.
Now, the piece of the upper column with the thread on for the steering wheel fits over this and locates to a two inch splined section on the top of the inner shaft we just mentioned, the lower end has a small splined area that fits in the other torque sensing ring in the torque sensor unit.
What you now have is the upper steering column made of two parts, connected by spline at the top only, when you turn the wheel, if there is resistance, the INNER steering shaft turns slightly along its length (it twists !- it is designed to) due to the applied torque - it is only located at the top to the outer shaft sleeve, so the outer shaft sleeve transmits that turning amount (twist) to the upper torque ring in the sensor (attached at the base of the outer shaft sleeve, the lower torque ring is attached to the inner shaft. There will be a difference in position of the two torque sense rings when torque is applied, this is measured by a mechanical wiper resistance inside the unit. there are two wipers, T1 and T2, they produce a voltage out in relation to the movement as you can see in the T1 T2 output pattern. at rest, they both are equal - with 5V applied, the output at rest on both T1 and T2 will be 2,5V.
The torque can be calculated by the ecu by the difference in voltage between t1 and t2, also the torque direction.applied
Thats it for torque..
Next we can look at position sensing....
Any questions on that bit please ask..

This video from Dorman will show you the shaft arrangement for torque - it may take a wee bit to get the grey matter around it, but it is fairly simple really.'The panda unit is VERY similar to this.

Also a huge thank you to Most Easterly Pandas (Jon) for the excellent price and service for a LHD unit to experiment with ! - cheers Jon.

theres quite a few italian you tubes about the torque sensor and setting it up
i have been saying for 5 years that reconditioners on thewhole are basically swapping components round and then as you say making sure voltage at straight ahead for torque sensors is nil
i never believe they are buying these as they are still €100+ even if you buy direct
if you could build a tool to align the sensor up i would have bought one but unfortunately this analogue technology is now giving way to newer stuff
but i appload your work
i too like a challenge

s and b said:

theres quite a few italian you tubes about the torque sensor and setting it up
i have been saying for 5 years that reconditioners on thewhole are basically swapping components round and then as you say making sure voltage at straight ahead for torque sensors is nil
i never believe they are buying these as they are still €100+ even if you buy direct
if you could build a tool to align the sensor up i would have bought one but unfortunately this analogue technology is now giving way to newer stuff
but i appload your work
i too like a challenge

Click to expand...

Cheers man,
There is no problem using digital technology, but actually analogue is actually easier to work with.
The industry is slowly changing to non contact - hall effect and strain gauge type, both give analogue or pwm outputs, either can still be easily tested.
Building a test unit is actually quite easy (in theory) I program similar systems in the context of waveform sensing and position sensing for a UK power products company - (Automotive and Marine).
I have seem a lot of the Italian videos, the guy has some nice kit too. Notice the boxes of sensors he has lol ... I bet he is only paying 20 / 25 euros a sensor if that.. and one size fits MANY units. Often it is just the plug on the end for the ECU that is different.
Even if they put the new sensor in, I think in many cases there is nothing at all wrong with the old one..
As you will see my friend, in the next bit, centring the unit is simplicity if using the 3 wire (3 P output) Units, and still relatively simple if using the 2P output units. (in some ways easier ...

I think it will be a long time before we get sensors with a can output directly due to an additional node cost, the move will, I believe, be simply to contactless with either PWM or frequency or indeed, analogue output still. Whenever you look at a change of state, analogue is best imho - but at the microprocessor end, it doesn't really matter.
I can build a tool to set any of the current and planned ones up and test them. Also publish plans and release the code for someone to build their own unit. cant see it costing a lot in parts either, a simple PIC micro, some buffered ADC inputs and a cheap 2 * 16 lcd display . The sockets and adapters for the differing plugs will cost the most lol.. I have also written to a major manufacturer asking for samples under their education program. so will see what happens. It keeps the grey matter active mate.. and hopefully helps someone. I just got a bee in my bonnet when my light came on for the steering, I was determined I would only pay for a new or recon unit if I absolutely KNEW I had problems. I suppose being a 'person of leisure' helps I try to avoid real work.....
Seriously, I hope it helps someone - it is all a learning process.

Ok, lets look at position now.
From the datasheet, we have 3 analogue signal outputs, P1 2 and 3.

P1 lags P2 by 90 degrees phase shift.
P3 steps are corresponding to P1 min and max range.
also they correspond to P2 mid range to mid range as a sanity check.

It is therefore a simple matter to determine position by looking at the step value of P3, then the analogue value of P1 or P2 (actually both) to be a valid reading.

It can be seen that position is easy to determine, also, direction of movement is by looking at current value in relation to old value. Delta (or rate of change) is simply the speed that the position changes (how fast the wheel turns!)

The ECU looks at the speed of the vehicle to determine the amount of assistance required, It also looks at the torque applied to determine effort and provide power to overcome, hence on slippery surfaces power is reduced. at higher speeds power is reduced. The steering geometry is different to a 'normal' car so there tends to be a more neutral steering with less castor effect and return to centre. The rack is also more like those used in rally cars -- around 2 turns lock to lock.! as opposed to 4 turns in a standard non powered car. As can be deduced, loss of power leads to HEAVY steering with little return to centre. The power unit controls the return to centre much more than a non powered unit and actively assists this.

It should also be fairly easy to see that tests of functionality on such a unit could be accomplished by looking at the values output over the range and checking for continuous signal (no gaps!) and also the built in error checking of the 90 degree phase shift and stepped dc voltage function.
This could be done with a multimeter (preferably Three !! analogue meters) but see later . A dual channel oscilloscope and a multimeter would be fine but not something the average mechanic has. Best thing is a simple microprocessor unit that all 5 signals plug into, torque can be checked by locking the lower shaft and turning the upper and comparing linearity and consistency of signals, Position, the same way, linearity, phase shift and consistency. Setting centre is a simple matter of 0v on P2, 2,5V on P1 and P3.
0 and 2.5V can be replaced with Min reading and (Max-Min / 2)
So, it reality, it could be mostly tested for common faults with a single multimeter, Position AND torque ! and also centred with a single multimeter.
Not too difficult really, - a small microprocessor tester would be far faster, but for the home user, not cost effective unless it could be borrowed or hired.
Cap.

HI. Today i have just successfully replaced my steering sensor in situ on my 2005 panda. I was getting the vibrations though the wheel and eventually the EPS fails and the warning light is displayed. Car is still drivable and i unplugged the sensor for a while and albeit heavy the car is drivable

After a good bit of research i bought the replacement sensor (with a GREEN sheath) and 6 wires in an 8 hole plug from Portugal for £60. from a company called INOVAUTO.COM.

I have taken loads of pictures showing how i did the swap and the only thing i had to make was a huge pair of circlip pliers, so out with the welder and an old pipe wrench and two old 2.5mm Allen keys to fabricate this.

The new sensor comes LOCKED at the Straight ahead position. So provided the steering wheel shaft remains in this position, everything falls into place with keyways inline. then remove the wire lock.

After putting the new sensor in, and replacing everything, the steering eps light remains off, and electrical alignment has not needed any programming via the service port.

New to this forum so need some guidence where to share my pictures.

Mike

bluepanda said:

HI. Today i have just successfully replaced my steering sensor in situ on my 2005 panda. I was getting the vibrations though the wheel and eventually the EPS fails and the warning light is displayed. Car is still drivable and i unplugged the sensor for a while and albeit heavy the car is drivable

After a good bit of research i bought the replacement sensor (with a GREEN sheath) and 6 wires in an 8 hole plug from Portugal for £60. from a company called INOVAUTO.COM.

I have taken loads of pictures showing how i did the swap and the only thing i had to make was a huge pair of circlip pliers, so out with the welder and an old pipe wrench and two old 2.5mm Allen keys to fabricate this.

The new sensor comes LOCKED at the Straight ahead position. So provided the steering wheel shaft remains in this position, everything falls into place with keyways inline. then remove the wire lock.

After putting the new sensor in, and replacing everything, the steering eps light remains off, and electrical alignment has not needed any programming via the service port.

New to this forum so need some guidence where to share my pictures.

Mike

Click to expand...

Hi Mike, and welcome to FF,

good stuff..!! how many miles on your panda..??,
as our '04 active has done 11 years and 80K with only a duff original battery messing-up the steering function,

BUT we've ( recently) also got a 2007 that's been Western Rebuilt already ( 50K) could well have been mis-diagnosed though..,

pics can be added after 5 posts..,
BUT , adding a GUIDE would be GREAT if you've the time..,
Cheers,
Charlie - Oxford

HI my panda 2005 Dynamic has done 126987miles.
Other than having 2 cambelts and regular servicing, still going strong.

I took a whole load of pictures in conjunction with Fiats ELearn Maintenance Manual.. The only thing that flummoxed me was how to get the air bag off, but soon got the knack.

Mike

bluepanda said:

After a good bit of research i bought the replacement sensor (with a GREEN sheath) and 6 wires in an 8 hole plug from Portugal for £60. from a company called INOVAUTO.COM.

After putting the new sensor in, and replacing everything, the steering eps light remains off, and electrical alignment has not needed any programming via the service port.

New to this forum so need some guidence where to share my pictures.

Mike

Click to expand...

Hey mike, id be interested in seeing how you did it, would you mind sharing?

Did you check the calibration with mes?

... also, did anything come of a reader for the steering torque sensor?

it seems that the garages are really charging for the setup of the sensor, which doesnt seem difficult, but you probably need the gadget to give you the readings...

zippeyrude said:

... also, did anything come of a reader for the steering torque sensor?

it seems that the garages are really charging for the setup of the sensor, which doesnt seem difficult, but you probably need the gadget to give you the readings...

Click to expand...

Does anyone know how to calibrate the torque sensor with a multimeter, or where a reasonably priced tool can be bought or hired for a one off job?

MultiECUScan will do it, but you need the licensed version.

If we knew where you were, someone might be nearby who could help.

portland_bill said:

MultiECUScan will do it, but you need the licensed version.

If we knew where you were, someone might be nearby who could help.

Click to expand...

Thanks for the reply Bill,

I have a proper registered copy of MES, but the wheel still self centers quicker when returning from steering right.

The car is genuine and hasnt had any knocks.

The I found the following on MES "current absorbed = too high". I have heard that the battery / earths can cause issues but I cant see how this would affect the steering in one direction only.

I checked the steering position and it is absolutely centered on the rack.

After I calibrated torque sensor there was no change...

This seems like a torque sensor issue, requiring replacement.

Normal repair is to replace the top half of the column. I did mine earlier this year, got one from Western Power Steering, see this thread: https://www.fiatforum.com/western-power-steering/342222-panda-electric-power-steering-problems.html
They give a lifetime guarantee. Replacement is not difficult, but lots to dismantle and refit. You should get two new lower column pinch bolts, these should not be re-used, and are grade 10, not the usual grade 8.8, so easiest way is from your dealer. You don't need the steering lock pinch bolts as this comes with the column, you just swap your lock barrel.

Recently someone has sourced just the torque sensor and replaced it themselves. Search for the thread for all the details. This is cheaper, but little or no warranty.

The eps torque sensor connector can affect the analogue voltage readings before they are turned digital and computed. I have fixed a meriva that uses the same brand EPS (DELPHI e-steer) and same torque sensor. My fix could apply to many FIAT models. If the eps gives less assistance in one direction + intermittent + works smoothly in other respects, then this can be a very cheap fix.


From thread Meriva EPS light


I belive that bringing the EPS back from fairy constant reduced left assistance proves that a sensor connection has been improved. The qlitches that it started with could be anything, but as they are also gone it must also have been related to the connector.

Connections I have fixed with deoxit:
1. Spotlight in kitchen fan/extractor (9 years old) intermittent connection - maybe 15 applications during 2 years. Due to constant humidity from cooking I expect it to fail again, and apply deoxit again.
2. Eps connector Meriva (12 years old)- First intermittent connection, then constant bad connection evry second day (more resistance/lower voltage read from left hand torque signal). Cause - bad quality/bad production connector as with the Hdmi connector ?
3. Hdmi connector TV (3 years old), intermittent connection for two years, one application and fixed.

Since fixing the connector I have read more about EPS systems and voltages, and feel more sure about what was wrong. I never managed to read related error codes, but voltages explains it all. Think about the steering as some kind of analog joystick (If it was, then 0.25v would be f.ex full left, 4.75V full right, 2.5V straight ahead. Bad connection would steeer you to the left. As it is set up the EPS will never steer you towards lower voltage.
It has two separate voltages as a safety measure.

The connector has 6 active pins (8 total, 2 empty) and 3 voltages.
Left hand (turns left when voltage is above 2.5 Volts, range 0.25 - 4.75 volts). When steering wheel is "squeezed" to the left this voltage increases, and the other decreases.
Right hand (turns right when voltage is above 2.5 Volts, range 0.25 - 4.75 volts)
Position (not sure about how it works).
If Left hand or right hand drops to 0 in a glitch, EPS is deactivated.
If left hand has a "neutral" voltage of f.ex 2.2 volts because of established, stable bad connection, turning to the left will be harder than to the right (you have to be above 2.5 volts for anything to happen, and you will never reach 4.75 volts).
Improve the connection with a good contact cleaner, f.ex deoxit, or just move the connector back and forth. Bad connection will degrade this connector very slowly because it is very low ampere, maybe 30-40 milliampere.