phendriks

Hi All,

Seen the warning light today (the horror), just as I was shutting down. Restarting hasn't caused it to come on again, although I haven't brought it back on the road. Question is, is there a scenario that will trigger this other than loose belt? Should I wait to see if it comes on again, with an immediate shutdown? Have had absolutely no elec gremlins before. Belt is 5 years old, 11 thousand kilometers on it, checked tension couple of seasons ago. Or is time to get the Kempf tool out?

Car is a '93 GTS auto.

Cheers, Phendriks

OTR18WHEELER

Just a wild guess here, im sure others will have a full page of contrary suggestions.
..! Light coming on at shut down could be due to excessive belt slap if timing were off a tad bit.
Like, if it may have backfired due to valve timing, but the knock sensors detected this and shut down the offending cylinder bank at the same moment the ignition was turned off, and that unspent fuel got a little spark.

FredR

It is a bit un-nerving when the belt alarm goes off but as with any instrumented function, temporary glitches can and do occur and the toothed belt warning system is not exempt from that. If the belt is in a genuine alarm condition then one might expect that the alarm will not reset and if it does, logic suggests the chances are that it is a spurious alarm.

That being said, prudence suggests that some investigation is warranted and if you have a Kempf tool then by all means pop the cover, check the tension and take a look to see if there are any signs of distress. You can also take a look through the vent holes to see if the belt appears to be tracking correctly.

There should not be any issues with your belt at this point in its life cycle but other factors can cause distress to it so prudent to ensure nothing untoward is going on.

In 18 years of ownership I have had two alarms on the annunciator and both times I accepted the alarm and the light went out never to return. As this invariably happens when you are out on the road what do you do? Both times, after accepting the alarm [that disappeared] I checked the vent holes that suggested nothing was amiss. I then drove home and checked the tension - found nothing wrong.

If there was a real world tension problem my expectation is that the alarm would remain- anyone who has had a genuine problem would experience such or so I would think. If the alarm does not cancel then shut the motor down and flat bed it.

Alan

There is an approx 3 min delay before the sensor is active - so if you ran it for less than 3 min you wouldn't have seen it. Get it in for maintenance ASAP or evaluate further DIY. I'd evaluate the belt though the top air ports and see how it looks (belt condition & tracking on the sprockets) before starting it again.

Alan

Mrmerlin

check the belt tension cold and at TDC,

dont run it till the belt has been checked.

Disconnect the battery now

FredR

When I posted to this thread yesterday I did so to see if I could stimulate some discussion. I half expected to find a pile of posts this morning trying to pull apart what I posted yesterday and all I saw was a couple of posts from Alan and Stan offering understandably [conservative?] safe advice.

The 928 is meant to be a long distance high speed GT and in a place like the USA the designers would expect the car to be far from anywhere in normal use. During the heated debates about the PKT tensioner this device was criticised because "it did not have a belt alarm". During the debate it was mentioned that the 928 was the only car at that time that had such an alarm. Ken's point of view was that as his tensioner progressively took up tension no such alarm was needed on his system. Thus the question "what is the alarm intended to warn of?".

Now we see in this thread the OP uses the term "horror" to describe his sentiment when he saw this alarm doubtless conditioned by the thought of what damage could happen if the valves clatter the pistons. The question I would ask is "How realistic is this"? Did Porsche intend this alarm to be a harbinger of imminent self destruction or did they simply intend it to be a cautionary advice? When the alarm initiates the graphic states something like "toothed belt service"- it does not state "tough luck your motor is about to grenade". This tells me that Porsche's intent was that the owner should take the car to the nearest dealer as soon as reasonably practical. Does the handbook advise the owner to pull over and call a tow truck? For sure Porsche never intended for DIY inclined hacks like ourselves to stop by the side of the road and do some on the spot analysis of the situation.

Doubtless we can dream up catastrophic failure scenarios such as water pump bearings seizing and the belt slipping over the seized pulley causing it to overheat, stretch and then fail but if Porsche thought this was a likely scenario would they announce it by "toothed belt service". I suspect their intent was to cover things like initial stretch [covered by service instruction to check new belt tension after 1000 miles] or more likely, unusual belt stretch.

Thus I revert to my original postulation- if an alarm goes off the first thing to do is accept the alarm by pulling the stork and if the light goes out, surely this means that a transient [false] condition has corrected itself? If there was a genuine lack of tension in the belt surely the alarm would not cancel?

If there is a sudden and catastrophic failure the alarm is academic and valves on the interference motors are already bent. In the process industries we have two types of alarm- a pre-alarm to warn the operator that something is not quite as it should be and a trip alarm that automatically shuts something down to prevent damage. It seems to me that the toothed belt alarm was intended to be a pre-alarm but somehow [folklore?] has "evolved" into a trip alarm.

I see a conundrum here- any thoughts?

phendriks

Thanks for the input guys. For the peace of mind, I've pulled the cover, initial Kempf test is in the window, but I still have to spin it to TDC and do a proper test. Took an hour to get it this far, not worth the worry. Plus I can swap the plugs WYAIT.

Fred raises a valid point. But if I nonchalantly accepted this as a pre-alarm, I still wouldn't know the cause/decay rate to trip alarm status. How long would you ignore it? Unfortunately, there is no trip alarm other than bent valves.

Cheers, Phendriks

FredR

The main point to note is that the TB check has to be done with No1 cylinder on TDC so make sure you have the correct stroke engaged.

The main point I raise is whether there is an issue of concern if the alarm cancels upon pulling the reset stalk. I suspect that most owners see the alarm and more or less in panic mode immediately turn off the motor [in your case an alarm showed as you turned off the motor -correct?] such is the perception of the belt alarm. If the alarm cancels and does not reappear I take that to imply a flase alarm but it does not stop me from checking things when I get back home. If the alarm remained on after resetting it, the 928 would be on the first tow truck out of there.

SeanR

When using the Kemph tool your belt tension needs to be at the tight end of the window. Any less and it will trip the warning light.

Alan

Fred,
There are two general causes of a warning - 1) the belt has become a little loose as a result of gradual stretching and has just tripped the tension warning dynamically and it may not do that again for a while, this indicates as you say an early warning to take care of it eventually and in general it won't likely trip again for a while anyway. An evaluation is in order ASAP within reason and appropriate action then. If it cancels without reappearing while running I think what you suggested is true.

However option 2) is that something potentially catastrophic has just started to happen and you are in major jeopardy. Given the occurrence on shutdown its not easy to disprove this possibility so an evaluation of how the belt looks and tracking on the sprockets seems a more urgent thing here than in case 1).

Given the ~3 minute timeout before tension monitoring starts - you have 3 more minutes of jeopardy after starting before you get any more indication.

WRT the Porkensioner I have installed a tensioner warning configuration for this on my GTS that will trigger for over-extension of the Audi tensioner piston (running out of tensioner range due to excessive belt stretch) or if the tensioner piston fails to maintain spring pressure within its operating range (e.g. binds up, spring breaks etc). This is described in another thread and was from an initial idea Ken proposed. Seems to be working well (no false warnings). An easy and worthwhile feature for any future PKT install I think.

Alan

FredR

Alan,

Agree with your above mentioned observations but the problem I have is how one should interpret the situation the OP experienced or indeed any alarm that occurs on a well maintained example.

As I understand the OP was running his car, pulled up, switched off the ignition key and as he did this the belt alarm appeared.

Shutdown is a transient condition, the fuel is cut off and so is ignition- the crank is still spinning but there are no power pulses whllst the motor runs down so the transient belt tension profile will change. Normally there is no reaction but if the belt is a bit down on tension I can visualise how switching off might just be enough to trigger the alarm. The OP sees this and quite naturally is "alarmed".

In such circumstance I cannot think of any reason why one should suspect a catastrophe is about to happen given the motor has been working seemingly OK until the point when the motor was turned off, especially if the belt/water pump system has a credible maintenance history as this one did. So, what should the owner do? I would take a quick look down the cam cover vent pipes and if there are no obvious signs of distress and the belt is sat snugly where it should be I would fire it back up and let it run for a few minutes whilst still warm to ensure the alarm does not repeat.

Whereas there is an understandable sense of paranoia regarding belt failure in our community, in reality how often if ever does this actually take place? Doubtless it happens but not long ago it struck me just how little we actually hear about such failures.

My take on the stock tensioner is that it is somewhat bullet proof providing the rollers/bushes are up to scratch. I suppose the lock nut could be left slack by mistake and then the tension bolt might work itself loose but that would eventually generate an alarm and stay there. Thus my theory that if we see an alarm and resetting it causes it to disappear, why should we be overly concerend as a true alarm brought about by lack of tension cannot just "go away"?

If an alarm goes off, is ignored and repeats then it should be no surprise if a catastrophe occurs.

Whilst on the subject congrats on the PKT alarm- that can only add value to the original PKT system

mj1pate

I suspect that Porsche related the tensioner warning to an "early warning" system as long as the belt was maintained as required in the owners manual,
Differences in our cars vs the 928s referred to by Porsche's owners manual might include that our water pumps may not be new Porsche pumps. Ours may be Laso or rebuilt "other". Maybe the new Porsche pump shafts and bearings *never* failed in a catastrophic way within the lifetime of the timing belt. I certainly don't know the answer to that proposition. In any event, the word "aged" applies to many of our cars and that alone prompts me to be more concerned about signals from early warning systems,

Yes, my tensioner light came on several hundred miles after belt installation. Reset resolved it during the ride home and Kemf retention long-term resolved it once I arrived home.

M. Requin

Do you know the brand of the belt? Consensus here seems to be that Contis experience more belt stretch, ergo more warning light incidents,

davek9

As stated the built in 3 min delay could be deadly in this case.

Look at the belt tracking (static), it should be a few mm's from the front edge of the gears on both of them.

If one side or both are on the front edge, do not run the engine, at least I would not as it could be a failing WP or tensioner bearing.

The test you did is meaningless, as stated it needs to be done cold and at top dead center, belts just like chains have a "tight" and a loose spot in their travel.

Dave K

FredR

The OP opened up this thread with a couple of very reasonable questions quote:
" is there a scenario that will trigger this other than loose belt?"
"Should I wait to see if it comes on again, with an immediate shutdown?"

The stock answers for sure are fail safe but for an owner who has no mechanical abilities are these answers reasonable advice given they imply a tow truck is needed? In a litigious society such response may be the only practical advice.

However, for an example that has been well maintained is it appropriate to suggest that the owner calls a tow truck and takes it to whoever? What realistically can happen that warrants not running the motor for a few more minutes bearing in mind that the car has just been out on the roads for whatever time, possibly having a few short blasts and having spent say 30 minutes on the motorway at 80 mph or whatever. The owner pulls onto his driveway, the car is idling, he turns it off and as he does the alarm comes on for the first time in his ownership. How realistic is it that in that specific split second something untoward has been manifesting itself such that the alarm triggered at that moment and the belt is now likely to let go in the next few seconds of operation? Has anyone ever seen or heard of an alarm going off and a few minutes later the belt snapped? If a water pump shaft were to seize would it do so without warning or would it be preceeded by loss of coolant due to seal failure and then the bearing being washed out of grease or whatever it runs in [i.e pre-cursors]? If the pump bearing was failing or had seized would it do so quietly or would it make a racket?

In designing systems engineers are faced with many dilemmas- there is the consideration of "what could happen" and more importantly, there is the consideration of "how likely is it to happen". It does not take an engineer to figure out the former but it takes a damm good one to figure out the latter. That we have the "luxury" of an alarm is one thing, that we have the potential to mis-apply it is another.

I do not know the answer to this conundrum and there may not be a definitive answer but perhaps we can think a little more about this.

As an adjunct to this thread Alan pointed out the 3 minute warm up period prior to the alarm being armed. Does anyone know for sure whether it is timed or whether it is conditioned on engine temperature that coincidentally takes about 3 minutes to warm up? Given the inputs are available I am damm sure if I was designing this system it would be temperature driven [80C?].