Weird raw water pump issue (Universal 40)

[Second Star]

An out of the box thought: if the shaft key was too small for the slot machined into the shaft (which is kinda crescent moon shaped and not even depth over its length) then at odd times it would "disappear" into the slot and the impellor wouldn't turn. Next time, or many times after it would engage the impeller. Seems weird but if you use the same key with many different impellers and nothing else has changed ...

 

[Joliba]

Eric,
I’ll repeat what I said before along with a suggestion. Sometimes during the first spring start up, my pump will not draw water even if the hose to it is primed with water. Adding PTFE grease is an instant cure that lasts all season for me.
If your problem is that the water drains out your thru-hull somehow, there might be an air leak between the thru-hull and the pump. The most likely spot would be the gasket or the drain on your raw water strainer.
These are very easy things to check before removing or replacing your Sherwood pump.
Mike

 

[EricFox]

Hmm, interesting idea - I'll check that when I take it apart this week.

 

[EricFox]

Well, I finally have a hypothesis about what is happening and to share and what looks to be a solution.

I brought the pump home, bought some clear tubing so I could see where water was (or was not) flowing, made a tool to drive the pump with my drill, then began to play. When pulling on air (i.e. startup) the pump can't create enough vacuum to pull the water column up unless I really get the RPMs up. If I can get to a point where all of the air in the pump chamber has been removed then it pumps, but clearing it of air isn't happening readily. My first suspicion was the seals to the outside - O-ring on the one side, and paper gasket on the other. But there are no leaks there. However while playing around I noticed that water inserted at the input barb leaks out the output barb, and vice versa. That suggests a leak within the pump chamber that would act as a "short circuit" between input and output when pulling on air. I confirmed the leak by getting the chamber filled and then, with hoses attached, adding water to input or the output hose - in each case the added water flows through the pump to equilibrate the levels in the hoses.

There is no obvious damage in the chamber to account for the leak - my best guess is that over the years the metal at the bottom of the chamber has been worn away just enough that the chamber is now slightly wider than the impellor. In fact, if I remove the O-ring the ability of the pump to lift water improves somewhat.

That gave me an idea. The pump has been installed on the engine oriented such that the input and output are at the bottom. If water falls out of the input tube when heeled over then the chamber becomes filled with air. And with the through hull back in the water, the short circuit prevents the pump from pulling up water. But if I re-orient the pump so that input/output are at the top, then I should be able to keep the chamber filled with water. And since water is much more viscous than air, then what is a short circuit within the pump for air may not be a short circuit for water (recall that once filled with water, the pump actually pumps). I re-oriented the pump on the bench and found that I can in fact pull water up the input hose that is initially filled with air, even at very low pump RPMs. And once that air column is removed the pump flows water to the output as it should.

I could try to remove some material from chamber to make it narrower, but since the pump seems to be flowing adequately in the new orientation, I think I'll try that first rather than risk damaging the pump. I have room on the front of the engine to re-orient the pump. Longer term I may need to remove material (or buy a new pump) if the gap within the chamber increases to a point where the gap is acting as too much of a short circuit even for water, but I expect that is a good many years away - probably the pump will fail before then at the seals surrounding the rotating parts.

Thank you to all of you for the helpful suggestions.

 

[Tooluser]

BraVO! That’s fantastic work. Thanks for sharing the result and the process.

 

[K2MSmith]

Well, I finally have a hypothesis about what is happening and to share and what looks to be a solution.

I brought the pump home, bought some clear tubing so I could see where water was (or was not) flowing, made a tool to drive the pump with my drill, then began to play. When pulling on air (i.e. startup) the pump can't create enough vacuum to pull the water column up unless I really get the RPMs up. If I can get to a point where all of the air in the pump chamber has been removed then it pumps, but clearing it of air isn't happening readily. My first suspicion was the seals to the outside - O-ring on the one side, and paper gasket on the other. But there are no leaks there. However while playing around I noticed that water inserted at the input barb leaks out the output barb, and vice versa. That suggests a leak within the pump chamber that would act as a "short circuit" between input and output when pulling on air. I confirmed the leak by getting the chamber filled and then, with hoses attached, adding water to input or the output hose - in each case the added water flows through the pump to equilibrate the levels in the hoses.

There is no obvious damage in the chamber to account for the leak - my best guess is that over the years the metal at the bottom of the chamber has been worn away just enough that the chamber is now slightly wider than the impellor. In fact, if I remove the O-ring the ability of the pump to lift water improves somewhat.

That gave me an idea. The pump has been installed on the engine oriented such that the input and output are at the bottom. If water falls out of the input tube when heeled over then the chamber becomes filled with air. And with the through hull back in the water, the short circuit prevents the pump from pulling up water. But if I re-orient the pump so that input/output are at the top, then I should be able to keep the chamber filled with water. And since water is much more viscous than air, then what is a short circuit within the pump for air may not be a short circuit for water (recall that once filled with water, the pump actually pumps). I re-oriented the pump on the bench and found that I can in fact pull water up the input hose that is initially filled with air, even at very low pump RPMs. And once that air column is removed the pump flows water to the output as it should.

I could try to remove some material from chamber to make it narrower, but since the pump seems to be flowing adequately in the new orientation, I think I'll try that first rather than risk damaging the pump. I have room on the front of the engine to re-orient the pump. Longer term I may need to remove material (or buy a new pump) if the gap within the chamber increases to a point where the gap is acting as too much of a short circuit even for water, but I expect that is a good many years away - probably the pump will fail before then at the seals surrounding the rotating parts.

Thank you to all of you for the helpful suggestions.

My pump doesn’t always seem to prime easily and wears out impellers in less than a year and it is also mounted upside down . So maybe you’re onto something. Btw, did you check to make sure the impeller spins with the cover off ? Maybe an obvious question but that was a problem I had with a new impeller they was defective.

 

[Loren Beach]

There is no obvious damage in the chamber to account for the leak - my best guess is that over the years the metal at the bottom of the chamber has been worn away just enough that the chamber is now slightly wider than the impellor

In the "FWIW" department of experience, on our prior engine my mechanic advised me to replace the OEM Oberdorfer pump after about 20 years. Our river water picks up very fine, but sharp, glacial sediment on its way from Canada to the sea and this slowly erodes the pump chamber and reduces the throughput. We also had to lap the cover a couple times to get rid of grooves and improve the flow.
Perhaps not relevant to your environment, but worth noting, seems like.
This subtle abrasion of the pump housing can reduce the seal and also the flow. I recall that after installing the new pump the water exiting out the exhaust visibly increased in volume.

 

[bigd14]

I recall that after installing the new pump the water exiting out the exhaust visibly increased in volume.

Yes! Major increase in flow with a new pump installed this past spring. I was quite surprised at the difference.

 

[EricFox]

K2MSmith, yes, I confirmed that the impellor is rotating properly. And I was getting the same problem with old and new impellors. I'm fairly certain the impellors aren't the problem.

Loren, you discovered our sneaky plan to fill the water that we send south with abrasive material. But I don't think we have yet been smart enough to figure out how to make that work to our advantage in plans for world domination! On a more serious note, I recall seeing a post a while back about lapping the pump - probably it was one of your posts - that alerted me to the fact that wear within the pump could be an issue. Thanks for that!

bigd14, if this doesn't work then I will replace it with a new pump. And I expect that will solve the issue regardless of orientation. For now, the pump is flowing more than enough water to keep the engine cool once it gets rid of the air lock.

Joshua, thanks - it was a long journey that started last spring with multiple hours cramped in the engine compartment when my first suspicion was that there was a blockage in the cooling system downstream of the pump. It's a good thing that I am recently retired and have time for these pursuits ;-)

 

[EricFox]

Just closing the loop on this - I got through the entire summer with zero problems - the pump drew like a champ every time I fired up the engine, even after significant heeling. I don't know how long this fix will last, but turning the pump upside down has allowed me to stave off the purchase of a replacement pump at least for a while.