mark kibort

plastic impeller came off due to a poor design. the knurling on the shaft and plastic casting is so fine, that the plastic over time will give way to the shaft and the impeller then spins freely around the input shaft. maybe high rpms, stress over the first couple of years was too much for it, but looking at the failure, it was pretty apparent that it is a poor design. However, I have not heard of any metal impellers failing this way, but if the bearing goes, then the metal imeller can shift its orientation and start digging into the block.
Such simple technology, such stupid designs!!

anyway, the holbert car had an original pump that i should have left in place!!!! instead, we changed the pump to a used pump, it leaked, then did the entire job again with a new pump, and it lasted 2 seasons and the impelller blew. now i got a new pump again , after the bearing started to go and was making some noise when cold. new pump in now and no issues. Hey, ive gotten real good changing pumps and doing timing belts though!

Mk

6.0-928S

When I worked at Holbert's, Cardone used to order new parts from me as they came available every year & tool up for their rebuilds. As soon as Porsche came out with a new part or year change mod they'd call & order one. They wanted to be ready when requests started rolling in. I was at their plant (I was the wholesale manager) & they seemed to be a quality operation. That was 10-12 years ago.
FWIW.
Hammer

jns

If I had the time, money and a 928 I'd build an electric water pump system for it since I can't stand the thought of a water pump failing and causing a valve train disaster. I built one for my 88 951 and you 928 guys have twice as many valves to lose. I saw a black shark with those wide rear lights(maybe an S4?)and knew I'd have to have one someday. Awesome cars!

mark kibort

the pump isnt the real problem, its the bearing. so, you change to electric, and now you have to design your own bearing roller to keep the belts original design. i think i would trust these pumps more than some new design for a new belt roller bearing design.

all in all, its a decent design, just very unforgiving for defective equip and neglect.

MK

toofast928

Fabio421, Ah ment to spell "impeller"
And Bill is right on the money. The metal impeller will gouge the block if the bearing begins to fail and the shaft gets some runout. This means a water pump failure could cause the block damage! The plastic impeller won't damage the block if the same situation.
The main reason for water pump failure on the 928 is sudden acceleration in the low gears. The timing belt side loads the water pump shaft, damaging the bearing.
Preventive maintenance is to inspect the water pump weep hole. If it weeps coolant, it's time to replace the pump.

heinrich

You're joking right? This is not correct.

toofast928

Nope not kidding, full throttle first and second gear shortens the life of the WP.

heinrich

Hmm no Dude, not so . Water pump life is shortened only by stress from the belt, which happens only when the belt is not tensioned right. Too tight or too loose.

mark kibort

i think thats a stretch. (no pun intended!)

I think the reason that its probably not a factor as there may be some calculation of the radial load based on the different acceleration rates in first or second gear. (we are talking times of 1-3 seconds) I wonder what difference of side loading (radial load ) on the bearing static vs under acceleration. I wonder what the different loads would be. If it was a factor, down shift matching revs would be much worse!

interesting.

MK

heinrich

The speed (why the f am I getting into another one) .... of the belt doesn't change axial strain. It's a too-tight belt that constanty forces, and a loose belt that violently attacks, the axis of the pump roller. Trust me, I've dealt with a few faiures and NOT ONE was a result of actually using lower gears. ALL were the result of poor maintenance.

Bill Ball

Then mine should have failed a long time ago.

SharkSkin

OK, now wait a minute. It may sound like nonsense to some, but in point of fact the valve train is a drag against the crank at all times, more so while accelerating. Let's look at this. If your engine is steady state, 3000 RPM, not accelerating or decelerating, there is some tension on the belt. Max tension would be on the segment of the belt between the crank and the oil pump. Slightly less tension on the segment between the 5-8 cam gear and the oil pump. Less for each segment, on around till the segment between the tensioner and the crank, which has the least tension on it. Now, this tension is all essentially constant at a constant speed, and is essentially the sum of the drag in all bearings, oil pump resistance, and valve springs -- disregarding the lumpiness that the cams would impart to a graph of load/time on each belt segment. This load would be nearly the same at most any engine speed, disregarding the slight increase in drag at higher RPM. This is what Heinrich means(I think) by "the speed of the belt doesn't change the axial strain".

Now, as you accelerate, things change. You not only have the aforementioned drag, you also have additional load on the belt as you induce force into the moving assemblies to overcome their inertia. So, during the time that you are accelerating from 3K to 4K RPM, the load increases by the amount of inertia that you are trying to overcome over a given time period. As the time taken to transition from 3K to 4K approaches an infinitely small duration, the load on each segment of the belt approaches an infinitely large value. To say that the belt/WP/OP/etc will be under a higher load when accelerating in first gear than in second or any other gear(and so on) is absolutely true.

That's simple physics, you can take that to the bank. The faster you accelerate, the more load there will be on all parts, Yes, including the water pump bearing. Will it make it wear out faster to redline more often in lower gears? Maybe so, maybe not. My guess is that it probably will, simply because of the higher load. But how much faster? I have serious doubts about whether the difference could really be measured, and doubts that the difference would really be that significant.

I for one don't really care if I take 5% off the potential life of the water pump by having a little fun in the lower gears. But Tony, if you have data to back up a 30% or higher reduction in pump life I'd sure love to see it!

jns

the pump isnt the real problem, its the bearing. so, you change to electric, and now you have to design your own bearing roller to keep the belts original design. i think i would trust these pumps more than some new design for a new belt roller bearing design.

all in all, its a decent design, just very unforgiving for defective equip and neglect.

MK


I don't have a 928 so I don't know how the timing belt and water pump are configured. On my 951, I don't have a timing belt roller there at all and have a different kind of belt that is re-routed so there are no pulleys to cause it to fail, just age and mileage on the belt itself. Now I can drive without worrying when my engine will grenade on me. I'll have to look at a 928 engine to see what it would take to convert it like my 951.

SharkSkin

jns, below is a pic of the layout(lifted from PorKen's page). All 928's have essentially the same layout, whether 16V or 32V. In order of proximity to the crank:

11:00 position - tension roller
1:00 position - WP pulley
3:00 position: Oil pump
Then of course, the cams.

heinrich

Dave, I'd not be so sure. Remember, the enemy here is inconsistent OR excessive belt tension in its length, because this is what pushes OR oscillates he axis of the pump against the travel direction (90deg). So ... knowing that most tbelt failures occur at low engine revs (makes perfect sense thinking inertia), and adding the factor of compression resistance at low revs, I would have to conclude that it is healthier for your water pump to drive always at higher revs .... and because the lower gears achie that on average more quickly .... I would have to conclude that the exact opposite (of originally stated theory) is true. low gears == less water pump failure.