Multiecuscan

There is a costly shortcoming within the actuator test of the above mentioned ECU - Selespeed Fiat Stilo Abarth.
It's the Clutch Solenoid Valve test.

The test actualy works however, - IT WILL DESTROY THE COAXIAL CLUTCH CYLINDER! (a.k.a. central bearing or clutch actuator).

This happened to me twice with two brand new clutch cylinders.

Apparently, this test only triggers a command for the clutch solenoid for a fixed time interval (approx. 4 sec).
During this time, pump pressure is not controlled - the pump comes up simultainiously and runs pressure up as high as the pump can deliver (+70bar).

It takes less than one second to fully open the cylinder, for the remaining three seconds the cylinder runs 'up against the wall'.
Unfortunally, this is more than the cylinder can bear. Its simple design can not withstand this pressure and the inner cylinder wall runs over the stopping circlip resulting in damage to the lower seal...
In order to get to the clutch cylinder it's necessary to take the entire gearbox off - which is no fun on a Stilo.

The same function on Fiat's Examiner does not do that...
A simple fix to this problem would be to shorten the time interval of the test to say, one second only.

Martin

Hi Martin,

Did you run any of the other tests before the Clutch SV Test?
Engine running or stopped?
Errors present?

The pressure should always be monitored and controlled by the ECU.
There is no requrement for this function to be executed quckly or to monitor the pressure during execution.

I will limit the open time to 0.5 seconds in next release, but I think that the problem is not in the time.
I remember that you had problems with the pressure sensor.

It is interesting, however, that examiner works fine. I have a log from examiner executing this function and it does not monitor the pressure during execution. It only checks for errors, RPM and Pressure before running the test. I it simply does not run any test if the engine is running or the pressure is low.


Yani

Yani - thanks for getting back to me.

Yes, I did run a good number of tests before this happened.
Bear in mind, this was also the second time it occurred.

What I do when I had a gearbox out and before I ever start the engine:

- Clutch drain (several times, intermittent with mechanical venting directly on the cylinder)
- Line-End calibration
- Switching up and down through all gears several times

If this is satisfactory, THEN I run the engine up and attempt a clutch self-calibration.
(BTW. - this doesn't seem to work properly in FES - although I get the 'completed = 00' result. But this is another subject.)

In this case I did a test ride for approx. 50km. The gearbox was working fine however, the clutch slipped after every shift for up to 6secs. Recording of the clutch pos. parameters in drive proved that this wasn't caused by the clutch itself (i.e. remaining air) as clutch SV current was in sync with the slipping. So I assumed insufficient calibration.
Several more Line-End (Production/Service final calib.) later, the slipping period during shifts had dropped to approx. 3 secs. Which was better but still not good enough.

It was then that I got the idea to run the clutch actuator test - just in order to see if clutch position will be in the released position the very moment solenoid current is back to zero.

Since I run the registered version I’m able to monitor selected parameters during actuator tests. Pump press, clutch pos and clutch SV current were selected.

This test isn’t possible with a running engine so the engine was off.

When I confirmed the test, I could immediately see clutch pos. moving to the disengaged position (which is normally around 7mm), but pump press - which I expected to drop slightly – did in fact climb to a whooping 70bar. SV current (normally around 1Amp during activation) came up to 2Amp – and clutch pos. dropped down to zero and then became even negative.
Before the test period was elapsed, pump press started to drop really quick – the clutch cylinder had given in and the car was standing in a puddle of sele-oil….

I agree – pressure should be always monitored and controlled by the ECU. Alas, it doesn’t in FES.
When it happened to me the first time, I didn’t monitor pressure and assumed the clutch cylinder had a factory fault. With the second failure I think we can exclude this.
The same actuator test was carried out with an Examiner before all this happened and didn’t yield any problems.

Yes, - I had a problem with the pos. sensor. Which was the whole reason to change the clutch cylinder since the sensor isn’t available without the cylinder. A missing – or even disconnected – pos. sensor does not trigger any error.
However, I now know that this got nothing to do with FES – in fact the Examiner will not show an error either. This is down to the SW version of my ECU. Later ECU’s do store this error in memory…

You said the Examiner is logging rpm during actuator tests? As indicated, clutch SV actuation isn’t possible while the engine is running in FES….? At least not in my version, which is the latest (2.2) Or do you mean rpm are monitored in order to determine a running engine and thus block the test?

Unfortunately, I currently don’t have an Examiner in order to direct compare with FES. I’m also reluctant to repeat the exercise since I’m in no mood to take the gearbox out again.

Other than this – great work on your part. You did a great job with your SW!

Martin

Hi,

I couldn't find anything in the documentation regarding this problem.
I think that some of the other test that you ran has left the selespeed ECU in an uncontrolled state or it might be a bug in the ECU software that leaves it in this state.
If you turn the key to OFF, wait at least 5 seconds and then turn it back ON then the ECU always goes back to normal operating state. However there is no requirement for this action after any of the selespeed tests in the documents.

Anyway, I limited the execution time of "Clutch Solenoid Valve" test to 0.5 seconds. This should be enough to prevent any damage.

Yani

Check the black ball reservoir on the selespeed unit. If the rubber diaphragm inside is fractured
the pump overpressurises the system (Pressure overshoot ) The Diaphragm damps the on/off switching
of the pump. I had the same problem on a Alfa 147 Selespeed. You can check this by watching the pressure
changes with the pump switching on and off controlled by the Control unit and not by FES.
devc

Believe me, this got nothing to do with the accumulator. Nothing wrong there.
The fault as decribed couldn't happen on an Alfa either as they got a proper clutch cylinder. It's more a lack in design of the central bearing/cylinder in the Stilo than FES that causes this problem. It just can't take any pressure once it's fully extended which, under normal circumstances, doesn't happen.
Reducing the actuator test interval to .5sec should cure the problem, though.

@Yani: No, this happened twice with two different clutch bearings. The ECU is doing what's supposed to. The bearing will also seize in the fully extended position if pumped up by an enerpac (or other hydr. unit) on a workbench, even without over pressurizing (approx. 40bar). I have meanwhile checked this.

Martin

Hi Martin,

I spent a lot of time on this and still cannot find anything wrong with this particular function. I even checked how exactly this function works in the Examiner and it seems to work the same way as in FES, except for the fact that it does a few checks before starting the procedure.
Here is what Examiner does:
1. Check if engine is running (RPM=0) and abort the procedure if it is
2. Check if vehicle is moving and abort the procedure if it is
3. Check if hydraulic circuit pressure is low (below 40bar) and abort if it is
4. Send a command to clutch solenoid
5. Wait 10 sec (or until user click STOP button)
6. Send a stop command to clutch solenoid

I still think that the problem could be caused by some of the other procedures which has failed to execute correctly and has left the ECU in an intermediate state.
This reminds me that you wrote about your slipping clutch and "Clutch degradation index calculation" which was not producing any results. I actually found that you need to take some additional actions in order to run this procedure. This procedure must be performed if the actuation unit and/or clutch assembly is replaced and if the clutch engages and disengages in an uneven way. Here is how it should be performed:
1. Select "Clutch degradation index calculation" and press the Execute button (engine must be OFF)
2. Start the engine
3. Do a pick-up in 1st gear with the accelerator pressed ~25%
4. Stop the vehicle after ~300 metres
5. Turn the key OFF, then wait ~10 seconds and turn the key back to MAR.


Yani

Thanks Yani, for looking into this - it's valuable info.
There is - of course - also the possibility that the gearbox ECU of this particualr car has a problem of some sorts allthough there is no indication for it in any other function.
The story about the 'Clutch degradation index calculation' is correct however, this has been recheckt on other cars, incl. Alfas.
I also believe that the clutch central bearing generally should be able to withstand any pressure that could normally occure for any length of time. As the fault happened twice - with two different but brand new bearings - there is also the remote possibility that I a had the great fortune to catch a bad batch...

Anyway, thanks for your info.

Martin