Hi all,
I have been looking into the EPS system on the panda since I had a dash light illumination and loss of steering (cured on restart - probable fault - battery charge and connections)
I am a marine Eectronics Engineer, I retired at 47, but still do small jobs for friends etc. I want to look into the EPS side and come up with some basic fault finding methodology for normal folks at least skilled enough to use a multimeter. - my first thoughts are ....

Firstly, there are MANY places offering a cheap replacement service for a supposedly reconditioned unit. At the proces they charge (which seem quite reasonable) it is my opinion that there is very little, if anything, done to the vast majority of units apart from a strip and clean. In certain case, the torque sensor can need replacing, these are available quite reasonably, with examination of a used unit, I am sure the part(s) could be obtained at very little cost.

There are a few areas that could cause issues and the simple act of replacing a unit will, in effect, cure a lot of these issues due to plugs and connectors being removed and replaced (which can often cure an intermittent fault on it's own)
First thoughts after consulting fiat documentation and other sources.
If you have problems with the steering, check the following.
Connector from torque sensor to EPS unit (probably green cable sheath) - remove, clean with switch cleaner (NOT WD 40 ETC!!) and replace. Same for any other connection to EPS control unit (ECU).
From the EPS Control are two direct cables, a ground, to the battery, which should also then be itself connected to the body AND the engine block - (These should be removed at the battery and cleaned, small amount of vaseline, then replace, then remove and clean the engine earth and body earths, there is also a power feed coming from fuse 05. remove the fuse and check the location is clean - use switch cleaner if needed, a toothbrush is also useful. Place a tiny smear of vaseline on the blades before replacing
There is also a dashboard earth, this is an important one, it is located in the right hand side (on a LHD car) footwell. (Will try and get pictures). Also, I believe there is a ground connection near the steering column base. ! remove, clean, vaseline and replace.
The other possible culprit is the speed sensor - you may well see a warning in MES related to this !!. this is located on the gearbox near the diff. It is fed from fuse F42 so check and clean the fuse, and also the connector to the unit.

Also check and clean F24 F23 and F51 which form part of the main feed circuits for the body computer and ignition.
Refer to handbook, haynes manual and also elearn available from the downloads here.
I am sure with careful approach most problems can be cured.
Will update this as I find out more and try to get some images - forecast is for nice sunny weather next week.

ALL battery terminals and leads should be sparkling clean, protected with vaseline, and covered - if no cover, use a small amount of clingfilm to stop dust adhering to the vaseline.
As far as cleanliness and checking, same goes for all ground connections, all positive connection, alternator connections, starter motor connections.
Do all this with battery off and preferably wait 10 mins so no risk to airbag circuits.
Carefully pull connectors to main body ECU also and other engine sensors, small amount of switch cleaner, small amount of vaseline on cover, replace.

Doing this should hopefully cure 90% of all problems, and also be preventative maintenance for the future if all is good.
(Cap. Joe)

have you seen this
captainslarty [ame="http://www.liveleak.com/view?i=f14_1393634596"]LiveLeak.com - Trouble with the steering wheel[/ame]
good post by the way but you do get strange looks in this country when you buy your vaseline off the counter at poundland

Thanks sir,
That is a classic torque sensor issue or ecu gone crazy. What it does show is how damn quickly the eps can respond. - it is an amazing system, until it goes bang lol

Warning, long fairly techy post follows, if you get bored easily, give up now lol
Apologies in advance for any typos or errors.....

Ok, to expand a wee bit on the above, let us look at an actual example of an EPS failure and cure, the errors reported, and the logical deductions that can be made to aid a fix.

My own panda, a 2003 1.2 8V, Dynamic. On a roundabout, steering stopped getting power assistance and warning light on on dash, ignition off, wait 10 seconds, on again, no light, all ok. Car had been sluggish starting for a few days as if battery was weak.

Ecu (EPS Delphi) shows following with MES MUX and CanTieCar -

ECU ISO Code: 52 80 86 02 2F
21/02/2014 15:59:23
CONNECTED TO:
Fiat Panda '03 1.2 8V
Delphi Electric Steering
--------------------------------------------------------------

ECU ISO code: 52 80 86 02 2F
FIAT drawing number: 735374944
Hardware number: 2609994607A
Hardware version: 00
Software number: 2610184805A
Software version: 4201
Homologation number:
ECU programming date: 09/25/2003

READING ERROR CODES:
1: C1004 - Alternator
No help available
The reason for this fault is that ECU did not receive signal from the sensor. The fault is not detected now, but it is stored in memory. Clear fault codes, and observe for future appearance of the same fault. Dashboard warning light was activated for this fault.

2: C1003 - Vehicle speed (from CAN)
Speed signal anomaly.
The signal is sent by the electronic tachometer or by the on board computer, hence verify the tachometer proper functioning.
If it works properly then check the ECU connections to the instruments board or to the body computer.
The reason for this fault is that ECU has received invalid signal from the sensor. The fault is not detected now, but it is stored in memory. Clear fault codes, and observe for future appearance of the same fault. Dashboard warning light was activated for this fault.

3: C1007 - Battery voltage
The ECU supply voltage is below or above the specifications.
Check the voltage between all the ECU harness positive and negative terminals key on, engine off, during cranking, and when the engine runs. Further, measure the voltages with and without the electric loads turned on, and verify that there are no ground differences when under load. Check the charging system voltage and the battery and cable condition.
NOTE: Disconnected battery or cranking the engine with weak battery may also set the code.
The reason for this fault is that ECU has received very low signal from the sensor. The fault is not detected now, but it is stored in memory. Clear fault codes, and observe for future appearance of the same fault. Dashboard warning light was activated for this fault.

4: C1002 - Steering column motor torque position sensors
A problem is detected in the sensor(s) or the power steering motor assembly, as indicated by the code.
Check the system fuses and the connectors and wiring for opens, shorts and weak connections.
If the wiring is in good condition and the problem can be directly pointed to an internal component, it may be necessary to renew the entire servo motor assembly, since repairs are usually not possible.
The reason for this fault is that ECU has received invalid signal from the sensor. The fault is not detected now, but it is stored in memory. Clear fault codes, and observe for future appearance of the same fault. Dashboard warning light was activated for this fault.
____________________________________________________________

Let us examine the above, what is strange is we have 'faults' 1 and 2 from the body node (ECU) - 3 possibly also from the body node via canbus. but could be internal to the eps ecu.
4 is an eps node error, but non specific when you look at the others.

What the car saw was a loss of speed signal (this goes from the gearbox sensor to the body node, then via canbus to the EPS node, hence a lot involved in the chain - unlikely to be related to EPS, it just needs speed sense to function.
Also a non charging warning (flagged as alternator) this is monitored again by the body node and sent via canbus to the EPS node.
It doesnt add up really does it ?

Now, lets us look at the voltages reported by the body node and also the eps node (fault cleared) engine running.
Voltage measured at battery with Fluke meter 14.090 V

ECU ISO Code: 0D 07 BC 83 A8
21/02/2014 16:12:58
CONNECTED TO:
Fiat Panda '03 1.2 8V
Magneti Marelli IAW 4AF/4EF/59F/5AF EOBD Injection
--------------------------------------------------------------

ECU ISO code: 0D 07 BC 83 A8
FIAT drawing number: 55192636
Hardware number: IAW5AFHW607
Hardware version: 00
Software number: 20B6W61
Software version: 0000
Homologation number: 5AFSP
ECU programming date: 09/26/2003

READING PARAMETERS:
Battery voltage: 13.7 V

Now note the steering ecu voltage !

ECU ISO Code: 52 80 86 02 2F
21/02/2014 16:18:15
CONNECTED TO:
Fiat Panda '03 1.2 8V
Delphi Electric Steering
--------------------------------------------------------------

ECU ISO code: 52 80 86 02 2F
FIAT drawing number: 735374944
Hardware number: 2609994607A
Hardware version: 00
Software number: 2610184805A
Software version: 4201
Homologation number:
ECU programming date: 09/25/2003

READING PARAMETERS:
ECU supply voltage: 14.087 V

Note the readings !!! EPS is almost identical to calibrated fluke meter!.
BUT, body node is reading 0.3V LESS than the battery.

That indicates a fault between the battery and body node and or a bad ground, the body node has its own ground, the supply comes from the battery, via a fuse, to the ignition switch, then via a fuse and then to the ecu..

With a differential like that, it is no wonder erroneous faults appeared,
Strip and clean all connectors and grounds, fuse holders and fuses and voltage readings returned to normal.

Fault cured. (we also found out that the battery was being drained by a faulty radio never turning off, but that in itself did not cause the errors!)

That is one example only - where access to information in depth can confuse or help, in this case, the 4 logged faults at the time of event had no logical conclusion apart from electrical glitching. No definite pattern.

Voltage checks (With MES) identified a problem, one could have traced the problem to the exact point with a meter, but it was worth stripping and cleaning ALL connections.
A multimeter on its own could have identified the voltage readings and differentials showing a definite ground or supply fault, but not, of course, the error codes.

Sometimes, you may find the problem of a low supply to something is not related to ground, it is in the actual hot line - it could well be the ignition switch itself ! much power passes through here, my car has been started over 18000 times !! (I know that for a fact as it is logged in the ABS node ecu - 18000 turns of the key, 18000 contacts made and unmade, you can see how a single component can cause issues.

ECU ISO Code: 25 86 13 02 15
12/02/2014 11:59:10
CONNECTED TO:
Fiat Panda '03 1.2 8V
Bosch ABS 8
--------------------------------------------------------------

ECU ISO code: 25 86 13 02 15
FIAT drawing number: 71718157
Hardware number: 0265800306
Hardware version: 00
Software number: BB36803
Software version: 0200
Homologation number:
ECU programming date: 06/02/2003

READING PARAMETERS:
Startups counter: 18070

(Amazing what you can find out from the right diagnostic kit)

Definitely a shopping and school run car !- it was also one real lady owner but well serviced and in a warm country with no salt.. (The CAR ! - not the lady owner lol )

Slight differences in supply voltages caused by a minor resistance in something like the ignition switch, or a relay for example, are not too much of a concern (unless they are breaking down and causing drop outs - very rapid disconnect and reconnect - a spike as such -or dip), The big problems are caused by ground potential differences.
Engines are mounted on rubber mounts to reduce vibration, this also insulates them, that is why many sensors and the engine itself has grounding straps to the vehicle body..

Hope that is of some use to some.

On diagnostics, here is a reading on the first panda I had before it went back, a 2010, raving mint, but, 90,000 Km had magically dissapeared from the odometer... (Clocked !) - money back
Caveat emptor !!!

How to tell if your car is clocked......

Number plate 02-IX-59
VIN: ZFA16900001667971
MODEL: ACTIVE 1.1 FIRE MPI 54CV 5M E4+EOBD ECO
FUEL: PETROL
CUBIC CAPACITY: 1108 CC
INTERIOR COLOUR: BLACK FABRICS
EXTERIOR COLOUR: GREY
POWER: 39 KW - 53 CV
BODY: 2 VOLUMES
DRIVE: LEFT

ECU ISO Code: 0D 07 C1 83 AD
10/01/2014 13:26:05
CONNECTED TO:
Fiat Panda 1.1 8V
Magneti Marelli IAW 4AF/4EF/59F/5AF EOBD Injection
--------------------------------------------------------------
ECU ISO code: 0D 07 C1 83 AD
FIAT drawing number: 51834074
Hardware number: IAW4AFHW407
Hardware version: 00
Software number: 20BBWR0
Software version: 0000
Homologation number: 4AFSF
ECU programming date: 02/12/2010

READING PARAMETERS:
Odometer: 120565 km (Actual True Reading from main ECU)
Odometer at last programming: 0 km
Number of rewrites: 1
-------------------------------------------------------------------------------------------------------

ECU ISO Code: 32 83 13 02 1F
10/01/2014 13:21:23
CONNECTED TO:
Fiat Panda 1.1 8V
Instrument Panel Marelli (Base)
--------------------------------------------------------------

ECU ISO code: 32 83 13 02 1F
FIAT drawing number: 51871010
Hardware number:
Hardware version: 00
Software number: 04000005205
Software version: 0403
Homologation number: 000000
ECU programming date: 02/22/2010

READING PARAMETERS:
Odometer: 32227 km (Altered Reading on Odometer Clock)

Cap.