ABS Fault Warning

[stevenlizuk]

Hello All,

I have seen this mentioned on the site before and it caused a bit of a stir as no-one else seemed to believe the claimant.

My ABS light started going on and off today so there is plainly an intermittent fault somewhere.

On approaching a roundabout I pressed the brake pedal and after the brakes initially started to come on, the pedal fluttered and the brakes totally released. The brake pedal felt as though the ABS was activated and I could feel the pistons bouncing through the steering.
For the rest of the duration of the pedal being depressed the brakes were totally ineffective.

There was no-one in front of me and the roundabout thankfully was clear, otherwise I would be spending tonight sorting out insurance, or worse.

Be warned. The ABS is supposed to take itself out of the equation if a fault presents itself to the ECU, but plainly in this case it didn't.
Whether this was caused by a sensor flicking in and out of a working state or not I don't know. At the time of the incedent the ABS light on the dash was out, afterwards it was on again.

I am no electronics wizard but this seems to be dangerous to say the least.
If any of you can offer an explanation I'd love to hear it.

Cheers
Steve

[jeremy]

Is the accumulator sound? It can be tested by running the engine for a couple of minutes, turning off and sitting in the boot. The back of the car should sink, then after 30 seconds rise to its original position.

[Mandrake]

Hello All,

I have seen this mentioned on the site before and it caused a bit of a stir as no-one else seemed to believe the claimant.

I believe you!

Your problem sounds very similar to the one I had that eventually turned out to be a faulty sensor wire, although I didn't have any "brown pants" moments, probably by pure luck.

There seems to be far too much faith placed in the ability of the ABS system to detect faults and respond in a safe fashion by some people - although ABS has all kinds of failsafes its impossible to build a 100% failsafe system, and knowing a bit about both electronics and programming it surprises me that ABS is as reliable in practice as it actually is...

What you say about the ABS light being off during the "incident" and then on again afterwards tallies with my experience of the problem - clearly what is happening is a fault condition is occuring and in the short time that it takes the computer (algorithms) to realise something is wrong - seemingly about half to one second - the brakes are inappropriately released.

As soon as the computer realises there is a fault the warning light comes on and the brakes are re-applied, because the computer has switched over to failsafe mode. Again, from my experiences, once this happens the computer will keep it in failsafe mode (light on) until the ignition is turned off and on again.

If the intermitant fault happens to be absent the next time you turn the ignition on the computer does its self test and sensor test, shrugs its shoulders and activates again, despite the fact that this intermitant problem may have occured on dozens of previous occasions.

What it SHOULD do is if the number of intermitant failures exceed a "safe" threshold, it should go into permanent failure mode which can only be reset by a diagnostic computer - thus forcing the owner to get the fault fixed, and preventing it from causing more havoc.

The willingness of the computer to "forget" about intermitant faults in the hope that they have "gone away" seems like a bit of a design flaw to me.

Regards,
Simon

[AndersDK]

Steve -

You are right that this weird problem has been reported before - and that in at least one case the problem was only finally sorted by replacing the ABS ECU box.

Must repeat that one or more - intermittent or steady - failing sensor(s) or their cabling - should NEVER - in any circumstances - set off the ABS function.

The ECU is build to detect abnormalities in the sensor circuits like cable/sensor disruptings or short circuiting - in which case the ABS function is disabled and the warning light is set on. This is because if at least one sensor circuit failing you will never get reliable differential speed signals for an "ABS event" to be correctly initiated.

[jeremy]

I entirely agree Anders - which is why I'm a bit sceptical on the subject.

On the BX if there is an intermittent fault (like a sensor lead) the thing shuts down when it senses a problem and stays shut down until it re-starts. This does work - as I had a front sensor lead fray due to excessive flexing and under certain curcumstances ( a big bump) it would break the circuit and the system would shut down.

I understood that the normal practice was for these things to shut down and stay shut down when a fauly was discovered. The ZX (1994 - presumably a more advanced system than the BX) states 'If the light stays on or comes on during driving, the ABS computer has detected a fault and has disconnected the ABS system. . . '

So why am I saying all this? I think the correct function for ALL is that on startup it monitors and if all components are present it puts the light out. When being driven it monitors - and if a fault is detected - it shuts down and stays shut down.

So how do people report the light coming on for a while and other things. I think the ECU is resetting itself - which means that the power supply must be interupted! - This interruption could be external (even a sparking or unstable ignition switch I suppose) or a fault in the ECU - probably a poor joint somewhere near the beginning of the circuit.

[stevenlizuk]

Hi Mandrake and others who have commented,

The cars main ECU can hold faults in it's memory until cleared, it strikes me that the brake ECU should do the same.
For the system to be able to reset during a journey seems like a system that hasn't been thought through all that well.
In a world where logic prevails the brake ECU should be told that if the car hasn't been switched off for any length of time and a fault is detected and suddenly self repairs then something is amiss.
Surely the system should log problems, and if a fault is detected and then vanishes without the car even being turned off then the system should take itself off line at least until that car has been restarted.
This would also negate the problem of an iffy power supply to the brake ECU.
I will be contacting Citroen about this to see if they can offer any explanation, until then the fuse is being pulled on the ABS.
Even then I'll be leaving myself a country mile to brake for anything!!

Cheers
Steve

[jeremy]

Right

Let me say in what I hope is a simple and completely understandable form - I DO NOT RECOMMEND OR ENDORSE THE RUNNING OF ANY CAR FITTED WITH ABS WITH THE SAID ABS SYSTEM NOT FUNCTIONING PROPERLY OR AT ALL.

ABS is a safety device and if it does not work properly it is one of the responsibilities of ownership to get it working properly before the vehicle is used again. Ifd you had a tyre with a lump on the wall you wouldn't say 'OK I'll carry on driving on it until the manufacturer can be bothered to reply to my letter' and I don't see that ABS is any different.

[windysurferuk]

I had exactly the same thing when I had an ABS fault.

The brakes would fail if the ABS fault light when from off to on or on to off as the brake where applied

Best not to drive the car and get the problem fixed asap.

Turn out to be a faulty sensor on my car.

[stevenlizuk]

Easy Jeremy,

Sounds like you need a cup of sweet tea and a sit down somewhere warm.

Removing the fuse from the ABS renders it totally inert, ie just like every other car out there with no ABS.

I'm not as daft or reckless as you seem to think.

Anyway, can't stop now, I'm off to shoot puppies into the thresher.

Cheers
Steve

[Mandrake]

Hi Mandrake and others who have commented,

The cars main ECU can hold faults in it's memory until cleared, it strikes me that the brake ECU should do the same.
For the system to be able to reset during a journey seems like a system that hasn't been thought through all that well.
In a world where logic prevails the brake ECU should be told that if the car hasn't been switched off for any length of time and a fault is detected and suddenly self repairs then something is amiss.
Surely the system should log problems, and if a fault is detected and then vanishes without the car even being turned off then the system should take itself off line at least until that car has been restarted.

This is exactly what happens - a fault IS logged to it's memory if the ABS light comes on, which can be read by a diagnostic computer, and the ABS IS disabled until the ignition is switched off and on.

However my complaint is that simply turning the ignition off and on makes the computer think "oh, everything is ok now, the sensor tests fine".

It should be realising that the fact that it logged 20 previous intermitant faults should be a sign not to trust the sensor anymore until reset by someone with a diagnostic computer

Instead of that it gives the sensor a new "chance" to prove itself or fail every time you turn the key on...

Regards,
Simon

[Mandrake]

Steve -

You are right that this weird problem has been reported before - and that in at least one case the problem was only finally sorted by replacing the ABS ECU box.

Must repeat that one or more - intermittent or steady - failing sensor(s) or their cabling - should NEVER - in any circumstances - set off the ABS function.

Hi Anders,

So you believe in the possibility of a totally infallable failsafe system ? If so you're a braver man than I....

The ECU is build to detect abnormalities in the sensor circuits like cable/sensor disruptings or short circuiting - in which case the ABS function is disabled and the warning light is set on. This is because if at least one sensor circuit failing you will never get reliable differential speed signals for an "ABS event" to be correctly initiated.

Yes of course, but what happens when the fault is not a simple straight forward fault such as a short circuit, open circuit or missing pulse ?

As a matter of fact in my case the fault caused an ohm reading of 500 ohms for the faulty sensor while good sensors were 1000 ohms. How could a pinched cable cause a 500 ohm reading instead of a short ? I still don't know to this day, but it did...

Regards,
Simon

[stevenlizuk]

Hi Simon,

I am totally with you on this, to my mind it's far too easy for the system to be fooled.

A series of unresolved sensor failures should end up with the system thinking that there might me an issue and to take a back seat 'til a man in a Citroen overall has pressed a magic button on his computer.

Thanks,

Steve

[AndersDK]

So you believe in the possibility of a totally infallable failsafe system ? If so you're a braver man than I....

Rubbish Simon -
You cannot cut & paste from my text and then interpret to your likings ...
My paragraph rigth above reads :

.... and that in at least one case the problem was only finally sorted by replacing the ABS ECU box.

At least you could do me a favour and read the complete subject before you splash in a hasty & smart reply ...

Steve -

The funny thing is that I've tried to explain DIY'ers what you have actually done now : de-activating the ABS system when its behaving weird.
I would rather drive the car with the good old fashioned brakes system (like I do now) - than drive a "clever" car that wont allow me to brake

[Mandrake]

So you believe in the possibility of a totally infallable failsafe system ? If so you're a braver man than I....

Rubbish Simon -
You cannot cut & paste from my text and then interpret to your likings ...
My paragraph rigth above reads :

.... and that in at least one case the problem was only finally sorted by replacing the ABS ECU box.

At least you could do me a favour and read the complete subject before you splash in a hasty & smart reply ...

Hi Anders,

Now you're trying to dodge the bullet I read your message carefully and what you said is quite clear:

Must repeat that one or more - intermittent or steady - failing sensor(s) or their cabling - should NEVER - in any circumstances - set off the ABS function.

You are saying that except in the case of a proven faulty ABS computer that there is no possible way that any type of faulty sensor could cause this problem to occur.

Unfortunately this problem HAS happened, not just to me, but to a number of other people, and in several cases the problem has been fixed ONLY by repairing the sensor or wiring and the ABS computer has NOT been replaced.

Why is it so hard to believe that it could be possible for the computer to be fooled by some specific types of faulty sensor data ? The crux of my argument that no failsafe system can be truly foolproof, and to deny the possibility is just silly.

Regards,
Simon

[Kowalski]

Must repeat that one or more - intermittent or steady - failing sensor(s) or their cabling - should NEVER - in any circumstances - set off the ABS function.

You are saying that except in the case of a proven faulty ABS computer that there is no possible way that any type of faulty sensor could cause this problem to occur.

"Should" and "Could" are two entirely different words.

In theory, the ABS system should failsafe, in practice it doesn't always.

I've had my ABS do slightly odd things when I had a dodgy cable on an ABS sensor. In my case, the ABS would occasionally cut in when it shouldn't BUT it only affected the circuit that had the dodgy sensor, i.e. the rest of the brakes worked properly BUT I did have to press the pedal harder.

I have a theory that a flat accumulator sphere could cause problems in combination with the ABS, but I don't know how the ABS system interupts the brakes and whether it needs additional pressure to operate. Somebody who understands the mechanics of the ABS valve block could no doubt rule out or confirm this theory as plausable or not.