Mike Frye

Thanks Bill. I don't know if it's possible on the impeller side, but I can see a reason to put a pin in the pulley side in a racing application, or even for a car that's frequently run in DEs.

Blown 87: Can you tell me which end came off the shaft in this case?

blown 87

I called one of my friends last night and he came down to the shop, he is a mechanical engineer with a lot of common sense.

We looked at it over a few beers and we have both come to the conclusion that the failure was that the shaft migrated aft in the bearing. This means that both the pulley and the impeller moved closer to the block.

There is also evidence of the impeller hitting the water pump housing in the front, so it looks like it had walked both ways on the bearing.

I am now left with the question, metal or plastic.

Mike Frye

So the shaft spun and freed itself from BOTH the pulley and the impeller?

blown 87

No, the impeller, pulley and the shaft stayed as one piece, it looks like they moved in the bearing, even though I am not sure how they could have.

Imo000

Nah… a spot weld (1 sec long) will hardly generate enough heat to warm the shaft up. The seal and bearing are so far away form the end of the shaft that the welding wont even reach them.

AO

Sounds like a political answer to me... C'mon Bill! Take a stand!

blown 87

Hell, I would vote for him.

Mike Frye

So there's play if you push on the whole assembly from front to back? OK, I think I can picture it. Thanks.

In this case fixing the impeller or the pulley (or both) to the shaft wouldn't have helped. So what's the fix? Anyone?

blown 87

That is one of the problems, there is no play.
I had to have either had the whole shaft move which I do not see how it could, but it sure looks like it did, or the impeller has moved back and forth on the shaft.

You would think that the shaft would have a step on it so the shaft could not move aft at the bearing, and the bearing cant move back because of the rear of the housing.

To me it would be a simple thing to make a pump that this could not happen, ie no movement unless the bearing failed.
Pin the pulley and impeller and three snap rings at the bearing, one to hold the bearing in the case and two to locate the shaft at the bearing.

I would think that some of our more talented members with machine shops could do this.

Mike Frye

I don't know, I'm just concerned about this weak point in a pretty complicated system.

I plan to check the paint mark on the front of my water pump the next time I check the tension. It stinks because I'll have to pull the center cover for that and I wouldn't have to do that otherwise.

John V, do you still have that center cover from your track car? Maybe you need to leave that little round section above the pulley disconnected so you can check the pulley/shaft status.

blown 87

There has to be a better setup (Mike, Porken or many others????) than we have right now.

mark kibort

Good point about the one type of failure where the impeller, WITH THE SHAFT, walks toward the engine. I guess, nothing we can do about that! However as long as both shafts are welded, the furthest it could walk inward would be the distance of the front pulley and the water pump housing.
I forgot to look to see scots pump, if the shaft had moved inward off the front pulley. it doest take much movement to start grinding the block. I think welding both sides would help here.
I think the most common failure is the impeller just coming off the shaft, at least it has been this way with the failures ive had.

one of the reasons the pump fails on race cars, is the sheer number of times the revs go from 4000rpm to 6000rpm. the load on the pump is pretty substantial. eventually the force weakens the press fit, and the pulley comes off. in the industrical world, you can do two things (or two things are usually done) deeper knurling of the press fit. deeper knurling on the plastic cast part, or roll pins/set screws.

What a stupid type of failure! of all the great things porsche designed into the 928, this surely wasnt one of them

mk

Mike Frye

+1, but then again, I'm sure the original pumps were spec.ed out properly, the aftermarket may not have carried this over, not knowing that it's a potential point of failure.

Does anyone know how much force it would take to spin the pulley on the shaft?

Does anyone think this could be a temperature thing? Maybe the pulley or impeller is just expanding because of heat and the shaft isn't?

How about if we try to gather information about all of the ones that have failed and see if there's anything common?
Was it only in track cars?
Is it all the same rebuilder?
Is it with only metal or only plastic impellers?

ROG100

Well I just looked at 3 different pumps.
Original Porsche, Laso & Rebuilt Porsche.
All have metal impellars.

All have the same assembly of shaft, roller and impellar.
All seem to be interferance fit only.

A spot weld would do the trick (as already said) on both ends.

What about the disimilar materials and welding. The shaft seems to be machined and hardened SS, the impellar looks like cast something and the roller is machined SS, either bare metal or plated.

AO

Here's an interesting bit of trivia I happened upon today. It appears the **** to plastic impellars happened in 1990. This was from the 1990 "Service Info by Year - 1990" pdf from Jim's CD.