Porsche 944 timing belt system 2

[RossABQ]

I've recently purchased one of these cars, an '84 normally aspirated model. These engines are famous for eating their valves when the belt breaks/slips/feels like it. Searching on-line I've found instances where qualified techs installed a new belt and idlers (all the bearings and belts are sold as a kit) and the engine ate the valves as little as 9,000 miles later, at idle, in the owner's driveway. It happens so often that the most common mod to these cars is to install a Chevy V8.

Do any of you in the industry have any insight as to the problems with this system? I see nothing unusual about the system, if anything it seems to have lots of wrap on the pulleys and good support along its runs.

Porsche subsequently went to chain drives on the turbo models of this engine, and they seem to have their own set of problems.

 

[JohnLear]

It's odd that Honda / Toyota (et al) can get over 100,000km reliably from rubber timing belts. Years ago Fiat (and I assume others) was getting 20,000km (miles?) reliably (and my undertsanding is that those old belts weren't nearly as good as more modern ones).

I wonder if the Porsche belts might live (and die) in a hotter engine bay?

 

[GregLocock]

For a start there are degrees of interference. The engine on my Toyota ran for two years with a one tooth error in the timing belt. On another design that might have caused an embarrassing metallic noise.

The belt on my car doesn't look much smaller than that on much bigger engines, so I suspect different manufacturers stress their belts to a different extent. Belts are available in various grades (three from Gates).

So between all that lot, you shrugs and read other people's experience.






Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[RossABQ]

I have heard there is a Mitsubishi-spec belt that is more durable, and is being used by some. I can't see anything that adds stress oveer any otheer OHC engine, although the sprockets are somewhat smaller in diameter, which directly relates to belt forces.

This engine has NO tolerance, in fact I am shocked at how thoroughly a job the pistons do on the valves -- double bends in the stems. Not cheap to repair!

 

[Bribyk]

The older Honda Prelude 2.0L FI engines would run with the timing belt out one tooth as well. Kinda lacked power though. I've heard of multiple people breaking a timing belt (at idle and at highway speed) on the Honda engines without damaging a valve, they must be zero-interference designs.

 

[evelrod]

I've had two belt failures in the past...one on a Fiat 124, twice! The first time, no harm, no foul. The second totally wasted the valves...I sold that sucker for twenty five cents!!! No joke. The Orange County Sheriff's deputy was going to have it towed and I sold it for 25cents.
Second was on our 83 Turbocoupe. As it turns out, the valves cannot reach the pistons on this Ford engine so no problem, just replace the belt taking care to line up all the pulleys, oil pump drive, dizzy, etc.
One other comment on reliability of the belt...When I rebuilt the 2.3 Turbocoupe engine to much higher hp (295 @ the wheels on the dyno) I chose the Goodyear belt recommended by JBA and changed it at 50k miles and it was still in the car at 100k miles when I sold it.

The last few cars I have purchased don't use belts...I just don't need the added aggravation at this point.

Rod

 

[dgallup]

Temperature will have a big influence on belt life just like any rubber component. The first gen DSM motors killed the timing belts on the turbo motors before 50k but they would go about double that on the NA motors. Mitsubishi had a service campaign and replaced the belts no charge with a higher grade belt.

I'm surprised Porsche couldn't get these to live. The devil is in the details.

 

[GregLocock]

Bear in mind that the forces in the belt are strongly affected by the camshaft torque and the speed of the engine as well. Pootling around at 2500 rpm all the time is different is to redlining at every traffic light.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[RossABQ]

Greg, that's what surprised me; at least three of the people on the Porsche forum I checked out had their belts fail at idle! Could be coincidence, but also makes me suspect a harmonic? The old GM 90-degree V6's in GMC trucks and Buicks used to break timing chains all the time due to the uneven cam loads.

It just ain't rocket science, I would have expected more from Porsche.

 

[BrianPetersen]

Idle speed is one of the worst-case conditions for variation in crank speed on each revolution.

 

[bub579]

Having replaced several 944 cam belts, I would like to comment. One didn't bend valves (broke at idle speed). The usual tensioning technique (quarter turn, moderate thumb pressure, longest span) would sing like a bird. Had to resort to adjusting at idle speed, not for the faint of heart. I suspect harmonics also. Good luck

 

[jcd06]

Is there any possibility to add a suplementary idler, idealy halfway the longest span?
Even with minimal tensioning it will reduce the belt orbiting or avoid it coming in resonance.

 

[RossABQ]

JCD, that is essentially what Porsche did in one of their attempts to fix the issue. They added a slipper-type guard on the water pump, to separate the two runs of the belt at their closest approach in the mid-span of the long tension-side run. They also made one idler a larger diameter. Looks to me like they suspected teeth on the two runs might be snagging each other when flapping around under variable tension -- just what you'd expect at idle.

They later went to a spring tensioner, then to a hydraulic one, but there is little improvement in belt endurance. Bottom line? They did not properly engineer the one system that needs to be bulletproof on an interference engine.

 

[jcd06]

RossABQ,

I just only noticed the picture with your first post.
The TB idler pulley prevents the belt to go orbiting but doesn't prevent it to deflect inwards.
The traction side of the belt goes straight from the crank pulley to the camshaft pulley. This is not the case in for example most of the DOHC engines.
If the idler pulley forces the belt to make an inward bend it will probably hold it better seated against the pulley. Unfortunately there seems to be very few place here for such a construction.
Is it possible to start the engine without the belt cover (or with one with a window) and film the belt with a high speed camera?

 

[RossABQ]

JCD, you are exactly correct, and Porsche addressed the problem accordingly in later versions. My take on the situation overall is this; there are several causative factors.

The original design had only static tensioners, so when the belt is properly set for running conditions, it is noticeably looser when the engine is cold (all aluminum engine). Many of the failures are idling in the driveway on initial start-up.

As you note, the tension run is long and largely unsupported. It also passes close by the opposing run, and at idle, I'm betting the long run flops around enough to actually touch the other run (cog to cog).

So what the engineers at Porsche did, in 3 distinct steps, was this: They added a guard rail, a smooth sheet of stainless, between the two runs at their closest approach. Then they replaced the static tensioner pulley with a spring-loaded one. Apparently that was also unsatisfactory, so they later made it a hydraulically-actuated (oil pressure) tensioner. (I said in my OP that they later went to a chain, that is incorrect)

The shield prevents contact, the active tensioners addressed the cold-to-hot variations.

What's also interesting is that this engine is half of the 928's V8, and the V8 models apparently were not prone to belt issues (because the two runs never approached each other closely?)

I found the maintenance records on my 944, which has 97,000 miles on it. It's on its 4th set of belts and tension rollers. The bill at each replacement was over $1,000 (parts and labor).

 

[PackardV8]

FWIW, I never owned an '84 944, but my '85 is run hard, belts changed as per schedule and never any trouble to this day. Also, I never knew of any reliable reports of dealer-installed OEM belts slipping or failing at 9,000 miles. IF this early failure occurred, I would suspect aftermarket belt, incorrect installation or adjustment There were improvements to later versions, but these cars are twenty-five years old now and most have been sold down the river to impecunious owners likely to ignore the schedule or use shadetree shops rather than the OEM dealer.

thnx, jack vines

 

[RossABQ]

Packard, most of my anecdotal info came off Rennslist and 944Online forums. Not the average "tuner" types on there, mostly pretty experienced D-I-Y'ers and racers.

I finished the water pump upgrade a week ago, with new belts, idlers, and tensioners. I tensioned the belts using my thumb and index finger, 90-degree twist method. I am about ready to open it up for the re-check after initial run, but no problems so far. The balance belt is quiet.

There really are no aftermarket belts, all are either Conti or Gates (OEM).

It's hard for me to accept that an '84 model is a 25-yr-old car, an antique in most states!

 

[evelrod]

You could say ---It's hard for me to accept that an '84 model is a 25-yr-old car, an antique in most states!

My wife's "new Lincoln" is the same age as MY first car, a '49 Mercury---and---at the time, I thought it was positively ancient, a vintage classic !!!

How time flies AND how much my attitude has changed.

Rod

 

[RossABQ]

Rod, I have a '52 F-1, near as I can tell it was made a month or so before I was born. I thoroughly enjoy seeing what things were like back then, for better and worse!

 

[evelrod]

I like the color. Needs a bit of pinstriping around the waist...thin red, maybe. Also a tailgate...I don't like to drive a truck without the tailgate and a good solid bumper!

Rod