Penn4S

Could not agree more with this statement. These cars have to be better than this for $100k+.

Jack F

ajskydiver

I just installed a dash cam today and fuse tapped into the Passenger footwell electric socket fuse. Not worried.

fastforddriver

$20,000 and $150,000 cars all use the same basic technology and they mostly depend on a 0-5v signal voltage range and the slightest change *can* cause odd behavior. I have no idea what's wrong with his car and I'd guess that 99% of the time there are no issues when tapping into OEM wiring but it is a risk every time you do. If you're just tapping into a power circuit I'd be surprised if that caused an issue but if you hit the CAN bus and maybe (as an example) tap into the speed signal wire things can go downhill quickly from there.

I have no idea what is wrong with his car obviously and I do hope they get it sorted out for him.

worf928

FYI, my background: 30-ish years designing and writing software and managing software teams in the area of 24x7x365 mission-critical systems ranging from factory control to embedded vehicle systems. 20-ish years repairing/restoring Porsche 928s for their owners.

So, basically 30 years finding and fixing bugs and coaching others on how to find and fix bugs and how to do things so that bug frequency is lower and how to do things so that bugs can be found in the field.

Pete wrote:
Kuma asked:
I responded:
I don't have 991-specific, or automotive (e.g. CAN, FlexRay, etc.) bus-system-specific examples. However, wiring is wiring and the bus systems have a few additional failure modes. And installation issues are installation issues.

The very short story: The kinds of issues Pete and I describe are not with the aftermarket stuff (AS for short.) The issues that provoke diagnosis are usually with something else. Diagnostics point to one or more non-AS components or sub-systems. When these components or sub-systems are examined evidence of problems with AS installation in the vicinity are uncovered. Note the use of 'evidence.' Without going back in time and being a fly on the wall one can never be 100% certain. However, Occam's Razor applies.

Intermittent issues are the most difficult to find. Efficient diagnosis first starts with attempting to make the intermittent failure repeatable at will. One must first characterize the environment in which an intermittent failure occurs and then attempt to re-create that environment. If you cannot provoke the failure at will you can never be sure if anything done to repair it has probably worked. If making the issue repeatable isn't possible - and sometimes even when it is - tests must be devised to gather data. Even in cases where failures are not intermittent, it is often necessary to devise tests to gather data to assist diagnosis.

Here's a synopsis of my most-recent example (not a 991 problem):

Owner brings me his 928. It has a serious battery drain issue (aka parasitic current draw). There is a simple process for isolating parasitic current draw for fused circuits. I determine that the interior light circuit is drawing current when it should not be. In this specific model year 928 (unlike other years), there are no single component 'silent' failures that cause parasitic draw. No light is always on. Nothing observable on the surface points to a cause. The interior light circuit is 'everywhere' in the car: doors, hood, hatch, etc. Damaged wires and components are not easily exposed in totality for examination.

For purposes of this exposition, I will skip the rest of the diagnostic procedure except to write that it involved black-box tests to gather data, making the problem repeatable, and determing the general physical location of the fault so that I knew what to disassemble. The location of the fault was the driver's door wiring harness.

Upon removing the door panel we - the owner was assisting in the process - discovered three stereo installs. The first was the factory install and the other two were AS installs. The second AS install didn't involve removing the wiring for the first AS install. Neither used the factory speaker wiring.

At some point the factory's routing of the door harness had been changed. This 928 had extensive service records none of which involved 'getting into' the doors. Thus, Occam's Razor suggested that one of the two AS installs was responsible for the change of the door harness's routing since the door panel had to have been removed to install extra speaker wiring.

In this case, part of the door harness had been stuffed into a cavity in the door instead of using the factory's (or anyone's) fasteners for securing it to the inner door-skin metal. Inside that door cavity is the window lift mechanism. Part of that mechanism is a thick crescent-shaped hunk of metal with gear teeth on the outer perimeter. Every time the window was lowered these teeth rubbed against the door harness. Over years, and who-knows-how-many window up/down operations, these teeth gradually cut through door harness's covering and into the individual wires. One of those wires was the power wire for the interior lights (which is constant 12v.) This cut wire was in partial contact with several other cut wires for other components in the doors. And thus, due to harness flexing, the power wire would intermittently - more often than not - sink current into other parts of the harness. I was mildly-surprised that this car didn't exhibit other 'spooky' behavior like the door mirror always moving up/down when the controls weren't touched.

There was nothing observably wrong immediately after the second AS install. It took several years for the AS installation issues to manifest as a seemingly-unrelated failure.

I spend many hours every single year finding and fixing things like the above on 928s. That's just the last gremlin hunt.

I don't think anyone has *insisted* that NM's gremlin's are due to the aftermarket stuff. If I read this thread correctly it's almost the opposite. Some questions were asked and now we're arguing about the extent to which gremlin's in general are possible, or not, due to aftermarket stuff.

I pointed out, a few posts ago, that the aftermarket stuff is - in my experience - not a cause of gremlin's but that it is instead, the prowess of the folks doing the installing that can be the ultimate cause of gremlin's. Properly, thoughtfully, installed it's not an issue. I have a mega-stereo in my '91 GT. It works great and after ten years I've no issues and no gremlins either. Of course, when I installed it, I was able to draw upon all the failure modes I'd seen and repaired and not repeat them.

This assumes at least three things:
1) The diagnostician is 'smart.'
2) The aftermarket device(s) were installed in a such a way that access to disconnection points is reasonably easy
3) Simple disconnection points actually exist.

In my experience the last two are rare. I almost always - when finding 'gremlin' issues - ask myself the question "WFT? Did the guy that did this think no one would ever, ever have to get in here again?" I have found aftermarket stuff installed in ways that required me to partially un-install it to get to something underneath or behind. Two years ago, I had to re-wire a subwoofer so that I could replace the 928's battery because the subwoofer had been wired and installed so that access to the battery box was prevented.


This is quite true and why these kinds of problems are not easily dealt with by dealers:
- work for repairs on in-warranty Porsches are only reimbursed if approved by Porsche
- dealer technicians are not always reimbursed for 'bug hunts.'
- nor will they (or the Dealer SM), generally, be willing to carry out an in-warranty repair without approval.
- Porsche doesn't make available sufficient 'system design' knowledge of, or tools for, the newer cars to enable a 'smart' motivated technician to efficiently carry out any possible bug hunt.
- On-board data gathering - i.e. logging - is possible, but the logs aren't very transparent.
- As of 2015 there were only three PCNA-employed persons authorized to interpret vehicle logs and use log data to approve warranty repairs. (Think about that: three folks covering all dealers in NA. This is, of course, why NM's car's been at the dealer for months: waiting for a Log Reader to approve a part swap or to order additional diagnostics. Then waiting for the part. Then getting pulled off NM's car by the SM to work on a higher-priority car....)

These last items I learned during a months-long 'bug hunt' on my wife's 981. There were about a dozen approved component replacements authorized by "Porsche's Log Readers." None of them fixed the issue - that I'd spent my time showing the SM how to make repeatable. In the end, I think one of the Porsche Log Readers threw a Hail-Mary and guessed right.

See above two responses to neanicu.

In general, besides motivation, in order to de-bug a complex system you have to have knowledge of how the system is supposed to behave and how it is designed so as to behave that way. With that knowledge you can consider how a failure mode might arise due to specific issues in hardware or software. If you can only treat a system as a black-box - i.e. you have little or no understanding of mechanisms that implement correct behavior - then it is very difficult to conduct a successful bug hunt. If there is no source of transparent operational data about a black-box system - no logs - then it becomes almost impossible to find a bug except through luck. (I've spent non-trival time over 30 years coaching software engineers how to writing useful logging code. It's a skill that comes from experience. It's not obvious.)

Porsche technicians are - to some extent - in the position of working on black boxes. They don't have available to them from Porsche sufficient knowledge of why the modern cars do what they do and the tools they have to gather data are poor. This is my big take away from my 981 bug hunt. We were looking at the log and I'd ask questions about specific log lines that were ambiguous: Does that mean this? Or that? It could mean either. More often that not the answer was "I don't know." And the reason was not because the tech wasn't smart. He was. And he was motivated. Porsche just hadn't made sufficient information available to enable him to use the log effectively. That knowledge was concentrated in PCNA's three Log Readers (and they, it seemed, didn't have enough system knowledge themselves to be efficient.)

A vast amount of function in modern cars is controlled by software. Even 'simple' things like oil pressure and thermostat opening are no longer purely mechanical. There's software involved. That software is a black box. It does things. We don't know what exactly. Neither do the technicians. Car companies - except Tesla? maybe? - haven't yet figured out how to turn their software into a white box that enables technicians to efficiently diagnose 'bugs.'

It's not the components. If correctly selected and installed. It's tools, knowledge and expertise. And the real question is "cause electrical issues that cannot be diagnosed or fixed in short time." Any bug can be found and fixed. The follow-on question is: who's paying for the time to find and fix it?

The operative answer to your question is: HELL yes. AS installs can and do cause problems that will not be diagnosed or fixed because no one is willing to pay to have them diagnosed or fixed. It happens. On rare occasions it's a built-in-at-the-factory gremlin not AS. Lemon Buybacks? Like NM's first one? In some cases cars go to the grave due to these issues because money and/or expertise is not at-hand. I've got plenty of 928 examples of too-hard-for-others-to-find electrical issues where owners were just about to jump off the cliff and get rid of their 928 before getting referred to me.

(And before all you rear-engine folks just say - it's a 928, their electrical systems are crap - in each and every case it was an issue caused by someone, not because 'it's a 928.')

Computing power doesn't matter if the wire to the computer is cut. Drawing power from a thoughtfully selected circuit will not cause problems.

991s (and the rest of the current generation) are a mix of discreetly-wired components (mostly engine) and FlexRay bus-wired components (mostly the stuff in the cabin.)

991 discrete wiring - e.g. a signal wire on a resistance-to-ground circuit to a temperature sender, or two wires to a switch (like a vacuum change-over valve...) - are no different from previous generations. They're pretty resilient as long as their conductors aren't cut or exposed and touching metal or touching other exposed conductors.

However, bus-based systems in our cars are just like a computer network. Wires aren't just carrying a voltage or sinking current. They are carrying complex signals. They are susceptible to noise caused by minor damage to the insulator (covering and inside) or at connections from conductor to connector. A piercing probe (something I never understood why anyone would use except as a last resort) can render a network cable unusable. Bending a network cable too-sharply can damage the conductor. Mis-routing a network cable so that it is too-near a source of electromagnetic interference can introduce noise and result in lost packets. Yanking on a connector rather than disengaging it can result in intermittent issues.

groundhog

I don't think this is the case and why should he be offered a GT3T? I see him as a symptom of where the brand is headed.

Also the "mainstreaming" of the product line has caused Porsche to worry about their retention rates, whilst still good, they're coming under increased pressure. The best ambassadors are the sports cars themselves. Unfortunately Porsche are trying to sell the Macan and Cayenne as sports car like (if you have ever been on an internal course they actually say this) adding further to the mainstreaming and gentrification of the brand in order to control retention rates and market share.

The reality is, I think for many, the brand as a sports car builder and the products it's producing are becoming less interesting. I suspect, the next Porsche I buy will be my last Porsche (probably a 718 GT4 or 991.2 GT3 clubsport).

Which brings me nicely to my initial point. Nick is a reflection of the current brand and its customer base but not a brand ambassador - for example I doubt Walter Rohrl or Mark Webber or Earl Bamber spend days on the next iteration of the configurator . (It has me totally beat as to why he does this along with the completely stupid and inept "street" pulls).

By the same token I hope his car gets sorted out and then again why does the car equivalent of munchausen by proxy come to mind.

dkhm3

This post makes a lot of sense- thank you for the effort in explaining.
It seems to me that the aftermarket installs done on Nick's 911 are the most likely culprit to his electronic gremlins. I had an S4 awhile back that kept saying my rear tailights were out, I kept on replacing the bulbs at the dealer under warranty- until they decided to look further and I saw the car basically dismantled and interior bits taken out to reveal that there was a small short in the wire somewhere causing the issue. It took someone a lot of time and effort to take everything apart to reveal the problem, I wonder if the Porsche dealer is even considering this as an option or simply throwing up their hands?

80p

smh at myself.

stumil

Bud Taylor

My SO called my 928 Christine...she was a very angry vehicle.

minthral

Based on my dealer service adviser, the don't get paid trying to find the problem...only when they're fixing it and Porsche approves warranty claim. When they don't know, they work with 'Porsche tech support.' Explains maybe why his car is at the dealer so long... its just sitting there as they go back and forth with Porsche, who probably has to consult parts supplier/designer. Cars are really complex today and dealer techs aren't trained/skilled enough or motivated to actually troubleshoot or even approved to do so. Even for a simple parts replacement they must jump through hoops.

Also explains why dealers are pain to work with when trying to track down intermittent 'I hear X sound but everything working' problems. They'd rather call it normal because they're basically wasting their time and its not worth it. If there is a clear failure, they got some meat to shew on. Honestly kinda screwed up system. In his case there are clearly check engine lights so its really surprising 'Porsche can't figure it out.'

neanicu

Get an OBD2 Jbox. Hook up your scope to the CAN signals. Get access to wiring diagrams. And start activating components,move electric seats,open doors to flex joints,windows...anywhere wiring is routed. Try to replicate the complaint. I guarantee the CAN signals get corrupted when the issue is happening! Is it a module that is corrupting the signal? Is it a short to ground,short to power etc? Do you have a temporary communication loss with a module?
Most technicians just follow manufacturer stupid flow charts that doesn't allow them to think.
If you suspect aftermarket stuff that is corrupting your signal,watch it in real time on the scope while going for a drive or try to replicate in the bay.

neanicu

I appreciate your contribution and knowledge.

About warranty time : even if you use programs like Alldata or Mitchell,you will see the same job that is paid flat rate...let's say 1.5 hrs,warranty is 0.8 or similar. Even if it gets approved by Porsche,it won't get approved to the actual time it takes : so technicians get screwed on warranty jobs.

Lack of information : YES! Big time! If they call the " hotline " they tell them to follow the flow chart like a robot- " measure here,unplug this,do you have 5V...if not replace this " You become brain dead after a while...

I bet you they'll just wash their hands on this one and buy it back. It's cheaper than pay for good diag. This is not an issue that gets diagnosed in 1 hr paid diag time.

Need to point this out : I do not imply anything regarding Porsche's reliability as a brand,stuff like this happens across the board. Nick's just unlucky perhaps...

ajskydiver

The 12v passenger footwell socket.

phm14

Agreed, and I for one hope he goes forward with the platform, onward into the 992, and continues sharing his experiences. Bringing problems like this to light should help fast track improvements, from the production floor in Stuttgart (assuming some of these issues are production/design faults), to dealer service departments, and benefit all Porsche buyers, promoting consumer confidence and continued growth of the brand. I tend to doubt that Porsche would be buying back cars if they could prove the issues were caused by mods...seems those aftermarket electronic installs would be top of the check out list...but what do I know?