It will be interesting to see how the alarm behaves if you reduce the belt tension to the correct setting. Over tension does the system no good at all and it is wise not to take liberties with such once you know where you stand. As I am concerned the margin for error lies between the mid point and the end point.

Last time, I filled my tensioner with STP- it was a right bugger trying to get the stuff into the tensioner- used my mityvac pulling a vacuum and nothing happened- in the end I force fed it in with the bleed nipple removed- it seemed to take forever. Hopefully it will stay in there given how difficult it was to get it in! Something thinner should be much easier of course.

Ok, scratch that. So we're back to what's happening, took the car out for a nice drive, no problems, went over 5k, light went on again. This time I was watching, it went on as I was accelerating, which means it's not going slack on deceleration (when the cams might lead the engine, hm).

Hm hm hm hm hm. Time to go back and look at that 928S trouble tree. We've verified tension, I will start going through the rest of it. Maybe there's a problem in the warning computer circuits...

I agree, will retension next weekend and back it off a bit. I also really want to take a hard look at exactly how it works, there has to be either something that is causing it to trigger or something that is causing the ground to interrupt. The fact that it seems to happen at higher RPMs and not when blasting over speed humps (I'm trying to see if it's caused by vibration) makes me wonder...

Then again maybe I need to hit some more speed bumps. Very low car though, don't want to ground out.....

Post #3 in the thread below talks about failure of the cam shaft caused by cracks in the spider that mounts the cam gear. This was preceded by multiple belt tension lights. I realize this is not common and the fact that the driver had been decelerating from speeds above 140 is also not common (at least to me)! But if nothing else shows up keep it mind.

Mark


https://rennlist.com/forums/928-foru...tensioner.html

Christopher,

If you are getting a repeat alarm you really should park the motor until you have got to the bottom of the problem. That the alarm is recurring with a similar pattern at high rpms suggests the alarm is genuine and given you know the belt is overtightened this should just not be happening.

That's a very good one to keep in mind, and I was wondering about something like that last night: Was something causing the belt to slack? However this one didn't pop on when decelerating, I was still accelerating when it came on (coming up around an on-ramp).

There is a pony in here somewhere. Hm.....

Ok. So this is not "normal" in any sense of the word.... :-)

Something is odd here. From what I can see we have two possible options:

1) The belt is somehow going slack at high RPMs enough so that the tensioner is moving in towards the center of the engine and the normally closed switch is disengaging

2) The belt is not going slack at high RPMs, the tensioner is not moving in towards the center of the engine and the normally closed switch is not disengaging.

If #1, how can I detect this using a second method. I can't run the engine without the top camshaft cover to watch the belt, the distributor cap is screwed into it if I recall. Is there another way to watch the belt and look for that slack? Some spot I can look with a phone camera or video recorder to see the belt in action?

If #2 I have a false alert happening somewhere. In which case I should pull off the wire connector and do a good hard look at it. If the timing belt alert light is triggered by a lack of ground on the signal line from the tensioner is there something that would indicate that lack of ground is coming from anywhere other than inside the tensioner itself? I think this is why Porsche says ground that tensioner wire and see if the light doesn't come on; if it does then something is wrong with the computer (which given the age of the electronics wouldn't surprise me).

This is a fun thought experiment. That said I'd like to have a bit of an answer before loosening the tension on the belt if possible: If it is #1, then I would expect loosening the tension would increase the chances of a belt jumping a tooth and resulting in the usual pull heads, lap and replace 14 valves per side, etc sort of problem.

Does a tool exist that would allow someone to monitor the position of the crank sensor and compare that to the spark pulse from a cylinder? If this was a cam changing position due either to a very oddly stretching belt or a crack somehow in the exhaust cam shaft then we should see it as a variance in pulses between the crankshaft and the camshaft. This is normal-ish with VVT, but the 928 doesn't have variable valve timing.

Another thought: I could put a scope and a bias signal (10k resistor to +12) on the stud that goes to the tensioner switch and watch it with the engine at idle and 5k RPM for spikes to 12v (the resistor would limit the current to the switch to a very small amount when closed, but if it opened it would show as a spike on the scope screen). Anyone ever done something like this?

Hm.

A lot of confusion with these pics and why would you practise at random positions of the crank? the pic I'm referring to says "the tool in place"

There is one pic that might be at 3 o'clock, it is not definitively framed; but bottom line I would recommend getting John Murphy's spiral bound tutorial (ebay) on the belt or search here for other step by step guides here. The shop manuals can help with things like the alignment of washers in the tensioner. Dare I mention the Porkensioner??? The original system works well, but all the pieces must be properly installed for it to function. As you can see it rather complicated.

You are in DC, you should make it Frenzy in Sept and get some clarity on this. Good luck.

Ok, I can see the confusion. That was a clear shot to show the tool in place on the belt while I was practicing. I wanted to ensure that I had the tool slotted properly on the belt, without nicking it or anything and low enough to ensure that I was getting an accurate reading.

I like taking lots of pics and it's helpful for me to check things at multiple points to look for patterns.

Note also my hand is resting on the distributor rotor in at least one of the pictures, it makes a handy handle while leaning in to work. :-)

Yep, will be there. Want to bring the car as well, I'd really rather not show up in a 944S. Although I could do a timing belt demo there on that :-)

Bigger question here is why is this happening. Either the belt is loose or it's not and I'd like a second way to see what the belt is doing while the engine is in motion. This evening I'm going to unplug the cable, test it to a known good ground, then clean the connectors and ensure they are snug. Then if that fails, we check tension again to see if there are any changes. If tension is good it's time to bring out the oscilliscope and figure out what's going on.

If one stands back and analyses the info presented to date the following seem to paint a bit of a picture:
1. Your belt has been running too tight
2. You are seeing a consistent alarm pattern occuring at high rpms.

So what does this tell us?
1. This suggests that the alarm system is working
2. Something is causing the earth signal to go to an interupted condition- two possibilities can cause this - something is faulty or tension is genuinely down when the motor is at full load.

The tension you see during a cold calibration is inferential given the belt is not being loaded. If the alarm is genuine then this would suggest that something is deflecting abnormally under loaded conditions thus the alarm is being generated. Belts that are too tight have been known to cause problems. Camshafts have been known to snap at the nose.

If there is a genuine condition causing your alarm without wishing to sound alarmist it would probably have to be something that has potential to cause a lot of damage thus why when one sees an alarm going off more than once to continue running the motor is tempting fate.

Hopefully this is no more than a false alarm but given we know the alarm system in general seems to be working, what are the chances this is the case in this example?

Sad to say if the belt has been running too tight there is a possibility it has induced something to start failing and equally possible it does not show under the static calibration check but under full dynamic load something is stating to flex causing a loss of tension.

Greg Brown has also advised that loss of oil in the tensioner can cause alarms.

Chance favours the prepared mind!

Inconsistent, it doesn't happen every time. Most times no. Occasionally yes, this last time I caught it after driving for a bit (yes, over 3 minute drives). Likewise the belt looks to be a hair tight, I've personally dealt with a set of cams that sheared teeth off the chain gears and broken several cam tower caps without damaging the shaft; those things are pretty thick. It's possible Porsche left no margin for error, but given that the slack varies more than the error amount over the revolution of the engine... Maybe.

I'm not convinced yet on 1: The light comes on but then again the TAIL light comes on as well if the right side bulb isn't perfectly shimmed. These are analog computers in a lot of ways, I'm *guessing* that the signal is an op-amp watching for a transient across that line to the tensioner with a +12 fed to it with a 10k resistor or something as a pull-up. Line opens, voltage goes up, op amp goes to an SCR and on comes the light.

2. is the trick: Assuming the computer is working, what could cause a transient on the line high enough to trigger the SCR gate? Well, I have a thought but more on that in a bit.

Agreed, that's 12 out of 16 valves bent per side. Which is not anywhere near a worst case scenario (that is a chain ramp failing, causing complete destruction of the head) but would take a while to fix. Then again it's do-able and at *least* I would know what's wrong :-)

I should test to see if the alarm system is working. See my next post for my attempt to follow the Porsche schematic and test that.

One person said that, another said that loss of oil would not cause an alarm. It's a bit tough to sort through everything, need to take it one step at a time and change as few variables as possible.

This will be figured out. But at least this is new territory. :-)

In the meantime I went out this afternoon to unplug the tensioner and see if I could test the wire back to the computer and make sure it's not intermittent.

I failed miserably. Trying to unplug the wire did not work, and trying to unbutton it from the plug got me this:


https://cimg6.ibsrv.net/gimg/rennlis...7ddeaa96e0.jpg

Well, it's plugged in at least....

It does not appear to come off by sliding out as some instructions have said. Nor is it a spade connection, was it different on 86.5 engine models?

Pushed it back in, but then took a picture of it back in place and went hm....


https://cimg4.ibsrv.net/gimg/rennlis...c3001c2473.jpg

What is wrong with this picture?

If you look closely you will see the wire to the dashboard with two spark plug wires sitting right on top of it. Because the spark plug wires were not screwed down and have probably been routed very badly. I wonder if those plug wires sitting on top of a signal line like this could induce a voltage large enough to be picked up by the computer. It would have to be a lot of sparks, like an engine running at high RPMs...

Hm.

That your alarm is activitating suggests the system is healthy but whether it has a fault potential remains to be seen. The alarm simply relies on a continuity to earth- if this continuity is broken the system alarms- what is known as a failsafe system in instrumentation parlance. Thus transient spikes should not cause an alarm.

I did not say your system alarms every time - I said it alarms in a consistent pattern- i.e. high revs under load or so I understand from your posts.

That connector sure looks awfully grotty- high revs = more vibration possibly causing discontinuity?

Oh- and Greg Brown is probably the most knowledgeable man on the planet when it comes to the 928 drive train.

Got it. However as soon as a transient happens the computer will lock that in via an SCR (a device that will normally not pass current till the gate trips, then pass current regardless of gate state until powered down). So even a momentary disconnect will cause the light to come on and stay on.

Any idea how that connection disconnects?

C

Time to stop guessing and tear it down. Best advice I can give you.