Correct doing this on a none valve bending engine is on thing doing it on a 32 valve ......

^^ I have to agree ... new are cheap.

My take on the failure sequence.

The back of the impeller started to rust behind the stainless base and pushed the ceramic seal assembly back along the shaft, and broke the seal (sealant at the interface) between that stainless base and the back of the impeller, and between it and the shaft, which allowed water to leak along that path. The rust on the back of the impeller on the seal spot face would support that theory. Another good reason why a plastic impeller is better than a metal one ... this failure mode can't occur.

I'm assuming that a sealant is applied (it should be) to the back of the impeller, and perhaps some on the shaft, to stop water taking that path to finally travel along the shaft to the bearings. Perhaps on this pump that was not properly or sufficiently applied.

Water under pressure then travelled along the shaft in sufficient quantity to defeat the bearing seal, and got to the bearings.

The bearings got sufficient axial play to diminish the pressure of the carbon seal against the ceramic section and the seal leaked. The ceramic sections of seals usually fail because of overheating when run without full immersion or no water at all, which causes them to crack, which would have also contributed to the volume of leakage. This seal might also have cracked as it was being pressed apart.

More water got to the bearings ... and it was all over.

I like this. Way to go but trying this on a 32V is pretty risky.

Thanks for the replies! Also for the negative ones... If it fails, well that's a good reason to install an electric water pump. But then again people will comment about not being original etc. ;)

Anyway, some more info:
- The pump wasn't installed with any sealant, despite the graphite side of the water seal, which was fixed to a stainless cup. The shaft itself didn't have sealant; only rust...
- I don't think the pumps are cheap; 400 dollars is a hell of a lot of monay for some aluminium and steel (and sometimes plastic). I think it's a stupid price, when pumps for other cars are 50 bucks or so. On top of that, I have to pay for shipment and customs, as in The Netherlands the pumps are even more expensive.
- But since the pumps still fail (why would people otherwise replace them?), there's opportunity for improvement:thumbsup:
- it's going to be fitted on a 16v, so the risk is a little less than on the 32v's

I appreciate this if only to learn more about how these pumps are built. Carry on!

Bart Jan for the record the new Laso pump Roger has made are about the best value for the money you can get.
In the long run the new Laso will last and cause less issues after a few years in service,
so now its only your time your gonna save if you bend your own wrenches.

So far I have installed 8 of these new pumps and not one failure, YMMV.
Yes what your doing is educational for all that are watching,
but the real deal is to swap in a fresh Laso for continued performance.

After making the design on paper, I had to make a new shaft to make the parts fit and seal propperly. This one is made of high grade stainless steel and all ready has the first part mounted in a groove: a small circlip holding the bearings in axial direction. On the left obviously the new one; right is the old one.
http://i1185.photobucket.com/albums/...e/IMG_0016.jpg

Also, a holder had to be made for the water seal (rubber + ceramic disc) as it had a different outer diameter compared to the inner diameter of the pump housing. This part can be seen third from the right in this image:
http://i1185.photobucket.com/albums/...e/IMG_0018.jpg
From left to right: single row needle roller bearing with oil seals, Needle roller/angular contact ball bearing, oil seal, water seal holder, water seal (rubber + ceramic disc) and finally the graphite part of the water seal.

The fitting of the parts is very tight, so the shaft was put in the freezer overnight. The inner race of the bearing heated up in the oven:
http://i1185.photobucket.com/albums/...e/IMG_0019.jpg
The inner race looks overheated, but that's only the oily bit that was still on.
After that the temperature was stabilised, the single needle roller bearing was assembled with some high temperature water resistant lubricant. It was put in the freezer overnight again, before putting the second race on:
http://i1185.photobucket.com/albums/...e/IMG_0020.jpg

looking pretty awesome

Keep it up Bart-Jan i love it. hope this works great.Ray

A few questions.
 

Nice work!

If I interpreted the original pump parts configuration correctly, the original had the ceramic half of the seal against the back of the impeller and the sprung carbon half in the pump body ... which is a normal pump seal configuration and allows water to cool the carbon/ceramic interface.

If I've read your post correctly you're putting the ceramic into your new adaptor in the pump housing and the sprung carbon against the back of the impeller ... have I got that right? If so it will mean that the ceramic is deep into the adaptor and will be a little shrouded from water circulation ... which could mean higher carbon/ceramic seal interface temperatures. Might be an issue ... might not.

Also, how have you established seal compression? Did the seal supplier give you an optimum seal compression figure for you to establish set up?

How will you be locating and sealing the sprung carbon seal assembly on the back of the impeller?

Thanks for posting.

Dave, You've analysed my design correctly. Strangely, all industrial water seals with ceramic discs I could find had the ceramic disc attached to the housing and the graphite part rotating, so to me it ssems strange that the original Laso pump had it the other way around. I agree that there'll probably be less water circulation around it, but since water is a very good conductor of heat, I'm not too worried of issues coming from that.

You are correct that there's a pre-determinded pressure (or move in the spring) that I have to commit to. Luckily, it could adapt for quite some tolerance, making production of the parts easy. Anyway, I've suited the design of the cup of the ceramic disc accordingly.

The sprung carbon seal has a rubber seal on the inside, that seals directly on the shaft, while positioning it and increasing the seal to the shaft during compression of the spring. Really a very nice and well thought of design by the pump manufacturer. It's a very nice setup, that water seal @ only 8,50 Euros...

OK, so I don't think I'll be trying to duplicate this rebuild. The most sophiticated metal cutting tool I have is a hacksaw. :)

To assemble the shaft assy from the last photo into the housing, an adapter piece needed to be made on the lathe to only press the outer race of the bearing and not putting any force onto the shaft itself. Here you can see it pressing in the assembly:
http://i1185.photobucket.com/albums/...e/IMG_0021.jpg
During pressing, I kept a close eye on the positioning, which I measured with a pair of callipers.

When pressed, it looked like this, with these pieces still to be assembled:
http://i1185.photobucket.com/albums/...e/IMG_0022.jpg

And again a step further:
http://i1185.photobucket.com/albums/...e/IMG_0023.jpg

Here, the water seal is completed:
http://i1185.photobucket.com/albums/...e/IMG_0024.jpg

Then, it was time to go over to the outside of the pump, to assemble the needle roller bearing with integrated oil seal:
http://i1185.photobucket.com/albums/...e/IMG_0025.jpg

After that, the spinner (or whatever it's called):
http://i1185.photobucket.com/albums/...e/IMG_0026.jpg

Finally, the impeller went on:
http://i1185.photobucket.com/albums/...e/IMG_0027.jpg

And to be sure it stays on:
http://i1185.photobucket.com/albums/...e/IMG_0028.jpg

Ready!

I still think that shrouding of the seal could give you overheating and premature seal failure ... hope I'm wrong.

:icon501:

Glad you have a 16V.