Speaking as someone with an IT background, a diagnosis of "corrupt software" would ring alarm bells. How did it get corruptedm when and why? If ROM based software becomes corrupted, this can usually only be due to a hardware problem. In other words, I wouldn't be surprised if it happens again.
The idea that the system updates its "user" parameters by flashing EEPROMs constantly fills me so much with horror that I can't believe it's actually true. That would be a truly shocking piece of design from a long term reliability viewpoint. Even the best EEPROM's are only good for about 10,000 "flashes" before they die.
Every system I've heard of uses RAM and with good reason.
Cheers
DP
|
The idea that the system updates its "user" parameters by flashing EEPROMs constantly fills me so much with horror that I can't believe it's actually true.
I too wouldn't believe it was constantly flashing the EEPROM, but most microcontrollers have functions which allow the program code to be re-written by the program (!), and this is where I'm suggesting the problem might lie. If it was a hardware issue then I would expect more reccurances of these failues. It happened to a friends LG phone - one minute the flash program was there, next it wasn't, and this is the only explanation I can think of.
--------------
Mike Farrow
|
Another reason not to buy French cars:-(
madf
|
|
A very helpful reply there madf..... The only trouble is the ECU's in these cars are German! (Bosch and Siemens systems)
|
|
|
|
Thanks Mike. I suppose thinking of it that way, it makes sense.
My background is in high end laser printers which I appreciate are quite different, but some of the principles apply. The engine microcode (which tells the printer controller board how to drive the hardware its bolted into) is part of a base firmware file which is fixed and doesn't change unless it is upgraded/downgraded to a revised level. The user parameters (preferences, resources etc) are stored on hard disc (not practical in a car, but RAM will do as power is theoretically never disconnected).
When the system starts up, the base code is always the same, and both lots of data are kept as two separate entities. Although the two interact as appropriate, nothing the user does to the machine can alter or re-write the base firmware.
Although what you say makes perfect sense, and sounds very plausible in this case, allowing user parameters to modify base code instead of being used to select which code to run sounds like a recipe for disaster to me. I'm amazed these failures are not more common.
Oh come back points and carbs..... ;-)
Cheers
DP
|
I have quite a bit of experience with petrol injection ECU's - although at least 5 years out of date now.
On the systems that I worked on the base code was stored in ROM and never modified, but the 'user parameters' such as long- and short-term fuel trim, nr. start cycles, fault flags etc etc were stored in battery-backed RAM. If the battery was disconnected then the ECU would have to re-learn the fuel-trims etc. I can't see any need to do anything different. There are long-term data (immobiliser codes, VIN number etc, injector calibration data) that might have to be put into EAROM - but they would only need to be re-written once or twice in the life of the product following major repair.
The tale of 'corrupt software' sounds like it might have its origins in a hardware fault. I wouldn't hold my breath on that one. £81 sounds a lot - mind you my local GM dealer charges about £50 to update their buggy factory software - which I think is a cheek given they are effectively correcting a 'manufacturing defect' and it only takes 15mins of technician time at most.
|
I too have experience with engine electronics - aircraft engines.
Programming or re-programming EPROMs is the easy bit. The difficulty comes with identifying why this had to be done in the first place. There are many factors that have to be correct before EPROMs can be spoken to. Some of the most difficult to root cause are the ones where the voltage drops out meaning that not all of the program is correctly loaded if at all. Depending on the part of the programme that has not been correctly loaded, the fault may not manifest itself for a long time. Also material defects in the EPROMs can over time and use cause a similar effect.
Best to keep the electronics as simple as possible and make the circuitry as sound as possible.
|
|
It is too early to say whether the fault has been rectified long term. The longest the car ran without the fault recurring, since it first occurred, is about 4 days. If it gets beyond the next week without recurring, I will feel a little more assured.
|
|
|
With regards to "corrupt" software in the ECU, I think what they are trying to say is the orginal software was buggy and the new software fixed these bugs. If the software was corrupt, the car would not function at all. As a hobby I dabble with ECU tuning and one of the major drawbacks when altering the file is correcting the checksums in the data which is stored on the EEPROM when you make an update. Whenever the ECU is initialised, it calculates the checksum for the data which is stored on the EEPROM and then compares it to the checksum which is stored within the data. If these match, everything is ok. If they dont, the ECU wont go any further and will flag a code. My point is, if the software was corrupt, the checksums wouldnt match and you would have a dead car, it wouldn't go anywhere. Thus, in my opinion, the software wasn't corrupt. A bug (glitch) in the software could easily cause phantom codes.
With regards to the program in the ECU "re-flashing" itself, then thats a no-goer as well, certainly for the EDC15 and EDC16 systems used in the PSA HDi's. As somebody said, variables which are calculated are stored in RAM and the original program code stays intact.
|
|
|
|
One of my neighbours runs a 5 yr old diesel Pug 306 bought from and always serviced at a local Peugot dealer.
Recently it failed to start (don't know exact details) but eventually diagnosed by the said dealer as a "virus" in the ECU. Over £600 for replacement ECU, + labour VAT etc around £900 total.
The owner argued that - since the car was supplied by the dealer and had never been to any other garage - the ECU must either have been faulty when supplied or else it had become corrupted in some way whilst being serviced.
The dealer was adamant "These things can just happen".......£900 take it or leave it basically.
You are completely at the mercy of the dealer in cases such as this
Jacks
|
Recently it failed to start (don't know exact details) but eventually diagnosed by the said dealer as a "virus" in the ECU.
.*********
--------------
Mike Farrow
|
.*********
That'll teach me to check my post's gone through before leaving the thread.
A virus? On a car? They mean 'bug', surely? Except they don't like using the 'b' word as it implies an inherent fault, which of course it is, as opposed to a malicious attack. They probably think people will be relieved to hear that they've got rid of this virus before they car drives them over a cliff or deletes all their MP3 files!
--------------
Mike Farrow
|
|
|
|
|
On the plus side at least the software is upgradeable - if it was burnt into ROM then you could be looking at regular replacement of on board computer at £500ish a go!
|
|
And people wonder why main dealers get such a bad reputation too ?
|
|
|
|
Just one question regarding the software, assuming it was at fault. Would it not be covered by the same warranty that covers the rest of the car? In other words, the manufacturer has no obligation to provide updates for free.
|
|
If the vehicle has a mechanical defect or design flaw that causes it not to run properly then the warranty would cover repair/replacement of the relevant part. I can't see why software should be any different.
|
I didn't phrase my question very well. What I meant to say was that the manufacturer has no obligation to update the software for free, once the car is out of warranty, which our C5 is.
With regard to the comment about Microsoft earlier, they only provide updates for their software for a limited period and they certainly don't provide free technical service for their products for very long. The same applies to all sorts of software products for PCs.
|
As I feared would be the case, the software upgrade has not fixed the problem. I went straight back to the dealer, when the fault occurred (after running OK for just one day) and had a chat to the 'technician' who worked on the car. I was very annoyed to find out that most of the information that I had given about the car's recent history, when it was booked in, had not been relayed to him. I told him that I was very happy with paying out £81 to achieve absolutely no progress in finding the fault, especially as he had said there was no anti pollution fault.
The car will be back there today.
|
|
Forgive the typo, obviously should be ''I told him that I was very unhappy''.
|
No problem with typo, thought you were being sarcastic in you post. Computers like cars are fantastic when they work, dreadfull when they don't.
In diagnosing mecanical faults I can see that people there may be some guess work in replacing parts. Why on earth does this occure with software.
|
|
|
|
|
I can't comment on how the software would become corrupted but share the scepticism of an earlier post on the subject also being from a technology background.
Most forms of flash memory or EEPROM type storage have limited numbers of read-write cycles before areas of the memory become 'worn-out'.
This is true of the memory sticks that people use with PC's.
It is possible that part of a program written to a bad or failing part of the memory could be the cause of such problems. However I would assume that a re-write of the software would avoid such areas of this memory and wouldn't neccesarily cause new problems.
|
Since my last post the car has been back to the main dealer (last Friday) and I made sure the technician, I spoke to on the previous day, was made fully aware of all the releveant rececnt history of the car.
The problem would appear to have been cause by a split in a small rubber pipe, that was connected to an EGR valve (engine gas recirculation). At least that is what I am hoping it was. The valve is similar to the ones the garage I normally use had been suspicious of. The car has been fine since then, which is all I can say at present. He also carried out some reconfiguration work on the engine and I am pleased to say that he was as good as his word, as I wasn't charged anything for the work carried out of Friday.
|
My limited understanding is that EGR is pointless on UK petrol engines, it deals with NOx which is not a requirement of UK emissions legislation. EGR has been a problem for many Mundano petrol engines, the fix is easy and cheap ..... so it appeals to most Mundano owners. Not certain if EGR can be disconnected for diseisels in the UK.
ECU software should be open source, plug the laptop in and adjust everything.
A split pipe! Astonishing.
|
My limited understanding is that EGR is pointless on UK petrol engines, it deals with NOx which is not a requirement of UK emissions legislation. >>
IIRC it also deals with unburnt hydrocarbons on the warm up cycle.
"Mundano" is just so not funny any more.
|
">"Mundano" is just so not funny any more.<"
Yes it is! I've had two and bought a third. The Mundano does what it should do, but no-one in their right mind would set the alarm to get up early on a summer's morning to go out for spin in a Mondeo.
Deal with unburnt HCs during warm-up? I thought EGR was disabled until the engine was warm? I am uncertain though, because uncertainty means that I am generally correct ;-)
|
>> right mind would set the alarm to get up early on a summer's morning to go out for spin in a Mondeo.
A few ST220 owners might disagree.
I use a motorbike for that however the Mondeo is pretty entertaing to drive, there are many, many cars that better deserve the tag "mundane".
|
">A few ST220 owners might disagree.<"
It's still a Mundano, it might be quick(ish) but the experience of reaching licence-losing velocities is tedious in the extreme. Hence Mundano.
|
Noooo - don't speak ill of the ST220 - I badly want one as my next car. So badly that it hurts.
The noise they make is fantastic and they handle amazingly well for such a big car.
3000cc = :-)
|
|
A split pipe! Astonishing.
What is astonishing, that it was a split pipe, or that the split caused the problem?
|
|
It seems that an oversight that occurs when using diagnostic machines is that preliminary checks are not carried out. I don't know whether engineers on here agree, but an instructor (maintenance of industrial FLT) I was talking to recently was concerned that his technicians were often blind to the limitations of the dignostic test machine, and went on wild goose chases because they hadn't eliminated the possibility of mechanical or other failures before the test was run. His main concern was speed of return of the FLT to productive service. For the punter in the grip of the car trade its that, plus the cost of redundant and expensive replacements. I would have expected a check of vulnerable components at least, to ensure the system was in working order before the leads were applied! Maybe it pays garages to misdiagnose?
|
|
|
">What is astonishing, that it was a split pipe, or that the split caused the problem?<"
It's astonishing that it took so long for the technicians/fitters/mechanics to identify the problem. That's doesn't mean that I would have spotted it though!
|
|
I wouldn't know how easy it is to identify a small pipe like the one that was found, as there are a multitude of them on an HDI. The chap who services the car ruled out a split pipe, on the basis that the fault only occurred above a specific speed and didn't recur straight after the fault had been cleared. It always took at least a day for the fault to recur. I imagine the techician at the main dealer had a bit more experience in dealing with this kind of thing on C5s.
|
|
|
|
|
|
|
|
|
