Coolant Pump Sealing / Design Practice

[ivymike]

I was performing some repair work this weekend on a friend's van, and I noticed some discrepancies between the way the vehicle was designed and the way I thought it should have been done. I was hoping that someone with some experience in this area might cast some light on whether the things that I thought should be done are really as important as I make them out to be.

The following are things that I noticed with regard to the water pump mating surfaces on this particular vehicle:

1) The interface between the water pump and the block used a fiber gasket between two metal surfaces; both metal surfaces were aluminum. The water pump surface that mated to the block was a rough "as-cast" surface. The block surface was machined (quite smooth when I got done removing the old gasket). I would have thought that machining the mating surface of the water pump would be a good idea. The gasket was a little under 1/8" (3mm) thick.

2) A metal tube inserted into a cavity on one side of the water pump. The tube was smooth and painted, and an annular o-ring was used to seal between the tube and the water pump. The water pump surface was as-cast (fairly rough, again), and as a result it was extremely difficult to insert the tube w/ o-ring into the cavity on the water pump. I would have thought that drilling this hole would have made a much better seal (less scuffing of the o-ring, fewer cavities to fill, etc)

3) Six bolts were used to hold the water pump to the block. The water pump gasket surrounded each bolt hole on both sides. Three of the bolts were long, and three were short. The long bolts penetrated all the way into the block water jacket, and coolant could flow easily from each hole when the bolts were removed. There were no special sealing features on the bolt heads (just washers). I would have thought that the bolts would either have to be isolated from the coolant, or they would require gasketed heads (think drain plug gaskets).

After assembly, the thing didn't leak. Interface #1 was leaking before the repair, which is why the repair was performed in the first place. I can see interface #1 sealing fine with enough gasket "smoosh" and a little luck. I can see interface #2 sealing fine if the o-ring was oversized enough that getting skinned up by the rough surface would not be a problem, and if the installer was lucky. I just don't get what keeps the bolts at #3 from leaking, barring sheer dumb luck. Those washers didn't look like the right way to do the job.

Comments?

 

[evelrod]

It's a giant conspiracy by Permatex, et al, Isaac.
Did the pump have a plastic (a la GM) impellar, too?

Rod

 

[ivymike]

nope, it was a metal impeller. The old pump looked fine, despite being a bit dirty. It seemed like a shame to replace it when the problem was just a leaky gasket, but it took the whole darn day to do the job (I HATE working on vans), and I figured going into it that replacing the pump while I was in there would save trouble later.

 

[Chumley]

Ivymike:

One thing I always try to remember when working on my
own cars is that when the folks originally assembled the
systems, they got them to seal up by doing it a certain
way, often a 1,000 times a day with better than 99.95%
success. That's one reason why I don't worry if I see
something that doesn't initially look quite right to me,
(in my humble opinion) What I don't like is when I see a
part designed radically different from what I'm removing
that has casting ID marks, a part numbering system, or
fittings on it that are clearly different from the OEM part
it is replacing. 9 times out of 10, that means trouble on
the way.

Of course nothing is perfect, but often, if parts are
getting damaged during assembly, or are difficult to
assemble, its because of the process. Going to a factory
and watching someone repeatedly do in 15 seconds what it
takes any one else 15 minutes to do can be eye opening.

It's quite common for threaded holes to daylight into water
passages, exhaust pipes, pressurized oil galleries and the
like. They normally seal up just fine as long as they are
torqued properly using the correct crush washers and/or
sealant is used and the sealing surfaces aren't damaged.
Oil drain plugs, spark plugs, and external oil/water pump
mounting bolts come to mind. Often the factory service
manuals will have notations on certain applications to the
tune of "do not reuse, always replace" or they specify a
certain thread locking/sealing compounds when applications
are critical.

Chumley

 

[patprimmer]

Isaac

Good old GM sealant does an excellent job of sealing threads which go through to water jackets.

Although the factory assembly line can quickly and reliably do a job, they are doing it under ideal conditions with everything dry and brand new and in a comfortable well lit environment, not hanging upside down with rusty water, oil and grease everywere, not enough room to move and restricted vision due to lack of access and light.

I would have thought that at least rough machineing of the pump mounting surface would have ensured accurate belt alignment

Regards
pat

 

[Chumley]

Hi Pat

"Although the factory assembly line can quickly and reliably do a job, they are doing it under ideal conditions with everything dry and brand new and in a comfortable well lit environment, not hanging upside down with rusty water, oil and grease everywere, not enough room to move and restricted vision due to lack of access and light."

OK. . . hold on for a second, . . . . it's all coming
back to me now. There, that's it!!! I think I worked in
a "different" factory than what you are/were familiar with!

It does help though to think of the factory assembly
process. So often, it's soooooo much easier to spend an
extra 10 minutes and remove something in the way, then
spend 10 minutes later to replace it, than to spend 2
hours working around it!

Hope this helps!

Chumley

 

[patprimmer]

Hi Chumley

OK, not all factories are an ideal environment and some are quite poor, but it is still a lot easier to assemble parts on an assembly line where the sequence and method are studied in detail and optimised, than it is with the engine between the chassis rails and surrounded by body and suspension parts, and were the correct sequence is unknown or not applicable.

I agree entirely with your comment that it's easier to remove more components than most repairers are prepared to do, so as to improve access. It not only often ends up faster, but also reduces dangers of injury and greatly improves the chances of a first rate job.

Regards
pat

 

[evelrod]

Back to the original subject---Leaking waterpump gaskets are not your typical problem, seal, yes---not usually gaskets. Most of the OEM gaskets are paper and assembled dry but some of the replacement gaskets are thinner than original (not usually a problem with addition of a sealer). The problem comes from non OEM parts (often) and/or not cleaning the mounting surfaces/mishandling of the new gasket. ie, improper installation (usual). Vans are a tough nut at best but removing the grill and rads can improve the situation regarding the water pump. Thankfully pumps last a long time given proper coolant service---just like changing the oil---the coolant needs service periodically to maintain proper lube and PH levels.
Besides, Isaac, the difficult jobs are the ones you will remember---trust me.

Rod

 

[strokersix]

Everyone seems to be going around the real reason for not machining the features in question. Design engineers at OEMs are under tremendous pressure to reduce product cost. If eliminating machining operations can be done with minimal impact on warranty costs then they will be eliminated.

 

[Marquis]

Thank you Stroker Six for talking sense and seeing the bigger picture!

 

[ivymike]

Thanks to those who helped to answer my question about the importance of machined sealing surfaces. I guess that it is fairly common amongst auto manufacturers to employ rough sealing surfaces such as those I mention above, and drillings that penetrate the coolant jacket, and they seem to get away with it often enough. Clearly it didn't work out so well in the particular case I was looking at (who ever heard of a water pump gasket failing before the pump? certainly not with an o-ringed water pump like Honda and others use), but perhaps that is a calculated risk from a warranty standpoint. The van was a whole 5 years old after all, with nearly 80,000 miles on it, so at least the seal made it out from under the warranty before it failed. Then again, I don't know whether the pump in question had been replaced earlier (but I sure hope not).

It's funny how we often end up looking at big pictures, vacation pictures, and all sorts of other pictures around here. Cost-based arguments don't really address the functionality of the parts in question, but I suppose that they'd be an important secondary discussion. I'd guess that most of the legitimate members of the site would agree that cutting cost is the reason for not machining the features. It would be interesting to hear from anyone who thinks the seals work better this way.

 

[evelrod]

Ok, a couple. I do not use super smooth finish on the head/block surface of my engines and I deliberatly roughen (scratch?) the mating surface betweed the block and transmission case of my 1380cc Mini race engine. Both surfaces have had gasket "migration" problems that have been effectively cured by using rougher finishes on the mating surfaces along with a suitable 'sealant' in the case of my bottom end. Given the correct gasket, most any finish can be sealed. Most newer assemblies, well, many at least, are being assembled sans gaskets effectivly. I have had few problems with this method (I am glad to see all those old cork gaskets go away) except in the pan/sump cover of certain UK Ford products which still need that gasket to seal.

Rod

I think most of the 'end users', especially those of us who have been able to "see" the other side agree that the $$$$$$$$ factor is THE overiding concern of management these days. Sorry, Marquis, I certainly did not mean to ignore this factor.

Rod

 

[patprimmer]

Marquis and Strokersix

Also sorry for not mentioning the cost down pressurers that auto company engineers and there suppliers are under.

At least here in Oz, the pressurers were very strong, with at least one engineer from a local manufacturer telling me they considered a perfect design was trouble free for the warranty period, but failed the next day. This was in 1988 and they offered 1 years warranty

By 1990 the same people were complaining of looseing market share to Japanese imports. They now offer 3 years warranty and last I heard were designing for at least 5 years life. They are winning back market share.

Also, I think that the OEM engineers design parts to be fitted on an assembly line, and give some but minimal consideration to insitu repairs. This policy is probably comercially viable as if they did their job right, most repairs should be done when the car is an old clunker, thus having minimal impact on the reputation of the new models.

I also agree with Rod. A smooth surface is not ideal for holding a gasket, but a flat surface with reliable dimension is helpfull in maintaining good belt alignment.

Regards
pat