Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Newer Vehicle Electronic Maintenance Nightmare!

Status
Not open for further replies.

bill190

Electrical
Mar 23, 2009
26
0
0
US
I'm learning about the electronic systems in newer vehicles and am SHOCKED at how this is designed! (In regards to expecting automotive service technicians to be able to troubleshoot these systems.)

In some cases you have 24 different electronic modules connected together via two different local area networks and a gateway module to connect the two networks.

AND in some cases, certain modules need to be programmed with certain software so they will work properly with the other modules AND have certain set-up procedures performed after installing the new software so the module and other modules will work properly.

And one or all of these procedures can be changed/updated as problems are found, so older documentation can quickly become wrong.

AND new software releases may require that electronic service tools and computers have the latest software versions installed for performing all this.

And they expect your typical automotive maintenance technician to do all this. Needless to say, this requires a LOT of reading and knowledge of how these things operate.

I've spent my life working in electronics and in the computer field and troubleshooting these systems would be difficult for me. And I could do this, but this would take quite a bit of time doing a LOT of reading as to the latest updates and procedures.

This is like asking someone to replace a video board in a desktop computer and install the correct new software driver by finding it on the internet and selecting the correct driver.

Now to make things even more complex, throw an intermittent problem into this mix! Say one loose LAN terminal on a connector right in the middle of the network. Or a loose terminal to one module.

It seems to me vehicle manufacturers would save a LOT of maintenance headaches by combining most or all of these electronic modules into ONE large electronic module. Then so far as cutting down on the wiring to this module, use "satellite" I/O modules each on its own network connection wire. Maybe sort of like a cat 5 wire to each. And these would not require any setup or software installation if replaced. All I/O units would be the same. Intermittent problem, swap out the wire and I/O unit. Simple!

If the network went down, the main unit software could disconnect each network connection one at a time, find the problem, then isolate that network wire/module from the rest of the vehicle, then set a diagnostic trouble code: The problem is here ->. Easy!

The only software installation/upgrading would be done to one unit. This could be done via a connection to the factory. Just like Microsoft upgrades software to your computer. Check for updates. Simple.

All settings would be on file per the vehicle VIN. No setup required.

Also it has been my experience that "corporate human resources" types will gear their training toward less intelligent types. The thinking is that anyone can be hired and trained to do the job, thus easier to recruit new employees and pay scales can be kept lower.

However with the current wave of electronic systems in new vehicles, highly skilled technical types will be needed to troubleshoot these systems.

This would be someone like me. And I will say I am "bored" by the training videos produced by these manufacturers. It goes too slow. I fall asleep and daydream when trying to watch these videos. I learn best by reading very technical books on the subject. I can skip forward and get right to the point. I learn best on my own, not in a school. And I know this is quite common in the computer field. There are tons of computer gurus out there who have never taken one class in computers. They learn on their own. They just need access to technical documentation, then away they go.

So my advice to the vehicle manufacturers is that if they want to attract the computer guru crowd to work on their vehicles, they would be best to start publishing detailed technical books which explain how all these things operate.

For example I found a book titled: Fundamentals of Computer Command Control by General Motors Service Technology Group. This is the type of thing which would interest me. But this book is hopelessly outdated and there is nothing new out there to replace it.

As the robot said in the movie Short Circuit; "Need input!"
 
Just recently, I've been tasked with investigating the reasons why engine control module suppliers reject warranty claims as 'TNI' - trouble not identified.
These TNIs are on modules returned from dealers that have gone through approved diagnostics routines resulting in indications that the engine control module was at fault.

I've been plugging away at diagnostics files and I regularly see CAN (and I suppose LIN) bus problems where there is no communication going on or where modules do not communicate with the diagnostics equipment.
In some cases, this seems to be because, although the latest engine modules may function down to 5V, the CAN system sometimes stops working at 8V, so there's a 3V 'gap' area where the engine can be started, but faults may be logged in the ECU because the CAN system hasn't woken up.


Bill
 
Bill,

Do you routinely come across vehicles that start with battery voltages below 8V? [ponder]

Dan - Owner
Footwell%20Animation%20Tiny.gif
 
Yes.

I've been working on FMC European vehicles which use geared permanent magnet starters(as most/all vehicles do these days, I think) where the CYCLING battery terminal voltages during a 20ºC start can, and do, go below 6V and the engine will still start.

Using PicoScope equipment, connected directly to the battery terminals and the ECU power supply, as close as possible to the ECU and the EC~U power relay, a colleague and I have been looking at a range of vehicle/engine combinations searching for instances where the power relay feeding the ECU drops out before the ECU reaches the limit of its capabilities, voltage-wise.

We deplete the battery by periodic dead-cranking and then go through a series of starts, monitoring the power relay, ECU and battery terminal voltages, until we get to a no-start.

The capability of the starter to actually turn the engine is still good at this point. In fact, for those of us familiar with a knackered battery on a carburetted vehicle of, say 30 years ago, the engines are typically turing over at speed that you would have been pleased with on that old carbed motor. Many vehicles were still cranking at 120-ish RPM, but not starting because either the ECU power relay could not hold in, or the ECU was re-setting when the cycling voltage approached 5V.

What I found interesting was the adoption of 'industry standard' CAN chips that could not meet these same voltage perfomance parameters and so stopped talking once about 7-8V was achieved, thus apparently throwing up diagnostic errors. It's a good example of non-system thinking.

Bill
 
Interesting... CAN specs are 5V nominal for a high, and every interface chip I've ever come across was designed to operate either on 5V or 3.3V. When you say "industry standard", is this a specific company manufacturing odd-voltage CAN interfaces?

Dan - Owner
Footwell%20Animation%20Tiny.gif
 
I'll find out. The CAN issue arose as a sideline to the original investigation so I'm not altogether familiar with tit, to say the least.

Bill
 
Right, I asked my local specialist and I had it a bit wrong. Let's suggest that this is a Bosch system. It is LIN function that's impaired below 8V. The CAN controller on-board the ECU is supplied by one of the regulators on the ECU PCB. This particular regulator is 5V and needs about 5.7 to 6V supply to the ECU to remain stable. The CAN regulated supply is 2.8V.
Other functions of the ECU take further regulated supplies of 3.3 and about 1.6V and may be viable after the CAN is disrupted.

Another item to observe in terms of diagnostics is the various voltages at which engine sensors may start to throw up diagnostics codes.
On one engine, the MAF will not operate much below 8V, for example. This particular MAF produces a PWM output to the ECU which stops if its supply falls below 8V.




Bill
 
That makes more sense. The CAN interface is obviously 3.3V (with a high signal being 2.7V+) and will remain stable even when the CPU regulator is out of range. 5.7V for stability of the CPU's 5V regulator is reasonable, so even at a battery voltage of 6V you should still be okay.

It's possible (and likely), however, that at such a low battery voltage the CPU's regulator (and even the CAN's) is falling out of regulation when a high current draw item is turned on (starter motor, lights, etc.).

Dan - Owner
Footwell%20Animation%20Tiny.gif
 
Bill, if you are really shocked at the idea of 24 modules on two communication networks, imagine 104 (nodes) on five networks, with two subordinate gateways and a master.

Now imagine "trouble tree" type diagnostic routines created by an engineer who adds a fault to the vehicle, and then observes how the system reacts, and then starts writing the diagnostic table, while knowing what fault he/she added to the circuit! Now imagine the engineer is sure that the technician that will have to follow his/her trouble tree has little to no experience in electronics so that he/she must ensure that each step will work universally. That's how we get trouble trees that send the technicians out on extended wild goose chases, where they have to do a few dozen tests including the infamous ones of check continuity on circuit XXX, and YYY instead of simply giving the techs the real information needed to diagnose the problem.

At one time Ford had a problem with EGR valves, where the position sensor voltage was too high at rest. The problem was that the valve could be held open by chunks of carbon. Pretty simple diagnosis with the correct information. Go and find the trouble tree for EVP sensor voltage too high, and see how long it takes you to even mock perform each step. Oh, and see if the chart actually ever tells you what the sensor voltage range should be (code set criteria) and if it ever leads you to the valve being physically stuck open.

Now back to the original question...

A big assist for technicians when having to diagnose problems like this is for the scan tool itself to connect to more than one bus, and even be able to be the bus master if necessary. But few manufacturers have done this reliably. On a recent lightning strike victim (2004 Saturn Vue) there were modules damaged to varying degrees on each of the three communication platforms, and the BCM which was was on all three networks was one of the modules that was damaged. There are no real world directions for us to follow as technicians to diagnose that car, we literally have had to teach ourselves how to deal with problems like that because even the scan tool was of no use until some type of communication from the car to it could be restored. That of course was only accomplished with an oscilloscope , the wiring diagram and product knowledge.


WGJ. Picoscope is a great tool, and there should be one in every technicians tool box. Yes, you are correct in that supply voltage can dip below 6v, cause communication drop outs, and then have the source voltage rise sufficiently with the engine cranking speed that we and have the car start at voltages below what some of the modules require for operation/communication. GM's will routinely set an ABS code for a weak vehicle battery/charging system (at <11v). Chrysler vehicles will crank happily away at 8v, but won't start below 9v. Where you run into your issue on CAN voltages, keep in mind the difference between what the CAN portion of the module requires to operate, VS what the module itself needs to perform the task for which it is designed. Also different modules in the same vehicle will have more or less tolerance than others which means the drop out points are all different, so long as we don't lose the bus master.

 
Forty years ago, and I suspect also today, documentation for new cars, including fault trees and (dis)assembly procedures, was done on brand new cars with no corrosion, dirt, or problems other than synthetic ones induced on conjecture. So far as I know, they have never sent a service manual writer, or a design engineer, to a junkyard, a local garage, or even a dealer shop.

Electronics made the vehicles more reliable, maybe by an order of magnitude, but they didn't change the isolation of the design process from the service process.



Mike Halloran
Pembroke Pines, FL, USA
 
Hi Mike. You are absolutely correct, electronics have made vehicles more reliable. One unfortunate side effect though is the randomness of the problems that we do see. Not all that long ago 99% of the diagnostics on a vehicle that visited my shop last week took little more than experience to handle. A cranks but won't start, no spark, started with a check of the gap on the points, and then a quick crank to see if any arcing was visible.

Also last week we dealt with a vehicle with a defective damper pulley CKP vane assembly. It had a good 3x, and 18x patterns from the hall switches, but the timing of the waveforms caused the module to not fire the coils. This was diagnosed by comparing the waveforms between this car and from a vehicle that ran that was provided by another technician. The "trouble tree" was useless to the point that it would have resulted in the technician replacing the ignition module every time.

The isolation of the repair and design processes is a very good description of why we see the problems that we do in our end of the automotive industry. Chrysler pulled one trick with it's transmission control modules that can serve as an example, this BTW is by no means the only one or even the worst it's just the one I choose to describe.

Imagine an intermittent circuit fault with the 2-4 solenoid in a transaxle. With no delay, the moment the TCM recognizes a fault with that circuit it would turn off the transaxle power relay. Diagnosing that fault was akin to trying to figure out how a magician just tricked you. By the time you realize that something happened, the relay opened and you only had the result of the failure, immediate fail-safe. The trouble tree's were fine for hard faults, and they could assist a technician with them. Any trouble tree that included the words "substitute known good part" should have been viewed as a disgrace for the writer. JMHO.
 
The biggest problem period with any and all electronics in automobiles, well actually any vehicals and heavy equipment, is the total lack of technical information on the "BLACK BOXES" no circuit descriptions, no schematics, just plain no information what so ever. It is very ignorant to make a device that is to be serviced yet offer no information on how to. If technicians had this kind of information it would make the job much easier.
But then thats part of the reason for the electronics, when the box is suspected or is bad, just replace it, at what ever astronomical cost that is tagged on it.
 
I suspect that once any given circuit terminates on a circuit board, further detail in the schematic becomes worse than pointless as the "black box" in question wouldn't be readily repairable on-site. Just more clutter to get through to get at the stuff that can be repaired/replaced.

What might be worthwhile for a particularly sharp and determined DIY'er to diagnose and repair (at $0.00/hour) frequently isn't for a flat-rate shop tech who's in the business to pay his bills. Even that assumes that individual internal component replacement is realistically possible.


Norm
 
The issue is not the info is missing to repair the internals of the computers, it's that there is often not enough info to determine if the computer or some other component is at fault. No info on what the expected output Y is when the computer is given input X.

But then, I feel the biggest problem in vehicle servicing is the lack of troubleshooting skills displayed by many so called service techs. It would be better to give a detailed description of how the system functions than a troubleshooting chart. If the individual can't use the description and wiring diagrams to determine the problem then they should not be attempting to fix the problem.

If someone does not understand what the system they are fixing does they how can they fix it?
 
As a tech we can test pretty much everything there are no gray areas just in the diagnostic equipment if something acts peculiure than we program the vehicle or affected module but everything has an expected reaction to a certain input.. most techs wont pinpoint and fix specific problems for liability sake... its always safer to replace the system than specific components.... because when it comes down to it people dont pay on repeat visits technicians do..
 
"most techs wont pinpoint and fix specific problems for liability sake... its always safer to replace the system than specific components"

Are you telling us that when I take my car into a authorized shop, with certified technicians, with an intermittent 'no stsrt' discrepancy, and the ECU gets replaced & paid for, maybe $1000.00, and two days later the same discrepancy rears it's ugly head; that the tech somehow "pays" for the rework? I think the customer now owns a new ECU AND an old 'no start' intermittent problem.
 
In most reputable shops the customer will have that issue resolved and the technician or that service department will eat the cost.... In certain situations concerning your safety... or mechanical damage to your vehicle.... it is imperative to make sure that the problem is resolved. considering for instance BMW restraint systems which use fiberoptics, with one central control module and several satellites i would rather hand you a quote for that system or part that system which is affected than have to attend your funeral... i would insist that people get to know there service department and ask questions... think of us as your doctors.... These cars are getting more complex and require an extreme amount of patience and knowledge...
 
ttf's point is that the current attitude towards troubleshooting often devolves to 'fit known good part, charge customer, see if it fixes problem'.

Which is more or less why an 8 year old BMW 7 series is worth less than a Camry

Cheers

Greg Locock

I rarely exceed 1.79 x 10^12 furlongs per fortnight
 
I am also referring to the use of complex technology for it's own sake. The "My Bimmer has fiber-optic seatbelt control modules" banter, sitting in the club bar after playing 18 holes will get you an approving nod & wink from your buddies, but soon the Piper must be Paid. I've been told BMW has a strict ethanol/gasoline that is performed on cars entering the shop. Any sample outside BMW's limits voids the warranty. Any truth to that?
 
Status
Not open for further replies.
Back
Top