|
I made a thread a while back asking why ECUs were so expensive. The answer I got was because they have to work in extreme temperatures and surve al the abuse it would get living in a car.
|
I made a thread a while back asking why ECUs were so expensive. The answer I got was because they have to work in extreme temperatures and surve al the abuse it would get living in a car.
So they make the chips by some wholly different process to to those in a PC? I doubt it.
I have knowingly never seen an ECU, but from the pictures they look like a small and simplified computer in a rugged case. Since the major manufacturers produce several million mass-market vehicles each year, I can't see any reason why they shouldn't be using very similar ECUs for all of them, just programmed a bit differently. There must be sufficient economies of scale here to make these things pretty cheap, even if they use entirely customised chips (which I really doubt).
A basic car like a Fiesta leaves the factory with a price tag of about £5,000, and I can't imagine for a moment that the makers would tolerate more than about £100 of that going on a little electronic box, and even that's probably too high. If replacement ECUs are being sold for many times that price, I can't be believe that reflects the actual marginal cost to the maker of building and delivering one more ECU.
Edited by NowWheels on 19/09/2009 at 18:56
|
I have knowingly never seen an ECU but from the pictures they look like a small and simplified computer in a rugged case. Since the major manufacturers produce several million mass-market vehicles each year I can't see any reason why they shouldn't be using very similar ECUs for all of them just programmed a bit differently. There must be sufficient economies of scale here to make these things pretty cheap even if they use entirely customised chips (which I really doubt).
As a software engineer, I would say that for an ECU, you are paying more for the software, rather than the hardware. I remember the thread that Rattle mentions, and it was pointed out at the time that there is a big difference in programming for PCs, where things don't have to happen in real-time and programming an ECU in a real-time environment, which is pretty tricky.
You're probably paying a R&D cost, rather than a "block of silicon chips" cost.
I agree, however, that the cost of a replacement ECU is:
a) Over-inflated
and
b) Often unnecessary
|
You're probably paying a R&D cost, rather than a "block of silicon chips" cost.
That makes sense, though I doubt the R&D costs per unit are particularly huge when spread over the number of units made. A bit of modular coding should ensure pretty high commonality across a range of cars, with smallish tweaks of basically standardised software to cope with the difference between, say, two diesels of difft sizes.
|
That makes sense though I doubt the R&D costs per unit are particularly huge when spread over the number of units made. A bit of modular coding should ensure pretty high commonality across a range of cars with smallish tweaks of basically standardised software to cope with the difference between say two diesels of difft sizes.
I agree. I know nothing about car ECUs, but I expect that there aren't all that many different ones around, and that the same make of car with different engines will use the same ECU.
|
|
|
I would have thought most the research is in the sensors them selves rather than the ECU itself which is simply a basic often 20Mhz computer with 16MB RAM (e.g a Circa 1994 PC with an expensive RAM upgrade).
I would imagine though that the biggest costs in research is the quest for that extra 1MPG or 1g less of CO2.
|
|
|
|
|
|
I agree with Altea Ego, so there's no point repeating it. I would never go back to tin boxes on wheels which rusted after 2 years and lasted about 80,000 miles before being scrapped. Along with carbs, points, chokes, distributors, condensors etc..
|
'I would never go back to tin boxes on wheels which rusted after 2 years and lasted about 80,000 miles before being scrapped'
So there's none of that these days then? I dunno about miles but it appears eight-year-old motors in the UK are now scrappers, to the benefit of the Far East motor industry and thanks to the generosity of the UK taxpayer...
|
|
Both the superminis in our household are over ten years old, both have over 82k on the clock, both going strong.
|
Re: ECUs, I agree their cost is completely disproportionate to their actual value, but how often do they go wrong, really? How many times are they immediately blamed for a problem, only to find it's actually a cheaper component or a wiring fault that's the real culprit? It seems to be an instant "fix" for clueless technicians or profit hungry dealerships who are unwilling or unable to find the real underlying problem. The Peugeot dealer I used to frequent with my 306 tried to blame the ECU for a repeated stalling problem. The actual fault was caused by a £60 stepper motor, which only needed cleaning anyway.
My 150,000 mile Volvo has electronic engine management, variable cam timing, drive by wire throttle, direct coil ignition, a turbocharger, CANBUS wiring and various other things that would make traditionalists back away in horror. It all works as intended, and has done consistently and reliably with zero maintenance since it left the factory 8 years ago, as of course do similar systems on almost all other cars. Correctly designed and installed, and given half decent wiring (the real cause of many issues on French cars IMHO), this stuff is nothing to worry about.
|
Since I am very old and have an elephant like memory I remember people saying EFI would be unreliable when cars grew older.
Spherical objects. Well designed EFI systems# just keep going on cars abused and rusting to bits.
Of course if you buy badly designed and built Mercedes with millions of electrical gadgets or a Renault with designed in design faults you will get electrical issues.
But any fool buying a second hand car can do some basic research and keep out of trouble.
It's rather like buying any electrical white goods made by Hoover. The reliability records speak for themselves.
Of course buying new cars is different. You get lots of useful things you really need like err I can't think of any.
# BL managed to route the cables on Montegos so water dripped onto the ECU...
And which car has a similar fault?
Edited by madf on 19/09/2009 at 17:09
|
It would be a bold manufacturer who, faced with known customer concern about noise and vibration from diesel engines did not fit a dual mass flywheel while all around were doing so.
One of the criticisms levelled against PD engines is their harsh behaviour - forum members would be complaining if manufacturers didn't fit dual mass flywheels, and calling them cheapskates!
I think that DMFs will sort out and become more reliable very quickly. For example, via an industry contact, I know that one manufacturer was caught out by not doing sufficient thermal analysis to realise that the reduced "effective" mass, and reduced thermal capacity of the flywheel would mean that clutch lining temperatures would rise to unnacceptable levels. They won't be caught out by this again!
As mentioned, most ECUs only fail in amazingly tiny numbers - of things to worry about this must be way down the list.
There is, however, a breed of mechanic and DIYer who, in case of a fault with a system, will always look at the most expensive, most complex part of it first. This has always been the case - mechanics around me seemed to be condemning far more carburettors and distributors than I could ever find serious fault with, but, I was clearing more blocked jets, and repairing more simple wiring faults - coincidence? I wasn't earning anything like the bonus they were though ;-(
It is true there are some ECU types which succumb to a common failure - for example some Vauxhall diesel ECUs had trouble with one soldered joint failing - these are a) the exception, and b) not usually repeated
I too have a long memory. I can remember people bemoaning the introduction of sealed crankcase ventilation systems - of course, engine life increased. In a similar vein, the more precise fuel metering of electronic fuel injection has given us another increase in engine life. All these "unasked for" improvements have given us supremely reliable cars - they don't bear any comparison to those we were running around in 40 years ago. I hope manufacturers continue to supply them.
Rose tinted nostalgia!
However, they can keep electronic handbrakes!
|
Does anyone appreciate how complex cars today are, and for the complexity and production volumes, how reliable?
Does anyone actually think about the components in a car, the kind of super-high-tech precision engineering that went into it, and is being produced at such amazingly cheap prices?
Having said that, HID is a rubbish idea. Just put in 90/100 halogen bulbs and tweak the circuitry a bit to handle that load ... worked wonders for my old Alto!!
|
Having said that HID is a rubbish idea. Just put in 90/100 halogen bulbs and tweak the circuitry a bit to handle that load ... worked wonders for my old Alto!!
Have you ever driven a car with HID lights? Personally I wouldn't buy a car that didn't have one.
Oh, and I suspect 90/100 lights are illegal...
|
Have you ever driven a car with HID lights? Personally I wouldn't buy a car that didn't have one.
Do they not sell them with two? ;o)
|
|
eight-year-old motors in the UK are now scrappers .......
The average lifespan of a car in the UK is 14 years. tinyurl.com/cw8azc
|
|
|
|
|
|
